WO2018111941A1

WO2018111941A1 - Compositions of high affinity reagents and methods of use thereof

Info

Publication number: WO2018111941A1
Application number: PCT/US2017/065914
Authority: WO
Inventors: Stephen P.A. Fodor; Ari CHANEY
Original assignee: 13.8, Inc.
Priority date: 2016-12-12
Filing date: 2017-12-12
Publication date: 2018-06-21

Abstract

The present disclosure provides methods for developing high affinity reagents to alter cellular interactions and high affinity reagent compositions thereof. The present disclosure provides methods for using molecular barcoding techniques to identify target molecules in single cells, methods for employing phage display techniques to screen, enrich, and identify high affinity reagents to the targets, and methods to alter cellular interactions with the high affinity reagent compositions identified therein. Altering of cellular interactions can lead to modulation of biological responses, such as an immune response or antitumor response.

Description

COMPOSITIONS OF HIGH AFFINITY REAGENTS AND METHODS OF USE

THEREOF

CROSS REFERENCE

[0001] This application claims the benefit of U.S. Provisional Patent Application No.

62/433,205, filed December 12, 2016, the entire contents of which are incorporated by reference.

BACKGROUND

[0002] Cell surface proteins can play a critical role in mediating signaling between interacting cells. This interaction can lead to cellular activation, proliferation, and differentiation, or conversely, can dampen or inhibit cellular activation, proliferation, and differentiation, in which the particular outcome of the interaction can be dependent on the repertoire of cell surface proteins present on each cell. Therefore, the cell surface protein repertoire on the interacting cells can influence the cell-to-cell interaction and impact a resulting biological process, such as an immune response.

SUMMARY

[0003] In some embodiments, a method of producing a high affinity reagent to a target comprises: a) identifying a target involved in an interaction of a first cell with a second cell by comparing a first cell surface protein repertoire of the first cell with a second cell surface protein repertoire of the second cell; and b) generating a high affinity reagent that binds the target.

[0004] In other embodiments, a method of altering an interaction between a first cell and a second cell, the method comprises: a) identifying a target involved in an interaction of a first cell with a second cell by comparing a first cell surface protein repertoire of the first cell with a second cell surface protein repertoire of the second cell; b) generating a high affinity reagent against the target; and c) administering the high affinity reagent to the first cell and the second cell.

[0005] In some aspects, the first cell surface protein repertoire and the second cell surface protein repertoire are identified by molecular barcoding and high-throughput sequencing.

[0006] In various aspects, a cell surface protein in the first cell surface protein repertoire is identified by a membrane signal sequence. In some aspects, a cell surface protein in the second cell surface protein repertoire is identified by a membrane signal sequence.

[0007] In some aspects, the high affinity reagent is generated using phage display. In further aspects, the high affinity reagent is an antibody. In some aspects, the high affinity reagent is a small molecule. [0008] In various aspects, a IQ of the high affinity reagent is less than 10 nM, less than 1 nM, or less than 0.1 nM.

[0009] In some aspects, the target is a receptor. In some aspects, the target is a ligand of a receptor. In various aspects, the target is expressed on the first cell. In various aspects, the target is expressed on the second cell. In some aspects, the target is secreted by the first cell or the second cell.

[0010] In various aspects, the first cell is an immune cell or a tumor cell. In various aspects, the second cell is an immune cell or a tumor cell.

[0011] In some aspects, the first cell is a T cell. In some aspects, the second cell is an antigen presenting cell.

[0012] In various aspects, the altering of the interaction comprises enhancing a biological response. In various aspects, the altering of the interaction comprises inhibiting a biological response. In various aspects, the altering of the interaction comprises dampening a biological response.

[0013] In some aspects, a biological response is an immune response, an anti-tumor response, cell proliferation, or cell apoptosis.

[0014] In some embodiments, a composition comprises a high affinity reagent, wherein the high affinity reagent is produced by the methods of the previous embodiments. In some aspects, the composition further comprises 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or more high affinity reagents, wherein each high affinity reagent binds to a different target.

[0015] In some embodiments, a pharmaceutical composition comprises the composition of any of the preceding embodiments and a pharmaceutically acceptable carrier.

[0016] In various embodiments, a method of treating a subject in need thereof comprises administering a therapeutic dose of the composition of any of the preceding embodiments or the pharmaceutical composition of any of the preceding embodiments.

[0017] In some aspects, the subject has cancer, an autoimmune disease, or dysfunctional cell signaling.

[0018] In various aspects, the composition or pharmaceutical composition is a administered intravenously, cutaneously, subcutaneously, or injected at a site of affliction.

INCORPORATION BY REFERENCE

[0019] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The novel features of the invention are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present invention will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the invention are utilized, and the accompanying drawings of which:

[0021] FIG. 1 illustrates an example strategy of identifying a cell surface protein (e.g. , a TCR as shown in this figure), screening for high affinity antibodies using phage display techniques, and administering high affinity antibodies to alter cellular interactions.

DETAILED DESCRIPTION

[0022] Disclosed herein are methods for developing compositions of high affinity reagents to mediate cellular interactions and the compositions of high affinity reagents thereof. The high affinity reagents can be developed by using barcoding techniques to identify cell surface protein repertoires of individual cells and then employing phage display to generate high affinity reagents that target molecules based on the cell surface protein repertoire. These high affinity reagents can be used to enhance or inhibit signaling for a cellular interaction. A composition of these high affinity reagents can be used as a therapy to manipulate subsequent biological processes, such as an immune response, that results from cellular interactions.

[0023] Cell surface proteins can play a critical role in mediating signaling between cells. This interaction can impact subsequent biological processes, such as stimulating cellular activation, proliferation, and differentiation, or conversely, dampening or inhibiting cellular activation, proliferation, and differentiation. The cell surface protein repertoire of a cell can influence the outcome of the cell-to-cell interaction. Therefore, altering how cells interact can affect the outcome of biological processes, such as an immune response, which can potentially be manipulated for the development of new therapies. However, the repertoire of cell surface proteins for an individual cell, and thus the combinatorial effect of the specific cell surface proteins that are present in the cell surface protein repertoire, has not been well characterized. Therefore, a streamlined process for identifying the cell surface protein repertoires of individual cells and then generating therapies to mediate interactions between these cells based on their cell surface protein repertoires is needed. The present disclosure provides methods of identifying the cell surface protein repertoire for generating high affinity reagents that can alter cellular interactions and provides compositions of these high affinity reagents thereof.

Cellular Interactions

[0024] Cellular interactions can play a key role in propagating molecular signals, which can lead to functional outcomes such as an immune response or tumor growth. The present disclosure provides methods for identifying targets that can be used in the generation of high affinity reagents to alter these molecular signals. These high affinity reagents can be used to block the binding of ligands to these targets and thus modulate propagation of downstream molecular signaling, leading to dampening, enhancing, or inhibiting the cellular response of interacting cells. These high affinity reagents can be combined into a composition to be used as a therapy.

Identification of Targets

[0025] A target for generating a high affinity reagent can be identified by analyzing the cell surface protein repertoires of interacting cells. For identifying a target for an interaction between a first cell and a second cell, the first cell surface protein repertoire can be characterized and the second cell surface repertoire can be characterized. The first cell surface protein repertoire can be compared to the second cell surface protein repertoire. This comparison can then identify a target that can be involved in the interaction between the first cell and the second cell. The target can be a molecule. The target can be a receptor. The target can be a ligand. The target can be on the first cell. The target can be on the second cell. The target can be expressed by or secreted by a first cell. The target can be expressed by or secreted by a second cell. The target can be used to generate a high affinity reagent. The target can be targeted by a high affinity reagent, which can alter how the target is involved in the interaction of the first cell with the second cell. The target can be targeted by a high affinity reagent, which can result in an enhanced, dampened, or inhibited response of an interaction of the first cell with the second cell in comparison to the response of an interaction of the first cell with the second cell in the absence of the high affinity reagent. A first cell and a second cell can be any cells that interact. A first cell can be the same cell type as a second cell. A first cell can be any cell type found in mammal. A first cell can be any cell type found in a human. A first cell can be an immune cell. A first cell can be a T cell or a B cell. A first cell can be an antigen presenting cell. A first cell can be a dendritic cell or a macrophage. A first cell can be a tumor cell. A second cell can be any cell type found in mammal. A second cell can be any cell type found in a human. A second cell can be an immune cell. A second cell can be a tumor cell. A second cell can be a T cell or a B cell. A second cell can be an antigen presenting cell. A second cell can be a dendritic cell or a macrophage.

[0026] The comparison of the first cell surface protein repertoire and the second cell surface protein repertoire can identify a target on the first cell, that when blocked by a high affinity reagent, can lead to an enhanced, dampened, or inhibited response after an interaction of the first cell with the second cell in comparison to the response after an interaction of the first cell with the second cell in the absence of the high affinity reagent. This comparison can also identify a group of targets on the first cell, that when blocked by a combination of high affinity reagents for that group of targets, can lead to an enhanced, dampened, or inhibited response after an interaction of the first cell with the second cell in comparison the response after an interaction of the first cell with the second cell in the absence of the combination of high affinity reagents. Conversely, this comparison can identify a target on the second cell, that when blocked by a high affinity reagent, can lead to an enhanced, dampened, or inhibited response after an interaction of the second cell with the first cell in comparison to the response after an interaction of the second cell with the first cell in the absence of the high affinity reagent. This comparison can also identify a group of targets on the second cell, that when blocked by a combination of high affinity reagents for that group of targets, can lead to an enhanced, dampened, or inhibited response after an interaction of the second cell with the first cell in comparison the response after an interaction of the second cell with the first cell in the absence of the combination of high affinity reagents. For example, for identifying a target for an interaction of a T cell and dendritic cell (DC), the T cell surface protein repertoire can be characterized and the DC cell surface repertoire can be characterized. The T cell surface protein repertoire can then be compared to the DC cell surface protein repertoire. This comparison can then identify a target on the T cell, that when blocked by a high affinity reagent, can lead to an enhanced, dampened, or inhibited response after an interaction of the T cell with the DC in comparison to the response after an interaction of the T cell with the DC in the absence of the high affinity reagent. This comparison can also identify a group of targets on the T cell, that when blocked by a combination of high affinity reagents for that group of targets, can lead to an enhanced, dampened, or inhibited response after an interaction of the T cell with the DC in comparison the response after an interaction of the T cell with the DC in the absence of the combination of high affinity reagents. Conversely, this comparison can then identify a target on the DC, that when blocked by a high affinity reagent, can lead to an enhanced, dampened, or inhibited response after an interaction of the DC with the T cell in comparison to the response after an interaction of the DC with the T cell in the absence of the high affinity reagent. This comparison can also identify a group of targets on the DC, that when blocked by a combination of high affinity reagents for that group of targets, can lead to an enhanced, dampened, or inhibited response after an interaction of the DC with the T cell in comparison the response after an interaction of the DC with the T cell in the absence of the combination of high affinity reagents.

Characterization of Cell Surface Protein Repertoire by Molecular Barcoding

[0027] A cell can be characterized by molecular barcoding and high-throughput sequencing. Molecular barcoding and high-throughput sequencing as described below can be used to identify and sequence the mRNA of a cell. Molecular barcoding and high-throughput sequencing can be used to identify and sequence mRNA of a first cell and mRNA of a second cell, wherein the first cell and the second cell can interact to produce molecular signals. Molecular barcoding and high- throughput sequencing can be used to determine the cell surface protein repertoire of a cell. The cell surface protein repertoire of a cell can be determined by examining the mRNA of the cell for membrane-tagged mRNA sequences, in which the membrane-tagged mRNA sequences can be mRNA sequences comprising a signal sequence that directs the resulting peptide or protein to the cell membrane. A key advantage of the technology of the present disclosure is the ability to rapidly and simultaneously sequence and identify all the molecules that can appear at the cell surface as membrane-bound or transmembrane moieties, thereby gaining a complete

comprehensive view of the cell surface protein repertoire of a cell. This additionally can allow for a comparison of cell surface protein repertoires between two cells, which can then be used to identify targets that are involved in the interaction of a first cell with a second cell for the generation of high affinity reagents.

Overview of stochastic labeling for molecular barcoding and counting:

[0028] The methodology underlying molecular barcoding can utilize a recursive Poisson strategy to implement single cell, molecular barcoding assays for large numbers of individual cells. For example, molecular targets from individual cells can be stochastically labeled with a cellular label (also referred to as a cellular index, barcode, or tag) and a molecular label (also referred to as a molecular index, barcode, or tag) by randomly associating individual cells with individual beads, wherein each individual bead comprises a plurality of attached stochastic labels. The stochastic labels attached to a given bead can be used to randomly label mRNA targets from an associated cell. Single cells can be randomly distributed into a plurality of microwells (e.g., a microwell array). A combinatorial library of beads, each comprising a plurality of tethered stochastic labels, can also be randomly distributed into the plurality of microwells so that a subset of the microwells contains both a single cell and a single bead. The beads can be deposited prior to depositing the cells or the beads can be deposited after depositing the cells. The stochastic labels comprising the cellular and molecular barcodes can further comprise a target recognition region that is capable of attaching to or hybridizing with molecular targets, for example, mRNA. The target molecules can be released from each cell, for example by lysing the cell, and then can be attached to or hybridized with the stochastic labels on a corresponding bead. The target molecules can also be released from the cells by cleavage, e.g., enzymatic cleavage. In some cases, such as when the target molecules are mRNA molecules, the beads can be retrieved from the microwells following hybridization of the mRNA target molecules to the stochastic labels, and pooled prior to performing reverse transcription, amplification, and sequencing reactions. [0029] The plurality of stochastic labels attached to a given bead can comprise a cellular label that can be identical for all of the stochastic labels attached to the bead, while the cellular labels for the pluralities of stochastic labels attached to different beads can be different. The plurality of stochastic labels attached to a given bead can comprise a diverse set of molecular labels selected from a set comprising a specified number of unique molecular label sequences. The plurality of stochastic labels attached to a given bead can comprise the same target recognition region. The plurality of stochastic labels attached to a given bead can also comprise two or more different target recognition regions.

[0030] The bead library can have a cellular label diversity (i.e. , a number of unique cellular label sequences) that can be at least one or two orders of magnitude higher than the number of cells to be labeled, such that the probability that each cell is paired with a unique cell barcode is very high. For example, the probability that each cell is paired with a unique cell barcode cancan be greater than 80%, greater than 90%, greater than 95%, greater than 99%, greater than 99.9%, greater than 99.99%, or greater than 99.999%.

[0031] The molecular label diversity (i.e. , the number of unique molecular label sequences) for the plurality of stochastic labels attached to a bead can be at least one or two orders of magnitude higher than the estimated number of occurrences of a target molecule species to be labeled, such that the probability that each occurrence of a target molecule (e.g. , an mRNA molecule) within a cell becomes uniquely labeled is also very high. For example, the probability that each occurrence of a target molecule is paired with a unique molecular barcode can be greater than 80%, greater than 90%, greater than 95%, greater than 99%, greater than 99.9%, greater than 99.99%, or greater than 99.999%. In these embodiments, the number of occurrences of a target molecule species in each cell can be counted (or estimated) by determining the number of unique molecular label sequences that are attached to the target molecule sequence. In many

embodiments, the determining step can be performed through sequencing of an amplified library of labeled target molecules (or their complementary sequences).

[0032] The molecular label diversity for the plurality of stochastic labels attached to a bead can be comparable to or low compared to the estimated number of occurrences of a target molecule species to be labeled, such that there can be a significant probability that the multiple occurrences of a given type of target molecule can be labeled by more than one copy of a given molecular label. In these embodiments, the number of target molecules in each cell can be calculated from the number of unique molecular label sequences attached to the target molecule sequence with the use of Poisson statistics.

[0033] In many embodiments, the target molecules of interest are mRNA molecules expressed within a single cell. Since cDNA copies of all or a portion of the polyadenylated mRNA molecules in each cell can be covalently archived on the surface of a corresponding bead, any selection of gene transcripts can be subsequently analyzed. A digital gene expression profile for each cell can be reconstructed when the barcoded transcripts are sequenced and assigned to the cell of origin (based on the cellular label identified) and counted (based on the number of unique molecular labels identified). An exemplary description of the assay methodology can be found in Fan, et ah, "Combinatorial Labeling of Single Cells for Gene Expression Cytometry", Science 347(6222):628; and Science 347(6222): 1258367. Additionally, the digital gene expression profile for each cell can be categorized to produce a digital gene expression profile of genes expressed at the cell surface. This categorization can be based on the presence of a signal sequence in the mRNA sequences that directs the resulting peptide or protein to the cell membrane. The resulting digital expression profile of genes expressed at the cell surface can therefore be representative of the cell surface repertoire of a cell.

Target molecules:

[0034] Suitable target molecules for analysis by molecular bar coding and high-throughput sequencing can include oligonucleotide molecules, DNA molecules, RNA molecules, mRNA molecules, microRNA molecules, tRNA molecules, and the like. Target molecules can be peptides, proteins, or protein fragments. The target molecules can be antibody heavy and light polypeptide chains, and/or receptor polypeptide chains {e.g., the alpha and beta chains of the T cell receptor).

Cell samples:

[0035] Suitable samples for analysis by molecular bar coding and high-throughput sequencing can include any sample comprising a plurality of cells, for example, cell cultures, blood samples, tissue samples in which the extracellular matrix has been digested or dissolved to release individual cells into suspension, and the like. The plurality of cells can be derived from a single sample, or from two or more samples that have been combined, and can comprise a plurality of cells of the same type, or a plurality of cells of mixed type.

[0036] Either cells, sub-cellular structures, or other nucleic acid containing particles can comprise suitable samples. For example, the samples can comprise cellular organelles {e.g. mitochondria, nuclei, exosomes, etc.), liposomes, cell clusters, or multicellular organisms, and the like.

[0037] The cells can be normal cells, for example, human cells in different stages of

development, or human cells from different organs or tissue types {e.g., white blood cells, red blood cells, platelets, epithelial cells, endothelial cells, neurons, glial cells, fibroblasts, skeletal muscle cells, smooth muscle cells, gametes, or cells from the heart, lungs, brain, liver, kidney, spleen, pancreas, thymus, bladder, stomach, colon, small intestine). The cells can be undifferentiated human stem cells, or human stem cells that have been induced to differentiate. The cells can be fetal human cells. The fetal human cells can be obtained from a mother pregnant with the fetus.

[0038] The cells can be rare cells. A rare cell can be, for example, a circulating tumor cell (CTC), circulating epithelial cell, circulating endothelial cell, circulating endometrial cell, circulating stem cell, stem cell, undifferentiated stem cell, cancer stem cell, bone marrow cell, progenitor cell, foam cell, mesenchymal cell, trophoblast, immune system cell (host or graft), cellular fragment, cellular organelle (e.g. , mitochondria or nuclei), pathogen infected cell, and the like.

[0039] Circulating tumor cells can be cancer cells. CTCs can be CD45-. CTCs can express cytokeratins such as 8, 18, and/or 19, or be cytokeratin negative. CTCs can be cancer stem cells and/or cells undergoing epithelial to mesenchymal transition (EMT). A CTC can be a metastatic cell.

[0040] The sample can comprise an immune cell. An immune cell can include, for example, T cell, B cell, lymphoid stem cell, myeloid progenitor cell, lymphocyte, granulocyte, B-cell progenitor, T cell progenitor, Natural Killer cell, cytotoxic T cell, helper T cell, regulatory T cell, plasma cell, memory cell, neutrophil, eosinophil, basophil, mast cell, monocyte, dendritic cell and/or macrophage, or any combination thereof.

[0041] A cell can be a T cell. A T cell can be a T cell clone, which can refer to T cells derived from a single T cell or those having identical TCRs. A T cell can be part of a T cell line which can include T cell clones and mixed populations of T cells with different TCRs all of which can recognize the same target (e.g., an antigen, a tumor, or a virus). T cells can be obtained from a number of sources, including peripheral blood mononuclear cells, bone marrow, lymph node tissue, spleen tissue, and tumors. T cells can be obtained from a unit of blood collected from a subject, such as an individual or patient, using Ficoll separation techniques. Cells from the circulating blood of an individual can be obtained by apheresis or leukapheresis. The apheresis product can comprise lymphocytes, including T cells, monocytes, granulocytes, B cells, other nucleated white blood cells, red blood cells, and platelets. The cells can be washed and resuspended in media to isolate the cell of interest.

[0042] T cells can be isolated from peripheral blood lymphocytes by lysing the red blood cells and depleting the monocytes, for example, by centrifugation through a PERCOLL™ gradient. A specific subpopulation of T cells, such as CD28+, CD4+, CDC, CD45RA+, and CD45RO+ T cells, can be further isolated by positive or negative selection techniques. For example, T cells can be isolated by incubation with anti-CD3/anti-CD28 (i.e. , 3x28)-conjugated beads, such as DYNABEADS® M-450 CD3/CD28 T, or XCYTE DYNABEADS™ for a time period sufficient for positive selection of the desired T cells. Immune cells (e.g., T cells and B cells) can be antigen specific (e.g., specific for a tumor).

[0043] The cell can be an antigen-presenting cell (APC), such as a B cell, an activated B cell from a lymph node, a lymphoblastoid cell, a resting B-cell, or a neoplastic B cell, e.g., from a lymphoma. An APC can refer to a B-cell or a follicular dendritic cell expressing at least one of the BCRC proteins on its surface.

[0044] The cells can be non- human cells, for example, other types of mammalian cells (e.g., mouse, rat, pig, dog, cow, or horse). The cells can be other types of animal or plant cells. In other embodiments, the cells can be any prokaryotic or eukaryotic cells.

[0045] The cells can be sorted prior to associating a cell with a bead. For example the cells can be sorted by fluorescence-activated cell sorting or magnetic-activated cell sorting, or more generally by flow cytometry. The cells can be filtered by size. In some instances a retentate contains the cells to be associated with the beads. In some instances the flow through contains the cells to be associated with the beads.

[0046] When loading cells, the concentration of the cell suspension (i.e., the number of cells per mL) can be adjusted so that the probability of having more than one cell settle into a given microwell can be very small. Typically, the concentration of the cell suspension can be adjusted so that the volume of cell suspension used to load, e.g., a microwell array, contains

approximately one-tenth the number of cells as the number of wells in the microwell array. The probability that more than one cell settles into a given microwell can be governed by Poisson statistics.

Stochastic label chemical structure:

[0047] In some examples of the disclosed methods, the stochastic labels (also referred to as barcodes, tags, or indexes) used for single cell molecular barcoding studies comprise

oligonucleotides, for example, oligodeoxyribonucleotides (DNA), oligoribonucleotides (RNA), peptide nucleic acid (PNA) polymers, 2'-0-methyl-substituted RNA, locked nucleic acid (LNA) polymers, bridged nucleic acid (BNA) polymers, and the like.

Stochastic labels can be attached to solid supports:

[0048] In many embodiments, the stochastic labels used in the disclosed methods can be tethered to beads, for example to synthesis resin beads, or other solid supports as can be described in more detail below. As used herein, the phrases "tethered", "attached", and "immobilized" are used interchangeably, and can refer to covalent or non-covalent means for attaching stochastic labels to solid supports such as beads. One non-limiting example of stochastic label structure is as follows. The stochastic labels comprise a plurality of 5'-amine modified oligonucleotides attached to a bead. The oligonucleotides comprise a 5' amine group, a universal primer, a cellular label, a molecular label, and a target binding region. The oligonucleotides can optionally further comprise one or more additional labels, e.g., a sample label for use in labeling all cells from a given sample when two or more samples are processed simultaneously. The stochastic label oligonucleotide sequences can be attached to a solid support at their 5' end. The stochastic label oligonucleotide sequences can be attached to a solid support at their 3' end.

Stochastic labels can comprise one or more universal labels:

[0049] The stochastic labels used in the disclosed methods, compositions, devices, kits, and systems can comprise one or more universal labels. The one or more universal labels can be the same for all oligonucleotides in the set of oligonucleotides attached to a given bead. The one or more universal labels can be the same for all oligonucleotides attached to a plurality of beads. A universal label can comprise a nucleic acid sequence that is capable of hybridizing to a sequencing primer. Sequencing primers can be used for sequencing oligonucleotides comprising a universal label. Sequencing primers {e.g., universal sequencing primers) can comprise sequencing primers associated with high- throughput sequencing platforms. A universal label can comprise a nucleic acid sequence that is capable of hybridizing to a PCR primer. The universal label can comprise a nucleic acid sequence that is capable of hybridizing to a sequencing primer and a PCR primer. The nucleic acid sequence of the universal label that is capable of hybridizing to a sequencing or PCR primer can be referred to as a primer binding site. A universal label can comprise a sequence that can be used to initiate transcription of the oligonucleotide. A universal label can comprise a sequence that can be used for extension of the oligonucleotide or a region within the oligonucleotide. A universal label can be at least about 1, 2, 3, 4, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50 or more nucleotides in length. A universal label can comprise at least about 10 nucleotides. A universal label can be at most about 1, 2, 3, 4, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50 or more nucleotides in length. A cleavable linker or modified nucleotide can be part of the universal label sequence to enable stochastic label oligonucleotides to be cleaved off from the solid support.

Stochastic labels can comprise a cellular label:

[0050] Stochastic labels can comprise a cellular label, e.g. a nucleic acid sequence that provides information for determining which target nucleic acid originated from which cell. In many embodiments, the cellular label is identical for all oligonucleotides attached to a given bead or solid support, but different for different beads or solid supports. At least 60%, 70%, 80%, 85%,

90%, 95%, 97%, 99% or 100% of the oligonucleotides on the same solid support can comprise the same cellular label. At least 60% of the oligonucleotides on the same solid support can comprise the same cellular label. In some embodiment, at least 95% of the oligonucleotides on the same solid support can comprise the same cellular label. There can be as many as 10 or more unique cellular label sequences represented in a plurality of beads or solid supports. There can be as many as 10⁴ or more unique cellular label sequences represented in a plurality of beads or solid supports. There can be as many as 10⁵ or more unique cellular label sequences represented in a plurality of beads or solid supports. There can be as many as 10⁶ or more unique cellular label sequences represented in a plurality of beads or solid supports. A cellular label can be at least about 1, 2, 3, 4, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50 or more nucleotides in length. A cellular label can be at most about 300, 200, 100, 90, 80, 70, 60, 50, 40, 30, 20, 15, 12, 10, 9, 8, 7, 6, 5, 4 or fewer nucleotides in length. A cellular label can comprise between about 5 to about 200 nucleotides. A cellular label can comprise between about 10 to about 150 nucleotides. A cellular label can comprise between about 20 to about 125 nucleotides in length.

Cellular labels can comprise error correction codes:

[0051] The cellular label can further comprise a unique set of nucleic acid sub-sequences of defined length, e.g., 7 nucleotides each (equivalent to the number of bits used in some Hamming error correction codes), which are designed to provide error correction capability. The set of error correction sub-sequences comprise 7 nucleotide sequences designed such that any pairwise combination of sequences in the set exhibits a defined "genetic distance" (or number of mismatched bases), for example, a set of error correction sub- sequences can be designed to exhibit a genetic distance of 3 nucleotides. In this case, review of the error correction sequences in the set of sequence data for labeled target nucleic acid molecules (described more fully below) can allow one to detect or correct amplification or sequencing errors. The length of the nucleic acid sub-sequences used for creating error correction codes can vary, for example, they can be 3 nucleotides, 7 nucleotides, 15 nucleotides, or 31 nucleotides in length. Nucleic acid subsequences of other lengths can be used for creating error correction codes.

Cellular labels can comprise two or more subunits:

[0052] The cellular label can comprise an assembly of two or more subunits (also referred to as subparts or components) that are assembled (or synthesized) in a combinatorial split-pool fashion to create a large number of unique cellular label sequences in a minimal number of assembly or synthesis steps. For example, three rounds of split-pool synthesis performed using a set of 6 unique cellular label subunits will yield 6 = 216 unique cellular label sequences, where each cellular label comprises an assembly of three subunits. More generally, a cellular label sequence comprising M subunits that have been assembled in a combinatorial split-pool fashion using a set of N unique subunits at each step will yield N^M unique combinations. Where the stochastic labels are oligonucleotides, the cellular subunit sequences can be assembled through the use of polymerase extension or ligation reactions. One or more linker sequences can be used to facilitate the assembly of the cellular label sequence subunits.

Stochastic labels can comprise a molecular label:

[0053] A molecular label can comprise a nucleic acid sequence that provides information for identifying the specific type of target nucleic acid species hybridized to the oligonucleotide. A molecular label can comprise a nucleic acid sequence that provides a counter for the specific occurrence of the target nucleic acid species hybridized to the oligonucleotide. In many embodiments, a diverse set of molecular labels are attached to a given bead. There can be as many as 10⁶ or more unique molecular label sequences attached to a given bead. There can be as many as 10⁵ or more unique molecular label sequences attached to a given bead. There can be as many as 10⁴ or more unique molecular label sequences attached to a given bead. There can be as many as 10 or more unique molecular label sequences attached to a given bead. There can be as many as 10 or more unique molecular label sequences attached to a given bead. A molecular label can be at least about 1, 2, 3, 4, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50 or more nucleotides in length. A molecular label can be at most about 300, 200, 100, 90, 80, 70, 60, 50, 40, 30, 20, 15, 12, 10, 9, 8, 7, 6, 5, 4 or fewer nucleotides in length.

Stochastic labels can comprise a target binding region:

[0054] The target binding regions can comprise a nucleic acid sequence that hybridizes specifically to a target nucleic acid {e.g., a cellular nucleic acid to be analyzed), for example to a specific gene sequence. A target binding region can comprise a nucleic acid sequence that can attach {e.g., hybridize) to a specific location of a specific target nucleic acid. The target binding region can comprise a nucleic acid sequence that is capable of specific hybridization to a restriction site overhang {e.g., an EcoRI sticky-end overhang). The stochastic label can then ligate to any nucleic acid molecule comprising a sequence complementary to the restriction site overhang. A target binding region can comprise a non-specific target nucleic acid sequence. A non-specific target nucleic acid sequence can refer to a sequence that can bind to multiple target nucleic acids, independent of the specific sequence of the target nucleic acid. For example, target binding region can comprise a random multimer sequence, or an oligo-dT sequence that hybridizes to the poly-A tail on mRNA molecules. A random multimer sequence can be, for example, a random dimer, trimer, quatramer, pentamer, hexamer, septamer, octamer, nonamer, decamer, or higher multimer sequence of any length. A target binding region can be at least about

5, 10, 15, 20, 25, 30, 35, 40, 45, 50 or more nucleotides in length. A target binding region can be at most about 5, 10, 15, 20, 25, 30, 35, 40, 45, 50 or more nucleotides in length. A target binding region can comprise any number of nucleotides within this range, for example, about a target binding region can be about 18 nucleotides in length.

[0055] The target binding region is the same for all oligonucleotides attached to a given bead. The target binding regions for the plurality of oligonucleotides attached to a given bead can comprise two or more different target binding sequences. For example, the target binding regions for the plurality of oligonucleotides attached to a given bead can comprise a mixture of oligo-dT sequences and copies of a single target specific sequence. The target binding regions for the plurality of oligonucleotides attached to a given bead can comprise a mixture of an oligo-dT sequence and copies of two different target specific sequences. The target binding regions for the plurality of oligonucleotides attached to a given bead can comprise a mixture of an oligo-dT sequence and copies of three different target specific sequences. In general, the target binding regions for the plurality of oligonucleotides attached to a given bead can comprise a mixture of between one and one hundred, or more, different target binding sequences, including, but not limited to, target specific sequences, random multimer sequences, sequences capable of specific hybridization to a restriction site overhang, or oligo-dT sequences, in any combination of sequences and in any combination of relative proportions.

Stochastic labels tethered to beads:

[0056] The stochastic labels disclosed herein can be attached to solid supports such as beads. As used herein, the terms "tethered", "attached", and "immobilized" are used interchangeably, and can refer to covalent or non-covalent means for attaching stochastic labels to solid supports such as beads. The stochastic labels can be immobilized within a small reaction volume, e.g., attached to a surface in a well or microwell, or to a different form of solid support rather than attached to a bead.

[0057] Pre- synthesized stochastic labels can be attached to beads or other solid supports through any of a variety of immobilization techniques involving functional group pairs on the solid support and the oligonucleotide. The oligonucleotide functional group and the solid support functional group are individually selected from the group consisting of biotin, streptavidin, primary amine(s), carboxyl(s), hydroxyl(s), aldehyde(s), ketone(s), and any combination thereof. A stochastic label oligonucleotide can be tethered to a solid support, for example, by coupling {e.g., using l-Ethyl-3-(3-dimethylaminopropyl) carbodiimide) a 5' amino group on the oligonucleotide to the carboxyl group of the functionalized solid support. Residual non-coupled oligonucleotides can be removed from the reaction mixture by performing multiple rinse steps. The stochastic label oligonucleotides and solid support are attached indirectly via linker molecules (e.g. , short, functionalized hydrocarbon molecules or polyethylene oxide molecules) using similar attachment chemistries. The linkers can be cleavable linkers, e.g., acid-labile linkers or photo-cleavable linkers.

[0058] In some embodiment, stochastic labels are synthesized on solid supports such as synthesis resin beads using any of a number of solid-phase oligonucleotide synthesis techniques known to those of skill in the art. Single nucleotides can be coupled in step-wise fashion to the growing, tethered oligonucleotide. A short, pre-synthesized sequence (or block) of several oligonucleotides can be coupled to the growing, tethered oligonucleotide. Oligonucleotides can be synthesized by interspersing step-wise or block coupling reactions with one or more rounds of split-pool synthesis, in which the total pool of synthesis beads is divided into a number of individual smaller pools which are then each subjected to a different coupling reaction, followed by recombination and mixing of the individual pools to randomize the growing oligonucleotide sequence across the total pool of beads. Split-pool synthesis is an example of a combinatorial synthesis process in which a maximum number of chemical compounds are synthesized using a minimum number of chemical coupling steps. The potential diversity of the compound library thus created is determined by the number of unique building blocks (e.g. , nucleotides) available for each coupling step, and the number of coupling steps used to create the library. For example, a split-pool synthesis comprising 10 rounds of coupling using 4 different nucleotides at each step will yield 4¹⁰ = 1,048,576 unique nucleotide sequences. Split-pool synthesis can be performed using enzymatic methods such as polymerase extension or ligation reactions rather than chemical coupling. For example, in each round of a split-pool polymerase extension reaction, the 3' ends of the stochastic label oligonucleotides tethered to beads in a given pool can be hybridized with the 5'ends of a set of semi-random primers, e.g. , primers having a structure of 5'-(M)_k-(X)i-(N)_j- 3', where (X)j is a random sequence of nucleotides that is i nucleotides long (the set of primers comprising all possible combinations of (X) , (N)_j is a specific nucleotide (or series of j nucleotides), and (M)_k is a specific nucleotide (or series of k nucleotides), wherein a different deoxyribonucleotide triphosphate (dNTP) is added to each pool and incorporated into the tethered oligonucleotides by the polymerase.

[0059] The number of oligonucleotides conjugated to or synthesized on a solid support such as a bead can comprise 100 or more oligonucleotide molecules. The solid support can comprise 1,000 or more oligonucleotide molecules. The solid support can comprise 10,000 or more

oligonucleotide molecules. The solid support can comprise 100,000 or more oligonucleotides. The solid support can comprise 1,000,000 or more oligonucleotides.

[0060] The number of oligonucleotides conjugated to or synthesized on a solid support such as a bead can be 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10-fold more than the number of target nucleic acids in a cell. The number of oligonucleotides conjugated to or synthesized on a solid support such as a bead can be 100-fold more than the number of target nucleic acids in a cell. The number of oligonucleotides conjugated to or synthesized on a solid support such as a bead can be 1,000-fold more than the number of target nucleic acids in a cell. In some instances, at least 10, 20, 30, 40, 50, 60, 70, 80, 90 or 100% of the oligonucleotides are bound by a target nucleic acid. In some instances, at most 10, 20, 30, 40, 50, 60, 70, 80, 90 or 100% of the oligonucleotides are bound by a target nucleic acid. In some instances, at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50, 60, 70, 80, 90 or 100 or more different target nucleic acids are captured by the oligonucleotides on a solid support. In some instances, at most 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50, 60, 70, 80, 90 or 100 or more different target nucleic acids are captured by the oligonucleotides on a solid support.

[0061] In one example of the disclosed methods, the plurality of stochastic labels tethered to a given set of beads can be designed to focus the downstream analysis of sequence data on selected subpopulations of cells. The plurality of stochastic labels tethered to a set of beads can be designed to exclude selected subpopulations of cells from the downstream analysis of sequence data. An example of a suitable approach for implementing these embodiments would be the inclusion of a subset of tethered labels attached to each bead that comprise one or more target- specific binding regions, where the one or more nucleic acid targets (e.g. , nucleic acid markers, genetic markers) are chosen to define the subset of cells to be included in or excluded from further analysis at the sequence data analysis stage. The tethered label molecules of the subset would each comprise the cellular label (the same sequence for all labels attached to a given bead), a molecular label (e.g. , a single, unique sequence selected at random from a diverse set of molecular label or barcode sequences included in the subset of tethered label molecules attached to a given bead), and a target- specific binding region, as well as one or more additional primer sequences, sample label sequences, etc. , as described above. A set of nucleic acid targets (nucleic acid markers or genetic markers), e.g. , mRNA targets, can be chosen to identify, for example, cells undergoing apoptosis (e.g. , by monitoring Bax, Bcl-2, caspase-3, and caspase-7 expression, or expression of other genes potentially involved in apoptosis, or combinations thereof), rapid proliferation (e.g. , by monitoring CKS 1B, CCNB2, CDC2, DLG7, BUB3, MAD2L1, DLG7, PLK4, KIF2C, MKI67, BRRN1, NUSAP1, ASPM, or KLF7 expression, or expression of other genes potentially involved in cell proliferation, or combinations thereof), or any other subpopulation of cells that can be defined on the basis of nucleic acid markers. Analysis of sequence data generated by performing the stochastic labeling or molecular barcoding assay then provides a list of the cellular barcodes associated with the specified subpopulation of cells so that further analysis can be focused on the selected subpopulation of cells, or the selected subpopulation of cells can be excluded from further analysis.

Beads & other encoded solid supports:

[0062] Any of a variety of different beads can be used as solid supports for attaching pre- synthesized stochastic label oligonucleotides or for in situ solid-phase synthesis of stochastic label oligonucleotides. A bead can encompass any type of solid, porous, or hollow sphere, ball, bearing, cylinder, or other similar configuration composed of plastic, ceramic, metal, or polymeric material onto which a nucleic acid can be immobilized {e.g., covalently or non- covalently). A bead can comprise a discrete particle that can be spherical {e.g., microspheres) or have a non-spherical or irregular shape, such as cubic, cuboid, pyramidal, cylindrical, conical, oblong, or disc-shaped, and the like. Beads can comprise a variety of materials including, but not limited to, paramagnetic materials {e.g., magnesium, molybdenum, lithium, and tantalum), superparamagnetic materials {e.g., ferrite (Fe₃0₄; magnetite) nanoparticles), ferromagnetic materials {e.g., iron, nickel, cobalt, some alloys thereof, and some rare earth metal compounds), ceramic, plastic, glass, polystyrene, silica, methylstyrene, acrylic polymers, titanium, latex, sepharose, agarose, hydrogel, polymer, cellulose, nylon, and any combination thereof. A bead can refer to any three dimensional structure that can provide an increased surface area for immobilization of biological particles and macromolecules, such as DNA and RNA.

[0063] In general, the diameter of the beads can range from about 1 μιη to about 100 μιη, or larger. The diameter of the beads can be at least 1 μιη, at least 5μιη, at least ΙΟμιη, at least 20μιη, at least 25μιη, at least 30μιη, at least 35μιη, at least 40μιη, at least 45μιη, at least 50μιη, at least

60 μιη, at least 70 μιη, at least 80 μιη, at least 90 μιη, or at least 100 μιη. In some embodiment, the diameter of the beads can be at most 100 μιη, at most 90 μιη, at most 80 μιη, at most 70 μιη, at most 60 μιη, at most 50 μιη, at most 45 μιη, at most 40 μιη, at most 35 μιη, at most 30 μιη, at most 25 μιη, at most 20 μιη, at most 15 μιη, at most 10 μιη, at most 5 μιη, or at most 1 μιη. The diameter of the beads can have any value within this range, for example, beads can have a diameter in the range of about 20 to 50 μιη. Beads can have a diameter of about 33 μιη. It can be desirable to use the smallest beads possible {i.e., that are compatible with the synthesis process used to create the plurality of stochastic labels attached to the beads and with other assay and device requirements), as larger beads will tend to settle faster than smaller beads of the same density, and therefore can result in less uniform distributions of beads across the microwell array.

For spherical beads, settling velocity can be calculated using Stokes' Law:

₌ 2Ga² (p₁ - p₂)

9η where V = settling velocity (cm/sec), G = the acceleration due to gravity (cm/sec ), a = bead radius (cm), pi = density of the bead (g/cm 3 ), p₂ = density of suspending media (g/cm 3 ), and η = coefficient of viscosity (poise; g/cm-sec). Thus, a 50 μιη diameter bead can settle 25-times faster than a 10 μιη diameter bead of the same density.

[0064] A bead can be attached to, positioned within, or embedded into one or more supports. For example, a bead can be attached to a gel or hydrogel. A bead can be attached to a matrix. A bead can be embedded into a matrix. A bead can be attached to a polymer. A bead can be embedded into a polymer. The spatial position of a bead within the support (e.g., gel, matrix, scaffold, or polymer) can be identified using the oligonucleotide present on the bead which serves as a location address.

[0065] Examples of beads include, but are not limited to, streptavidin beads, agarose beads, magnetic beads, Dynabeads®, MACS® microbeads, antibody conjugated beads (e.g., antiimmunoglobulin microbead), protein A conjugated beads, protein G conjugated beads, protein A/G conjugated beads, protein L conjugated beads, oligo-dT conjugated beads, silica beads, silica-like beads, anti-biotin microbead, anti-fluorochrome microbead, and BcMag™ Carboxy- Terminated Magnetic Beads.

[0066] A bead can be associated with (e.g., impregnated with) quantum dots or fluorescent dyes to make it fluorescent in one fluorescence optical channel or multiple optical channels. A bead can be associated with iron oxide or chromium oxide to make it paramagnetic or ferromagnetic. Also, beads can be non-spherical in shape. A flatter, disc-like bead can be used in some embodiments. The disc-like bead can substantially occlude the well volume in some orientations but permit cells to move freely into the well in other orientations. This large disc-like bead is beneficial for achieving two different functions, the first being confinement of the cell and materials from within the cell after lysis, and the second being the ability for a cell to be readily loaded during the assay. The orientation can also be controlled during operation, including automatically by the instrument, using an applied magnetic field. Other bead shapes, used in combination with other well shapes (which can be different from simple cylinders), can be used to improve the efficiency and speed of cartridge, of bead and cell loading, and of performing the assay.

Dual-encoded beads for stochastic labeling and molecular barcoding:

[0067] In one example of the stochastic labeling and molecular barcoding methods described above, it can be desirable that beads associated with individual cells exhibiting a predefined set of properties, or beads associated with more than one cell, be removed from further processing, or that the sequence data arising from said beads be used to focus the downstream analysis of sequence data. The sequence data arising from said beads can be excluded from any further sequence data analysis. Further analysis of sequence data can include only that sequence data arising from said beads. This can be achieved through the use of optically-encoded beads in a dual encoding scheme, e.g. , where individual beads are uniquely identified both by an optical code (e.g., by impregnating the beads with a spectrally-distinct set of fluorophores, quantum dots, Raman tags, up-converting phosphors, and the like; or by synthesis of an attached optical code through the use of solid-phase split-pool synthesis methodologies and a set of spectrally-distinct fluorescent building blocks) as well as a nucleic acid sequence (e.g., the cellular label) that is incorporated into the plurality of tethered stochastic labels attached to a given bead. Beads co- localized with cells exhibiting a set of predefined properties, or with more than one cell, would each be identified based on their optical code, and the sequence data arising from said beads would be subsequently identified by the corresponding cellular label sequence, thereby generating a list of sequence data to be included or excluded from further analysis. Individual cells or sub-populations of cells that exhibit a predefined set of characteristics, e.g. , that express a particular cell surface receptor (marker) or set of cell surface receptors, can be identified through any of a variety of suitable techniques, e.g., through immunohistochemical (IHC) staining of individual cells in a microwell array format using fluorescently-labeled antibodies directed towards the cell surface markers and fluorescence imaging techniques, or through the use of flow-cytometry and fluorescence-activated cell-sorting methods. If the location of each cell label sequence in a two-dimensional space can be identified and recorded, the assay can provide spatial information for single cell gene expression, and can be particularly useful for analyzing gene expression in, for example, the thin tissue sections routinely collected for pathological studies.

Dual encoding using array address codes:

[0068] Dual encoding schemes can be implemented by use of pre-deposited array address codes (e.g., nucleic acid barcodes that code for the location of a specific well in the array) instead of optically-encoded beads to implement dual encoding schemes. Array address codes can be deposited in wells using ink-jet printing techniques, microarray spotting techniques, dip-pen nanolithography techniques, and the like. The array address codes can be no n- specifically adsorbed to one or more inner surfaces of the microwells. The array address codes can be covalently attached to one or more inner surfaces of the microwells. The array address codes can be synthesized in situ by means of solid phase synthesis techniques, wherein one or more inner surfaces of the microwells are used as a solid support. In embodiments where the array address codes are covalently attached to one or more inner surfaces of the microwells, the attachment can comprise the use of cleavable linkers, e.g., acid-labile, base-labile, or photocleavable linkers, so that the array address codes can be released when desired and allowed to hybridize with a subset of the tethered stochastic labels attached to a bead. The array address codes can be used in combination with the plurality of stochastic labels attached to a bead that comprises a cellular label. The array address codes can be used instead of a plurality of stochastic labels attached to a bead, and can themselves comprise a cellular label, a molecular label, and one or more primer or adapter sequences. The array address codes can be used in similar fashion to that described above for optically-encoded beads in identifying subsets of cells to be included or excluded from downstream sequence data analysis.

Cell lysis:

[0069] Following the random distribution of cells and bead-based stochastic labels, such that an individual cell and individual bead are confined together within a small reaction volume, e.g., a well or microwell, the cells can be lysed to liberate the target molecules. Cell lysis can be accomplished by any of a variety of means, for example, by chemical or biochemical means, by osmotic shock, or by means of thermal lysis, mechanical lysis, or optical lysis. For example, cells can be lysed by addition of a cell lysis buffer comprising a detergent (e.g., SDS, Li dodecyl sulfate, Triton X-100, Tween-20, or NP-40), an organic solvent (e.g., methanol or acetone), or digestive enzymes (e.g., proteinase K, pepsin, or trypsin), or any combination thereof.

Attachment of stochastic labels to target nucleic acid molecules:

[0070] Following lysis of the cells and release of nucleic acid molecules therefrom, the nucleic acid molecules can randomly attach to the stochastic label oligonucleotides of the co-localized bead. Attachment can comprise hybridization of a label's target recognition region to a complementary portion of the target nucleic acid molecule. The assay conditions used for hybridization (e.g., buffer pH, ionic strength, temperature, etc.) are chosen to promote formation of specific, stable hybrids, as is well known to those of skill in the art.

[0071] Attachment can further comprise ligation of a label's target recognition region and a portion of the target nucleic acid molecule. For example, the target binding region can comprise a nucleic acid sequence that is capable of specific hybridization to a restriction site overhang (e.g., an EcoRI sticky-end overhang). The assay procedure further comprises treating the target nucleic acids with a restriction enzyme (e.g., EcoRI) to create a restriction site overhang. The stochastic label can then be ligated to any nucleic acid molecule comprising a sequence complementary to the restriction site overhang. A ligase (e.g., T4 DNA ligase) can be used to join the two oligonucleotide fragments. [0072] In one example of the disclosed methods, the labeled target nucleic acid molecules from a plurality of cells (or a plurality of samples) are subsequently pooled, for example by retrieving beads to which the stochastically-labeled nucleic acid molecules are attached. The distribution and/or retrieval of bead-based collections of attached nucleic acid molecules can be implemented by use of magnetic beads and an externally-applied magnetic field. Once the stochastically- labeled nucleic acid molecules have been pooled, all further processing can proceed in a single reaction vessel. Further processing (e.g., reverse transcription reactions (or other nucleic acid extension reactions) and amplification reactions) can be performed within the microwells, that is, without first pooling the labeled target nucleic acid molecules from a plurality of cells.

Reverse transcription:

[0073] In one example of the disclosed methods, a reverse transcription reaction is performed to create a stochastic label - target nucleic acid conjugate (e.g., a covalently-linked molecular complex or molecular conjugate) comprising the stochastic label and a complementary sequence of all or a portion of the target nucleic acid (i.e., a labeled cDNA molecule). Reverse

transcription can be performed using any of a variety of techniques known to those of skill in the art. Reverse transcription of the labeled-RNA molecule can occur by the addition of a reverse transcription primer along with the reverse transcriptase. The reverse transcription primer is an oligo-dT primer, a random hexanucleotide primer, or a target- specific oligonucleotide primer. Generally, oligo-dT primers are 12-18 nucleotides in length and bind to the endogenous poly- A tail at the 3' end of mammalian mRNA. Random hexa-nucleotide primers can bind to mRNA at a variety of complementary sites. Target- specific oligonucleotide primers typically selectively prime the mRNA of interest.

Amplification:

[0074] In one example of the disclosed methods, one or more nucleic acid amplification reactions can be performed to create multiple copies of the labeled target nucleic acid molecules.

Amplification can be performed in a multiplexed manner, wherein multiple target nucleic acid sequences are amplified simultaneously. The amplification reaction can be used to add sequencing adaptors to the nucleic acid molecules. The amplification reactions can comprise amplifying at least a portion of a sample label, if present. The amplification reactions can comprise amplifying at least a portion of the cellular and or molecular label. The amplification reactions can comprise amplifying at least a portion of a sample tag, a cellular label, a molecular label, a target nucleic acid, or a combination thereof. The amplification reactions can comprise amplifying at least 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 15%, 20%, 25%, 30%, 35%,

40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 97%, or 100% of the plurality of nucleic acids. The method can further comprise conducting one or more cDNA synthesis reactions to produce one or more cDNA copies of sample label-tagged nucleic acids, cellular label-tagged nucleic acids, or molecular label-tagged nucleic acids.

[0075] Amplification can be performed using a polymerase chain reaction (PCR). As used herein, PCR can refer to a reaction for the in vitro amplification of specific DNA sequences by the simultaneous primer extension of complementary strands of DNA. As used herein, PCR can encompass derivative forms of the reaction, including but not limited to, RT-PCR, real-time PCR, nested PCR, quantitative PCR, multiplexed PCR, digital PCR, and assembly PCR.

[0076] Amplification of the labeled nucleic acids comprises non-PCR based methods. Examples of non-PCR based methods include, but are not limited to, multiple displacement amplification (MDA), transcription-mediated amplification (TMA), nucleic acid sequence-based amplification (NASBA), strand displacement amplification (SDA), real-time SDA, rolling circle amplification, or circle-to-circle amplification. Other non-PCR-based amplification methods include multiple cycles of DNA-dependent RNA polymerase-driven RNA transcription amplification or RNA- directed DNA synthesis and transcription to amplify DNA or RNA targets, a ligase chain reaction (LCR), a QP replicase (Q 3) method, use of palindromic probes, strand displacement amplification, oligonucleotide-driven amplification using a restriction endonuclease, an amplification method in which a primer is hybridized to a nucleic acid sequence and the resulting duplex is cleaved prior to the extension reaction and amplification, strand displacement amplification using a nucleic acid polymerase lacking 5' exonuclease activity, rolling circle amplification, and ramification extension amplification (RAM).

[0077] In some instances, the methods disclosed herein further comprise conducting a polymerase chain reaction on the labeled nucleic acid (e.g., labeled-RNA, labeled-DNA, labeled- cDNA) to produce a labeled-amplicon. The labeled-amplicon can be double- stranded molecule. The double- stranded molecule can comprise a double- stranded RNA molecule, a double- stranded DNA molecule, or a RNA molecule hybridized to a DNA molecule. One or both of the strands of the double- stranded molecule can comprise a sample label, a cellular label, or a molecular label. Alternatively, the labeled-amplicon is a single- stranded molecule. The single- stranded molecule can comprise DNA, RNA, or a combination thereof. The nucleic acids of the present invention can comprise synthetic or altered nucleic acids.

[0078] In some embodiment, amplification can comprise use of one or more non-natural nucleotides. Non-natural nucleotides can comprise photolabile or triggerable nucleotides.

Examples of non-natural nucleotides include, but are not limited to, peptide nucleic acid (PNA), morpholino and locked nucleic acid (LNA), as well as glycol nucleic acid (GNA) and threose nucleic acid (TNA). Non- natural nucleotides can be added to one or more cycles of an amplification reaction. The addition of the non-natural nucleotides can be used to identify products as specific cycles or time points in the amplification reaction.

[0079] Conducting the one or more amplification reactions can comprise the use of one or more primers. The one or more primers can comprise one or more oligonucleotides. The one or more oligonucleotides can comprise at least about 7-9 nucleotides. The one or more oligonucleotides can comprise less than 12-15 nucleotides. The one or more primers can anneal to at least a portion of the plurality of labeled nucleic acids. The one or more primers can anneal to the 3' end or 5' end of the plurality of labeled nucleic acids. The one or more primers can anneal to an internal region of the plurality of labeled nucleic acids. The internal region can be at least about 50, 100, 150, 200, 220, 230, 240, 250, 260, 270, 280, 290, 300, 310, 320, 330, 340, 350, 360, 370, 380, 390, 400, 410, 420, 430, 440, 450, 460, 470, 480, 490, 500, 510, 520, 530, 540, 550, 560, 570, 580, 590, 600, 650, 700, 750, 800, 850, 900 or 1,000 nucleotides from the 3' ends the plurality of labeled nucleic acids. The one or more primers can comprise a fixed panel of primers. The one or more primers can comprise at least one or more custom primers. The one or more primers can comprise at least one or more control primers. The one or more primers can comprise at least one or more housekeeping gene primers. The one or more primers can comprise a universal primer. The universal primer can anneal to a universal primer binding site. The one or more custom primers can anneal to a first sample label, a second sample label, a cellular label, a molecular label, a target nucleic acid, or a combination thereof. The one or more primers can comprise a universal primer and a custom primer. The custom primer can be designed to amplify one or more target nucleic acids. The target nucleic acids can comprise a subset of the total nucleic acids in one or more samples. The target nucleic acids can comprise a subset of the total labeled nucleic acids in one or more samples. The one or more primers can comprise at least 96 or more custom primers. The one or more primers can comprise at least 960 or more custom primers. The one or more primers can comprise at least 9,600 or more custom primers. The one or more custom primers can anneal to two or more different labeled nucleic acids. The two or more different labeled nucleic acids can correspond to one or more genes.

[0080] An example amplification scheme for use in methods of the present disclosure can be carried out as follows. The first PCR reaction amplifies molecules attached to the bead using a gene specific primer and a primer against the universal Illumina sequencing primer 1 sequence. The second PCR reaction amplifies the first PCR products using a nested gene specific primer flanked by Illumina sequencing primer 2 sequence, and a primer against the universal Illumina sequencing primer 1 sequence. The third PCR reaction adds P5 and P7 and sample index to turn PCR products into an Illumina sequencing library. Sequencing using 150bp x 2 sequencing reveals the cell label and molecular index on read 1, the gene on read 2, and the sample index on index 1 read.

[0081] The downstream analysis of sequence data can be focused on selected subpopulations of cells by performing an amplification reaction using one or more target- specific primers, wherein the one or more target- specific primers are capable of specific hybridization with, for example, one or more genes or gene products that define a subpopulation of cells. A set of nucleic acid targets (nucleic acid markers or genetic markers), e.g. , mRNA targets, can be chosen to identify, for example, cells undergoing apoptosis (e.g. , by monitoring Bax, Bcl-2, caspase-3, and caspase- 7 expression, or expression of other genes potentially involved in apoptosis, or combinations thereof), rapid proliferation (e.g. , by monitoring CKS 1B, CCNB2, CDC2, DLG7, BUB3, MAD2L1, DLG7, PLK4, KIF2C, MKI67, BRRN1 , NUSAP1, ASPM, or KLF7 expression, or expression of other genes potentially involved in cell proliferation, or combinations thereof), or any other subpopulation of cells that can be defined on the basis of nucleic acid markers. A multiplexed amplification reaction performed using the one or more target- specific primers can be used to create multiple copies of the labeled target nucleic acid molecules attached to beads, which can then be sequenced to generate a list of cells comprising the one of more specified target nucleic acid molecules.

Sequencing:

[0082] In some aspects, determining the number of different labeled nucleic acids can comprise determining the sequence of the labeled nucleic acid or any product thereof (e.g. labeled- amp licons, labeled-cDNA molecules). In some instances, an amplified target nucleic acid can be subjected to sequencing. Determining the sequence of the labeled nucleic acid or any product thereof can comprise conducting a sequencing reaction to determine the sequence of at least a portion of a sample label, a cellular label, a molecular label, at least a portion of the labeled target nucleic acid, a complement thereof, a reverse complement thereof, or any combination thereof.

[0083] Determination of the sequence of a nucleic acid (e.g. , amplified nucleic acid, labeled nucleic acid, cDNA copy of a labeled nucleic acid, etc.) can be performed using variety of sequencing methods including, but not limited to, sequencing by hybridization (SBH), sequencing by ligation (SBL), quantitative incremental fluorescent nucleotide addition sequencing (QIFNAS), stepwise ligation and cleavage, fluorescence resonance energy transfer (FRET), molecular beacons, TaqMan reporter probe digestion, pyrosequencing, fluorescent in situ sequencing (FISSEQ), FISSEQ beads, wobble sequencing, multiplex sequencing, polymerized colony (POLONY) sequencing; nanogrid rolling circle sequencing (ROLONY), allele- specific oligo ligation assays (e.g. , oligo ligation assay (OLA), single template molecule OLA using a ligated linear probe and a rolling circle amplification (RCA) readout, ligated padlock probes, or single template molecule OLA using a ligated circular padlock probe and a rolling circle amplification (RCA) readout), and the like.

[0084] In some instances, determining the sequence of the labeled nucleic acid or any product thereof comprises paired-end sequencing, nanopore sequencing, high-throughput sequencing, shotgun sequencing, dye-terminator sequencing, multiple-primer DNA sequencing, primer walking, Sanger dideoxy sequencing, Maxim-Gilbert sequencing, pyrosequencing, true single molecule sequencing, or any combination thereof. Alternatively, the sequence of the labeled nucleic acid or any product thereof can be determined by electron microscopy or a chemical- sensitive field effect transistor (chemFET) array.

[0085] High-throughput sequencing methods, such as cyclic array sequencing using platforms such as Roche 454, Illumina Solexa, ABI-SOLiD, ION Torrent, Complete Genomics, Pacific Bioscience, Helicos, or the Polonator platform, can also be utilized. In some embodiment, sequencing can comprise MiSeq sequencing. In some embodiment, sequencing can comprise HiSeq sequencing.

[0086] The labeled nucleic acids comprise nucleic acids representing from about 0.01% of the genes of an organism's genome to about 100% of the genes of an organism's genome. For example, about 0.01% of the genes of an organism's genome to about 100% of the genes of an organism's genome can be sequenced using a target complimentary region comprising a plurality of multimers by capturing the genes containing a complimentary sequence from the sample. The labeled nucleic acids comprise nucleic acids representing from about 0.01% of the transcripts of an organism's transcriptome to about 100% of the transcripts of an organism's transcriptome. For example, about 0.501% of the transcripts of an organism's transcriptome to about 100% of the transcripts of an organism's transcriptome can be sequenced using a target complimentary region comprising a poly-T tail by capturing the mRNAs from the sample.

[0087] Sequencing can comprise sequencing at least about 10, 20, 30, 40, 50, 60, 70, 80, 90, 100 or more nucleotides or base pairs of the labeled nucleic acid. In some instances, sequencing comprises sequencing at least about 200, 300, 400, 500, 600, 700, 800, 900, 1,000 or more nucleotides or base pairs of the labeled nucleic acid. In other instances, sequencing comprises sequencing at least about 1,500; 2,000; 3,000; 4,000; 5,000; 6,000; 7,000; 8,000; 9,000; or 10,000 or more nucleotides or base pairs of the labeled nucleic acid.

[0088] Sequencing can comprise at least about 200, 300, 400, 500, 600, 700, 800, 900, 1,000 or more sequencing reads per run. In some instances, sequencing comprises sequencing at least about 1,500; 2,000; 3,000; 4,000; 5,000; 6,000; 7,000; 8,000; 9,000; or 10,000 or more sequencing reads per run. Sequencing can comprise less than or equal to about 1,600,000,000 sequencing reads per run. Sequencing can comprise less than or equal to about 200,000,000 reads per run.

Microwells used for entrapment:

[0089] As described above, microwells are used to entrap single cells and beads (one bead per cell) within a small reaction chamber of defined volume. Each bead comprises a library of oligonucleotide probes for use in stochastic labeling and digital counting of nucleic acid targets {e.g., the entire complement of cellular mRNA molecules) which are released upon lysis of the cell.

Alternatives to microwells:

[0090] Individual cells and beads can be compartmentalized using alternatives to microwells, for example, a single bead and single cell could be confined within a single droplet in an emulsion {e.g., in a droplet digital micro fluidic system). Alternatively, cells could potentially be confined within porous beads that themselves comprise the plurality of tethered stochastic labels.

Individual cells and beads can be compartmentalized in any type of container, microcontainer, reaction chamber, reaction vessel, or the like. Thus, single cell, stochastic labeling or molecular barcoding assays can be performed without the use of microwells. Single cell, stochastic labeling or molecular barcoding assays can be performed without the use of any physical container, e.g., by embedding cells and beads in close proximity to each other within a polymer layer or gel layer to create a diffusional barrier between different cell/bead pairs.

Distribution of beads within microwells:

[0091] The beads comprising libraries of tethered stochastic labels can be distributed amongst a plurality of microwells as part of the assay procedure. The beads can be pre-loaded in a plurality of microwells as part of the manufacturing process for either flow cells or cartridges that incorporate a substrate comprising a plurality of microwells. The percentage of microwells that contain a single bead can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain a single bead. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the microwells in the plurality of microwells can contain a single bead. Distribution of cells within microwells:

[0092] In many embodiments, cells can be distributed amongst a plurality of microwells as part of the assay procedure. The percentage of microwells that contain a single cell can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain a single cell. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the microwells in the plurality of microwells can contain a single cell.

Microwells containing both a single cell and a single bead:

[0093] In many embodiments, cells and beads can be distributed amongst a plurality of microwells such that a fraction of the microwells contain both a single cell and a single bead. The percentage of microwells that contain both a single cell and a single bead can be between about 1% and about 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain both a single cell and a single bead. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the microwells in the plurality of microwells can contain both a single cell and a single bead.

Cell and bead distribution targets:

[0094] When distributing beads amongst a plurality of microwells, any of a variety of predetermined levels can be targeted. For example, the percentage of microwells that contain a single bead can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain a single bead. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the microwells in the plurality of microwells can contain a single bead.

[0095] When distributing beads amongst a plurality of microwells, the percentage of microwells that contain two beads can be at least 1%, at least 2%, at least 3%, at least 4%, at least 5%, at least 6%, at least 7%, at least 8%, at least 9%, or at least 10% or more. In other embodiments, the percentage of microwells that contain two beads can be at most 10%, at most 9%, at most 8%, at most 7%, at most 6%, at most 5%, at most 4%, at most 3%, at most 2%, or at most 1%. [0096] When distributing cells amongst a plurality of microwells, any of a variety of predetermined levels can be targeted. For example, the percentage of microwells that contain a single cell can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain a single cell. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the microwells in the plurality of microwells can contain a single cell.

[0097] When distributing cells amongst a plurality of microwells, the percentage of microwells that contain two cells can be at least 1%, at least 2%, at least 3%, at least 4%, at least 5%, at least 6%, at least 7%, at least 8%, at least 9%, or at least 10% or more. In other embodiments, the percentage of microwells that contain two cells can be at most 10%, at most 9%, at most 8%, at most 7%, at most 6%, at most 5%, at most 4%, at most 3%, at most 2%, or at most 1%.

[0098] When distributing both cells and beads amongst a plurality of microwells, any of a variety of pre-determined levels can be targeted. For example, the percentage of microwells that contain both a single cell and a single bead can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain both a single cell and a single bead. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the microwells in the plurality of microwells can contain both a single cell and a single bead.

[0099] Bead and/or cell distribution can be dependent on size and/or density. For example, larger beads and/or cells can settle (i.e., into wells) at a faster rate than smaller beads and/or cells. For example, beads that are 33 micron in diameter can settle about 0.5, 1, 1.5, 2, 2.5, or 3 or more times faster than beads that are 22 microns in diameter (assuming equal or similar density). In some instances, beads that are 33 microns in diameter settle about 2.25 times faster than beads that are 22 micron in diameter.

[0100] Cells of different sizes and/or densities can settle at different rates into the wells of the substrate. For example, red blood cells can settle at least 0.5, 1, 1.5, 2, 2.5, or 3 or more times faster than white blood cells. Red blood cells can settle faster than white blood cells due to their higher density in spite of their smaller size.

[0101] A buffer can be flowed over the microwell array before and/or after cells or beads have been loaded. The buffer can be a lysis buffer and/or a wash buffer. In many embodiments, flow of the buffer will not substantially remove the contents of the microwells. Flow of the buffer can remove the contents of at most 1%, at most 2%, at most 3%, at most 4%, at most 5%, at most 6%, at most 7%, at most 8%, at most 9%, or at most 10% or more microwells.

[0102] The viscosity and/or density of a buffer can be adjusted to optimize the uniform loading of beads and/or cells into microwells. Varying the viscosity or density of the loading buffer can, for example, help protect cells from shear forces or provide positive or negative buoyancy to cells and/or beads to facilitate uniform loading. For example, the viscosity of a buffer used for loading beads and/or cells can range from about lx to about lOx that of water. The viscosity of the buffer can be at least lx, at least l. lx, at least 1.2x, at least 1.3x, at least 1.4x, at least 1.5x, at least 1.6x, at least 1.7x, at least 1.8x, at least 1.9x, at least 2x, at least 3x, at least 4x, at least 5x, at least 6x, at least 7x, at least 8x, at least 9x, or at least lOx or more times the viscosity of water. The viscosity of the buffer can be at most lOx, at most 9x, at most 8x, at most 7x, at most 6x, at most 5x, at most 4x, at most 3x, at most 2x, at most 1.9x, at most 1.8x, at most 1.7x, at most 1.6x, at most 1.5x, at most 1.4x, at most 1.3x, at most 1.2x, at most l. lx, or at most lx that of water. Those of skill in the art will recognize that the buffer viscosity can have any value within this range, for example, about 1.75x that of water.

[0103] Similarly, the density of a buffer used for loading beads and/or cells can range from about 0.8x to about 1.25x that of the density of the beads and/or cells to be loaded. The density of a buffer used for loading beads and/or cells can be at least 0.8x, at least 0.9x, at least l.Ox, at least l. lx, at least 1.2x, or at least 1.25x that of the beads and/or cells, or higher. The density of a buffer used for loading beads and/or cells can be at most 1.25x, at most 1.2x, at most l. lx, at most l.Ox, at most 0.9x, or at most 0.8x that of the beads and/or cells, or lower. Those of skill in the art will recognize that the density of the buffer used for loading beads and/or cells can have any value within this range, for example, about 0.85x that of the beads and/or cells.

[0104] The viscosity and/or density of a buffer can be adjusted to optimize the efficiency of retrieving beads from microwells. Varying the viscosity or density of the bead retrieval buffer can, for example, provide viscous drag forces or provide positive or less negative buoyancy to beads to facilitate efficient bead retrieval. For example, the viscosity of a buffer used for bead retrieval can range from about lx to about lOx that of water. The viscosity of the buffer can be at least lx, at least l. lx, at least 1.2x, at least 1.3x, at least 1.4x, at least 1.5x, at least 1.6x, at least 1.7x, at least 1.8x, at least 1.9x, at least 2x, at least 3x, at least 4x, at least 5x, at least 6x, at least 7x, at least 8x, at least 9x, or at least lOx or more times the viscosity of water. The viscosity of the buffer can be at most lOx, at most 9x, at most 8x, at most 7x, at most 6x, at most 5x, at most 4x, at most 3x, at most 2x, at most 1.9x, at most 1.8x, at most 1.7x, at most 1.6x, at most 1.5x, at most 1.4x, at most 1.3x, at most 1.2x, at most l. lx, or at most lx that of water. Those of skill in the art will recognize that the buffer viscosity can have any value within this range, for example, about 2.3x that of water.

[0105] Similarly, the density of a buffer used for bead retrieval can range from about 0.8x to about 1.25x that of the density of the beads. The density of a buffer used for bead retrieval can be at least 0.8x, at least 0.9x, at least l.Ox, at least l. lx, at least 1.2x, or at least 1.25x that of the beads and/or cells, or higher. The density of a buffer used for bead retrieval can be at most 1.25x, at most 1.2x, at most l. lx, at most l.Ox, at most 0.9x, or at most 0.8x that of the beads and/or cells, or lower. Those of skill in the art will recognize that the density of the buffer used for bead retrieval can have any value within this range, for example, about l. lx that of the beads.

[0106] Examples of buffer additives that can be used to adjust buffer viscosity and/or density include, but are not limited to sucrose, polyethylene glycol (PEG), Ficoll, glycerin, glycerol, dextran sulfate, histopaque, bovine serum albumin, and the like.

Magnetic field-assisted bead transport & manipulation:

[0107] Cells or beads can be distributed among the microwells, removed from the microwells, or otherwise transported through a flow cell or cartridge of an instrument system by using magnetic beads (e.g., conjugated to antibodies directed against cell surface markers, or as solid supports for libraries of stochastic labels) and externally-applied magnetic field gradients. When using magnetic fields to trap magnetic beads in microwells or to elute magnetic beads from microwells, an externally-applied magnetic field gradient can be applied to the entire microwell pattern simultaneously. An externally-applied magnetic field gradient can be applied to a selected area of the microwell pattern. An externally-applied magnetic field gradient can be applied to a single microwell. Permanent magnets can be used to apply time- varying magnetic field gradients by moving the position of one or more permanent magnets relative to the microwell array or vice versa. In these embodiments, the velocity of the relative motion can be adjusted to so that the time-dependence of the magnetic field gradient is matched to the timescale on which magnetic beads undergo magnetophoresis into or out of microwells. In some embodiment, time- varying magnetic fields can be provided by varying the current applied to one or more electromagnets. A combination of one or more permanent magnets and one or more electromagnets can be used to provide magnetic field gradients for transporting magnetic beads into microwells, out of microwells, or through the device. Cells or beads can be distributed among the microwells, removed from the microwells, or otherwise transported through a flow cell or cartridge of an instrument system by means of centrifugation or other non-magnetic means.

[0108] Beads (solid supports) can be removed from the microwells using one or more magnetic fields. Beads can be removed after lysis of cells in the microwells and/or attachment of nucleic acids to the pluralities of oligonucleotides immobilized on the individual beads. A magnet can be place on top of the cartridge and beads can be removed from the wells using the resultant magnetic field. At least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, or at least 100% of the beads can be removed. At most 70%, at most 75%, at most 80%, at most 85%, at most 90%, at most 95%, or at most 100% of the beads can be removed.

Real-time imaging & feedback to improve cell and/or bead distribution:

[0109] An imaging system and real-time image processing and analysis can be used to monitor the cell and bead distribution processes (i.e., the distribution of cells and/or beads within the plurality of microwells) and feedback can be used to adjust process steps accordingly, e.g., by prolonging or repeating some steps, by activating alternative cell or bead distribution

mechanisms, and the like, in order to improve cell and/or bead distributions, or to achieve pre- specified target distributions.

Real-time imaging & feedback to select sub-populations of cells:

[0110] Real-time image processing and analysis can be used to identify wells containing cells exhibiting one or more specified characteristics (as described in more detail below), followed by selection or exclusion of a subset of cells from further analysis. Real-time image processing and analysis can be used to identify wells containing two or more cells, followed by the exclusion of the cells in those wells from further analysis. Examples of mechanisms that can be used to select or exclude a subset of cells from further analysis (e.g., selection mechanisms) include, but are not limited to, (i) physical removal of selected cells or beads from the array, (ii) physical entrapment of selected cells or beads within the array, (iii) physical destruction of selected cells or beads within the array, or (iv) use of dual-encoding schemes whereby the sequence data that is generated for a given cell is selected for or excluded from further analysis.

[0111] One non-limiting example of a selection mechanism for physical removal of selected beads (wherein the beads are magnetic beads) from microwells (thereby preventing downstream sequence analysis of nucleic acid target molecules from the corresponding cell(s)) is the use of miniaturized magnetic probes (e.g., modified computer hard drive write heads) to pluck beads from wells containing cells that have been identified to match a pre-specified set of cellular characteristics. Modern hard drives can reach bit densities of over 800 Gbit/in", which

corresponds to a bit area of 8.1 x 10^" 4 μιη 2. Hard drive technology dating from the 1980s had bit areas of approximately 50 μιη or less, hence magnetic write head technology can be adapted for the purpose of removing beads from specified wells.

[0112] Another example of a selection mechanism for physical removal of selected cells or beads from microwells is the use of micropipettes and micromanipulators. In embodiments where the microwells are accessible, e.g., through the use of a removable chamber wall or optical window, a micromanipulator can be used to position one or more micropipettes, extract selected cells or beads from microwells (e.g. , by applying gentle suction to the one or more micropipettes), and moving them to a position in the cartridge or instrument (e.g. , a reservoir) where they can be sequestered for subsequent disposal or subsequent processing and analysis. Commercially- available micromanipulators provide sub-micron step resolution for precise positioning, while commercially- available micropipette pullers permit fabrication of tip diameters of about 5 μιη and smaller.

[0113] Another example of a selection mechanism for physical removal of selected cells or beads from microwells is the use of single light beam gradient force traps (e.g. , "optical tweezers"). These apparatus use a highly focused laser beam to create an optical trap (e.g., by generating attractive or repulsive forces, depending on the mismatch in refractive index between an object and the surrounding medium) to physically hold and move microscopic dielectric objects. Optical tweezers can be used to extract selected cells or beads from microwells and move them, either directly using the optical tweezers or through the simultaneous control of fluid flow through the flow cell or microwell chamber, to a position in the cartridge (e.g. , a reservoir) where they can be sequestered for subsequent disposal or subsequent processing and analysis.

[0114] Another example of a selection mechanism for physical removal of selected cells or beads from microwells can be the use of acoustic droplet ejection. Acoustic droplet ejection has been used for precision dispensing of small droplets of liquid (ranging in volume from several hundred picoliters to several hundred nanoliters), and has also been used to dispense cells, beads, and protein microcrystals. Focused acoustic energy generated by a microfabricated ultrasonic transducer might be used to eject and capture selected cells or beads from microwells in order to exclude them from downstream assay and analysis steps. Typically the size of the droplets can be controlled by adjusting ultrasound parameters such as pulse frequency and amplitude. The ejected cells or beads might be retained for selective use in downstream assay and analysis steps.

[0115] A related example of a selection mechanism for ensuring that data for selected cells is eliminated from downstream processing, without requiring physical removal of cells or beads from microwells, would be the use of photocleavable linkers for the attachment of stochastic labels to beads. Beads co-localized with specified cells would be illuminated with a focused light beam (typically UV light) to release the attached labels and allow them to be rinsed away and eliminated from further assay steps. This approach will likely require that suitable conditions be identified for achieving efficient photolysis while leaving the remaining cells in the microwell pattern intact. [0116] An example of a selection mechanism for physical destruction of selected cells or beads in microwells is the use of laser photoablation, in which focused laser light, e.g., focused C0₂ or excimer laser pulses can be used to selectively break bonds and remove material while causing little or no damage to surrounding materials.

[0117] One non-limiting example of a selection mechanism for physical entrapment of magnetic beads within microwells is the use of miniaturized magnetic probes, e.g., magnetic write head technology originally developed for computer hard disks can be adapted for the purpose of trapping beads in specified wells. One or more modified microfabricated magnetic write heads could be moved into proximity with one or more specified microwells and activated to hold the corresponding beads in place, thereby preventing them from elution and downstream assay steps. An array of microfabricated electromagnets can be fabricated on one surface of a microwell array substrate (or within the substrate itself) to create an addressable array of magnetic probes that can be used to trap selected beads.

[0118] Other examples of selection mechanisms for physical entrapment of cells or beads include the use of bead or microwell substrate materials that shrink or swell upon exposure to a localized physical or chemical stimulus. Beads can be fabricated from a suitable material such that selected beads, i.e., those associated with a specified subset of cells in the microwells, are subjected to a local stimulus and swell such that they cannot be removed from the microwells within which they are located, thereby effectively removing them from further assay process steps.

Alternatively, microwells can be fabricated from a suitable material such that selected wells, i.e., those wells containing a specified subset of cells, are subjected to a local stimulus and shrink such that the beads contained within cannot be removed, thereby effectively removing them from further assay process steps. Examples of suitable swellable or shrinkable materials can include thermoresponsive polymer gels (which exhibit a discontinuous change in degree of swelling with temperature), pH-sensitive polymers (which shrink or swell depending on local pH), electro- responsive polymers (which shrink or swell in response to local electric fields), and light- responsive polymers (which shrink or swell in response to exposure to UV or visible light).

Distribution of more than one cell type:

[0119] The system can include functionality for distributing more than one cell type over the microwell array. For example, the system can load the microwell array with a first cell type A, followed by rinsing and subsequent loading with a second cell type B, such that a plurality of microwells contain a single cell of type A and a single cell of type B. Such system functionality can be useful in studying cell-cell interactions and other applications. In general, the system can be configured to distribute at least one cell type, at least two cell types, at least three cell types, at least four cell types, or at least five cell types over the microwell array. The system can be configured to distribute at most five cell types, at most four cell types, at most three cell types, at most two cell types, or at most one cell type over the microwell array. The system can be configured to distribute complex mixtures of cells over the microwell array. In all of these configurations, the system can be set up to optimize the distribution of cells in microwells, and to identify wells having a greater or lesser number of cells than a specified number of cells, using cell distribution, real-time imaging, and feedback mechanisms as described above. In general, the percentage of microwells that contain more than one cell type, e.g., one cell each of types A and B, or one cell each from types A, B, and C, can range from about 1% to about 100%. The percentage of microwells that contain more than one cell type can be at least 1%, at least 5%, at least 10%, at least 20%, at least 40%, at least 60%, at least 80%, or at least 90%. In other embodiments, the percentage of microwells that contain more than one cell type can be at most 100%, at most 90%, at most 80%, at most 60%, at most 40%, at most 20%, at most 10%, at most 5%, or at most 1%. In specific embodiment, the percentage of microwells that contain more than one cell type can have a value that falls anywhere within this range, e.g., about 8.5%.

Cell lysis mechanisms:

[0120] Cell lysis can be accomplished by chemical or biochemical means, by osmotic shock, or by means of thermal lysis, mechanical lysis, or optical lysis. Cells can be lysed by addition of a cell lysis buffer comprising a detergent (e.g., SDS, Li dodecyl sulfate, Triton X-100, Tween-20, or NP-40), an organic solvent (e.g., methanol or acetone), or digestive enzymes (e.g., proteinase K, pepsin, or trypsin), or combinations thereof. One or more of these reagents (for example, bead suspensions) can be stored in bottles or containers that are connected to the microwell flow cell or cartridge when inserted into the instrument. The reagents (for example, bead suspensions) can be pre-loaded into the cartridge.

[0121] The instrument system can include mechanical cell lysis capability as an alternative to the use of detergents or other reagents. Sonication using a high frequency piezoelectric transducer is one example of a suitable technique.

Magnetic field control:

[0122] As indicated elsewhere in this disclosure, many embodiments of the disclosed methods utilize magnetic fields for removing beads from the microwells upon completion of the assay.

The instrument system can further comprise use of magnetic fields for transporting beads into or out of the microwell flow cell or chamber, or through other parts of the instrument system, or for retaining or trapping beads in particular locations after they have been loaded or distributed prior to the assay or during the assay. Examples of suitable means for providing control of magnetic fields include, but are not limited to, use of electromagnets in fixed position(s) relative to the cartridge, or the use of permanent magnets that are mechanically repositioned as necessary. The strength of the applied magnetic field(s) can be varied by varying the amount of current applied to one or more electromagnets. The strength of the applied magnetic fields can be varied by changing the position of one or more permanent magnets relative to the position of the microwell chamber(s) using, for example, stepper motor-driven linear actuators, servo motor-driven linear actuators, or cam shaft mechanisms. In other embodiments, the positions of magnets can be controlled in a linear (or non-linear) fashion, with speeds chosen to maximize bead collection efficiency, as opposed to performing transitions between just two fixed positions. The use of pulsed magnetic fields can be advantageous, for example, to prevent clustering of magnetic beads.

[0123] In addition to consideration of the strength and location of magnetic fields for

manipulating beads and other materials, it is important to design the system such that the magnetic field gradient is suitable for the task being performed. It is spatial gradients in magnetic field which exert translational force on magnetic materials and particles. Suitable gradients in fields can be achieved by the use of multiple magnets, the use of magnets or magnetized materials with particular edge and face geometries, and by designing magnets with appropriate spatial scale. Here, the term "magnets" refers to permanent magnets or electromagnets. Magnet assemblies comprising multiple magnetic domains, formed intrinsically or by design, can be used to generate magnetic fields with desirable field strengths and spatial variations. For example, patterns of small magnets with parallel or antiparallel field axes, or other relative angles, can be placed adjacent to the pattern of wells and fluidics, to achieve optimal trapping or manipulation of beads during the loading and operation of the device. When using magnetic fields to trap magnetic beads in microwells or to elute magnetic beads from microwells, an externally- applied magnetic field gradient can be applied to the entire microwell pattern simultaneously. Externally- applied magnetic field gradients can be applied to a selected area of the microwell pattern. An externally-applied magnetic field gradient can be applied to a single microwell. The magnetic field lines for an externally-applied magnetic field can lie at an angle relative to the plane of the microwell substrate of between about 30 degrees and 89 degrees. The angle of the magnetic field lines relative to the plane of the microwell substrate can be between about 45 degrees and 80 degrees. The angle of the magnetic field lines relative to the plane of the microwell substrate can be at least 45 degrees, at least 50 degrees, at least 55 degrees, at least 60 degrees, at least 65 degrees, at least 70 degrees, at least 75 degrees, or at least 80 degrees, or higher. The angle of the magnetic field lines relative to the plane of the microwell substrate can be at most 80 degrees, at most 75 degrees, at most 70 degrees, at most 65 degrees, at most 60 degrees, at most 55 degrees, at most 50 degrees, or at most 45 degrees, or smaller. Those of skill in the art will recognize that the angle of the magnetic field lines relative to the plane of the microwell substrate can have any value within this range, for example, about 52 degrees.

Interfaces with PCR thermocyclers, sequencers, & FACS instruments:

[0124] The instrument systems of the present disclosure can further comprise interfaces with PCR thermocyclers, sequencers, cell sorters, fluorescence-activated cell sorter (FACS) instruments, or other types of lab automation equipment.

[0125] An interface for PCR thermocyclers can be provided such that instrument system outputs labeled oligonucleotide libraries directly into tubes, strips, or plates that are compatible with commercially- available PCR instruments, for example, the Roche LightCycler® series of realtime PCR instruments, and the like.

[0126] An interface can be provided for cell sorters or FACS instruments such that sorted cells are deposited directly into a microwell array or cartridge. The interface for FACS instruments can, for example, include both hardware and software components, where the software provides the capability for simultaneous control of the FACS instrument and the single cell, stochastic labeling or molecular barcoding system. The software can provide analysis capability for identifying correlations between the FACS data (e.g., the presence or absence of specified cell surface markers) and the copy numbers for one or more genes in a specified sub-population of cells. FACS machines can be used to sort single cells directly into the microwell array of the disclosure.

[0127] An interface with lab automation equipment in general can be provided. For example, cartridges for use with the disclosed instrument systems can be configured to have inlet ports of the proper dimension and spacing such that samples and reagents can be dispensed directly into the cartridge using commercially-available pipetting stations and liquid-handling robotics.

Similarly, cartridges for use with the disclosed instrument systems can be configured to have dimensions that are compatible with commercially-available plate-handling robotics for automated storage, retrieval, or movement between other laboratory workstations.

System processor and software:

[0128] In general, instrument systems designed to support the automation of multiplexed, single cell stochastic labeling and molecular barcoding assays can include a processor or computer, along with software to provide (i) instrument control functionality, (ii) image processing and analysis capability, and (iii) data storage, analysis, and display functionality. System processor and control software:

[0129] In many embodiments, the instrument system will comprise a computer (or processor) and computer-readable media that includes code for providing a user interface as well as manual, semi-automated, or fully- automated control of all system functions, e.g., control of the fluidics system, the temperature control system, cell or bead distribution functions, magnetic bead manipulation functions, and the imaging system. The system computer or processor can be an integrated component of the instrument system (e.g., a microprocessor or mother board embedded within the instrument). The system computer or processor can be a stand-alone module, for example, a personal computer or laptop computer. Examples of fluid control functions provided by the instrument control software include, but are not limited to, volumetric fluid flow rates, fluid flow velocities, the timing and duration for sample and bead addition, reagent addition, and rinse steps. Examples of temperature control functions provided by the instrument control software include, but are not limited to, specifying temperature set point(s) and control of the timing, duration, and ramp rates for temperature changes. Examples of cell or bead distribution functions provided by the instrument control software include, but are not limited to, control of agitation parameters such as amplitude, frequency, and duration. Examples of magnetic field functions provided by the instrument control software include, but are not limited to, the timing and duration of the applied magnetic field(s), and in the case of

electromagnets, the strength of the magnetic field as well. Examples of imaging system control functions provided by the instrument control software include, but are not limited to, autofocus capability, control of illumination or excitation light exposure times and intensities, control of image acquisition rate, exposure time, and data storage options.

Image processing software:

[0130] The system can further comprise computer-readable media that includes code for providing image processing and analysis capability. Examples of image processing and analysis capability provided by the software include, but are not limited to, manual, semi-automated, or fully-automated image exposure adjustment (e.g., white balance, contrast adjustment, signal- averaging and other noise reduction capability, etc.), automated edge detection and object identification (i.e., for identifying cells and beads in the image), automated statistical analysis (i.e., for determining the number of cells or beads identified per micro well or per unit area of the microwell substrate, or for identifying wells that contain more than one cell or more than one bead), and manual measurement capabilities (e.g., for measuring distances between objects, etc.). The instrument control and image processing/analysis software can be written as separate software modules. The instrument control and image processing/analysis software can be incorporated into an integrated package.

[0131] The system software can provide integrated real-time image analysis and instrument control, so that cell and bead sample loading steps can be prolonged, modified, or repeated until optimal cell and bead distributions (e.g. , uniformly distributed across the microwell pattern at a pre-determined level for the number of wells containing a single cell, the number of wells containing a single bead, or the number of wells containing both a single cell and a single bead) are achieved. Any of a number of image processing and analysis algorithms known to those of skill in the art can be used to implement real-time or post-processing image analysis capability. Examples include, but are not limited to, the Canny edge detection method, the Canny-Deriche edge detection method, first-order gradient edge detection methods (e.g. , the Sobel operator), second order differential edge detection methods, phase congruency (phase coherence) edge detection methods, other image segmentation algorithms (e.g. , intensity thresholding, intensity clustering methods, intensity histogram-based methods, etc.), feature and pattern recognition algorithms (e.g. , the generalized Hough transform for detecting arbitrary shapes, the circular Hough transform, etc.), and mathematical analysis algorithms (e.g. , Fourier transform, fast Fourier transform, wavelet analysis, auto-correlation, etc.), or combinations thereof. As outlined above, examples of mechanisms for facilitating cell and bead distribution which can be controlled through feedback from real-time image analysis include, but are not limited to, rocking, shaking, swirling, recirculating flow, oscillatory or pulsatile flow, low frequency agitation (for example, through pulsing of a flexible (e.g., silicone) membrane that forms a wall of the chamber or nearby fluid channel), or high frequency agitation (for example, through the use of piezoelectric transducers). The instrument system can monitor the total number of cells captured in the microwells, as determined by image processing and analysis, and turn off the supply of cells when a pre-determined number of cells is reached in order to avoid loading an excess number of wells with two or more cells. The instrument system can monitor the number of wells containing single cells, and turn off the supply of cells when a pre-determined number of wells are reached in order to avoid loading an excess number of wells with two or more cells. The instrument system can monitor the number of wells containing single beads, and turn off the supply of beads when a predetermined number of wells is reached in order to avoid loading an excess number of wells with two or more beads. The instrument system can monitor the number of wells containing both a single cell and a single bead, and turn off the supply of cells or beads (or both) in order to avoid loading an excess number of wells with two or more cells or beads.

[0132] When using integrated real-time image analysis and instrument control to achieve optimal bead distributions, any of a variety of pre-determined levels can be targeted. For example, the percentage of microwells that contain a single bead can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain a single bead. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the microwells in the plurality of microwells can contain a single bead.

[0133] When using integrated real-time image analysis and instrument control to achieve optimal cell distributions, any of a variety of pre-determined levels can be targeted. For example, the percentage of microwells that contain a single cell can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain a single cell. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the microwells in the plurality of microwells can contain a single cell.

[0134] Real-time image analysis can comprise monitoring cell capture efficiency in the microwells (i.e. , determining the number of wells that have a cell in them and/or determining the percentage of cells that are in between wells). Cell capture can be improved by methods such as agitation, washing (i.e., flushing) fluid, and/or magnetic methods. For example, the percentage of cells that can be between microwells can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the cells can be between microwells (e.g., on the surface between microwells). At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the cells can be between microwells (e.g. , on the surface between microwells).

[0135] When using integrated real-time image analysis and instrument control to achieve optimal cell and bead distributions, e.g. , to maximize the percentage of microwells containing both a single cell and a single bead, any of a variety of pre-determined levels can be targeted. For example, the percentage of microwells that contain both a single cell and a single bead can be between 1% and 100%. At least 1%, at least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 99% of the microwells in the plurality of microwells can contain both a single cell and a single bead. At most 100%, at most 99%, at most 95%, at most 90%, at most 80%, at most 70%, at most 60%, at most 50%, at most 40%, at most 30%, at most 20%, at most 10%, at most 5%, or at most 1% of the micro wells in the plurality of micro wells can contain both a single cell and a single bead.

[0136] The system software can provide integrated real-time image analysis and instrument control, so that cells can be optically monitored and classified according to a pre-determined set of characteristics, and subsequently included or excluded from the downstream sequence data analysis. Examples of cellular characteristics that can be optically monitored and used for classification purposes include, but are not limited to, cell size, cell shape, live cell / dead cell determination (e.g., using selectively absorbed chromophores such as Trypan blue, or fluorescent dyes such as calcein AM, ethidium homodimer- 1, DiOC₂(3), DiOC₅(3), DiOC₆(3), DiSC₃(5), DiICi(5), DiOCi₈(3), propidium iodide, SYBR® 14, SYTOX® Green, etc.), cells exhibiting a specified range of intracellular pH (e.g. , using intracellular pH-sensitive fluorescent probes such as 2',7'-Bis-(2-carboxyethyl)-5-(and-6-)carboxyfluorescein (BCECF), 2',7'-bis-(2- carboxypropyl)-5-(and-6-)-carboxyfluorescein (BCPCF), etc.), cells exhibiting a specified range of membrane potential (e.g. , using membrane potential- sensitive fluorophores such as

FluoVolt™, di-3-ANEPPDHQ, Bis-(1,3-Dibutylbarbituric Acid) Trimethine Oxonol

(DiBAC₄(3)), DiBAC₄(5), DiSBAC₂(3), Merocyanine 540, JC- 1, JC-9, Oxonol V, Oxonol VI, Tetramethylrhodamine methyl and ethyl esters, Rhodamine 123, Di-4-ANEPPS, Di-8-ANEPPS, Di-2-ANEPEQ, Di-3-ANEPPDHQ, Di-4-ANEPPDHQ, etc.), cells exhibiting a specified level of intracellular calcium (e.g. , using Ca²⁺- sensitive fluorescent dyes such as fura-2, indo- 1, fluo-3, fluo-4, Calcium Green- 1, Quin 2, etc.), cells exhibiting one or more specified cell surface markers (e.g. , using fluorescently-labeled antibodies directed towards the cell surface markers), cells expressing fluorescent proteins (e.g. , GFP, bilirubin- inducible fluorescent protein, UnaG, dsRed, eqFP611, Dronpa, TagRFPs, KFP, EosFP, Dendra, IrisFP, etc.), and the like. In many embodiments, two or more dyes, fluorophores, or other optical probes having non-overlapping spectral properties (e.g. , non-overlapping excitation peaks, non-overlapping absorption or emission peaks, etc.) can be selected so that cells can be simultaneously characterized with respect to two or more properties. Real-time image processing and analysis can be used to identify wells containing cells exhibiting one or more specified characteristics, followed by selection or exclusion of a subset of cells on the array from further analysis. Real-time image processing and analysis can be used to identify wells containing two or more cells, followed by the exclusion of the cells in those wells from further analysis. As described above in more detail, examples of mechanisms that can be used to select or exclude a subset of cells from further analysis include, but are not limited to, (i) physical removal of selected cells or beads from microwells by means of miniaturized magnetic probes, optical tweezer apparatus, micromanipulators, photoablation, etc., (ii) physical entrapment of selected cells or beads within microwells by means of miniaturized magnetic probes, swellable beads, shrinkable wells, or (iii) use of dual-encoding schemes whereby the sequence data that is generated for a given cell is selected for or excluded from further analysis.

[0137] The system software (or a stand-alone software module) can provide image analysis capability for automated cell counting using a hemocytometer, thereby allowing users to determine how much cell suspension to load onto the microwell array substrate. The automated cell counting capability can be coupled with optional fluorescence image analysis so that cells can be characterized with respect to viability or other properties (using, for example, the optical probed described above) at the same time that counting is performed.

Data analysis & display software:

[0138] The system can comprise computer-readable media that includes code for providing data analysis for the sequence datasets generated by performing single cell, stochastic labeling or molecular barcoding assays. Examples of data analysis functionality that can be provided by the data analysis software include, but are not limited to, (i) algorithms for decoding/demultiplexing of the sample barcode, cell barcode, molecular barcode, and target sequence data provided by sequencing the oligonucleotide library created in running the assay, (ii) algorithms for determining the number of reads per gene per cell, and the number of unique transcript molecules per gene per cell, based on the data, and creating summary tables, (iii) statistical analysis of the sequence data, e.g., for clustering of cells by gene expression data, or for predicting confidence intervals for determinations of the number of transcript molecules per gene per cell, etc., (iv) algorithms for identifying sub-populations of rare cells, for example, using principal component analysis, hierarchical clustering, k-mean clustering, self-organizing maps, neural networks etc., (v) sequence alignment capabilities for alignment of gene sequence data with known reference sequences and detection of mutation, polymorphic markers and splice variants, and (vi) automated clustering of molecular labels to compensate for amplification or sequencing errors. Commercially-available software can be used to perform all or a portion of the data analysis, for example, the Seven Bridges software can be used to compile tables of the number of copies of one or more genes occurring in each cell for the entire collection of cells. The data analysis software can include options for outputting the sequencing results in useful graphical formats, e.g., heatmaps that indicate the number of copies of one or more genes occurring in each cell of a collection of cells. The data analysis software can further comprise algorithms for extracting biological meaning from the sequencing results, for example, by correlating the number of copies of one or more genes occurring in each cell of a collection of cells with a type of cell, a type of rare cell, or a cell derived from a subject having a specific disease or condition. In some embodiments, the data analysis software can further comprise algorithms for comparing populations of cells across different biological samples. In some embodiments, the data analysis software can further comprise algorithms for categorizing sequencing results in different categories based on the presence of a signal sequence, such as a cell surface protein category in which the sequences comprise a signal sequence that directs the resulting peptide or protein to the cell membrane.

[0139] All of the data analysis functionality can be packaged within a single software package. The complete set of data analysis capabilities can comprise a suite of software packages. The data analysis software can be a standalone package that is made available to users independently of the assay instrument system. The software can be web-based, and can allow users to share data.

System processors & networks:

[0140] In general, the computer or processor included in the presently disclosed instrument systems can be further understood as a logical apparatus that can read instructions from media or a network port, which can optionally be connected to server having fixed media. The system can include a CPU, disk drives, optional input devices such as keyboard or mouse and optional monitor. Data communication can be achieved through the indicated communication medium to a server at a local or a remote location. The communication medium can include any means of transmitting or receiving data. For example, the communication medium can be a network connection, a wireless connection or an internet connection. Such a connection can provide for communication over the World Wide Web. It is envisioned that data relating to the present disclosure can be transmitted over such networks or connections for reception or review by a party.

[0141] An example architecture of a computer system that can be used in connection with example embodiments of the present disclosure is as follows. The example computer system can include a processor for processing instructions. Non-limiting examples of processors include: Intel Xeon™ processor, AMD Opteron™ processor, Samsung 32-bit RISC ARM 1176JZ(F)-S vl.0™ processor, ARM Cortex-A8 Samsung S5PC100™ processor, ARM Cortex-A8 Apple A4™ processor, Marvell PXA 930™ processor, or a functionally-equivalent processor. Multiple threads of execution can be used for parallel processing. Multiple processors or processors with multiple cores can also be used, whether in a single computer system, in a cluster, or distributed across systems over a network comprising a plurality of computers, cell phones, or personal data assistant devices. [0142] A high speed cache can be connected to, or incorporated in, the processor to provide a high speed memory for instructions or data that have been recently, or are frequently, used by processor. The processor is connected to a north bridge by a processor bus. The north bridge is connected to random access memory (RAM) by a memory bus and manages access to the RAM by the processor. The north bridge is also connected to a south bridge by a chipset bus. The south bridge is, in turn, connected to a peripheral bus. The peripheral bus can be, for example, PCI, PCI-X, PCI Express, or other peripheral bus. The north bridge and south bridge are often referred to as a processor chipset and manage data transfer between the processor, RAM, and peripheral components on the peripheral bus. In some alternative architectures, the functionality of the north bridge can be incorporated into the processor instead of using a separate north bridge chip.

[0143] The system can include an accelerator card attached to the peripheral bus. The accelerator can include field programmable gate arrays (FPGAs) or other hardware for accelerating certain processing. For example, an accelerator can be used for adaptive data restructuring or to evaluate algebraic expressions used in extended set processing.

[0144] Software and data are stored in external storage and can be loaded into RAM or cache for use by the processor. The system includes an operating system for managing system resources; non-limiting examples of operating systems include: Linux, Windows™, MacOS™, BlackBerry OS™, iOS™, and other functionally-equivalent operating systems, as well as application software running on top of the operating system for managing data storage and optimization in accordance with example embodiments of the present invention.

[0145] In this example, the system also includes network interface cards (NICs) connected to the peripheral bus for providing network interfaces to external storage, such as Network Attached Storage (NAS) and other computer systems that can be used for distributed parallel processing.

[0146] The present disclosure provides for a network with a plurality of computer systems, a plurality of cell phones and personal data assistants, and Network Attached Storage (NAS). In example embodiments, computer systems can manage data storage and optimize data access for data stored in Network Attached Storage (NAS). A mathematical model can be used for the data and be evaluated using distributed parallel processing across computer systems, and, and cell phone and personal data assistant systems. Computer systems and cell phone and personal data assistant systems can also provide parallel processing for adaptive data restructuring of the data stored in Network Attached Storage (NAS). A wide variety of other computer architectures and systems can be used in conjunction with the various embodiments of the present invention. For example, a blade server can be used to provide parallel processing. Processor blades can be connected through a back plane to provide parallel processing. Storage can also be connected to the back plane or as Network Attached Storage (NAS) through a separate network interface. [0147] In some example embodiments, processors can maintain separate memory spaces and transmit data through network interfaces, back plane or other connectors for parallel processing by other processors. In other embodiments, some or all of the processors can use a shared virtual address memory space.

[0148] The present disclosure provides a multiprocessor computer system using a shared virtual address memory space in accordance with an example embodiment. The system includes a plurality of processors that can access a shared memory subsystem. The system incorporates a plurality of programmable hardware memory algorithm processors (MAPs) in the memory subsystem. Each MAP can comprise a memory and one or more field programmable gate arrays (FPGAs). The MAP provides a configurable functional unit and particular algorithms or portions of algorithms can be provided to the FPGAs for processing in close coordination with a respective processor. For example, the MAPs can be used to evaluate algebraic expressions regarding the data model and to perform adaptive data restructuring in example embodiments. In this example, each MAP is globally accessible by all of the processors for these purposes. In one configuration, each MAP can use Direct Memory Access (DMA) to access an associated memory, allowing it to execute tasks independently of, and asynchronously from, the respective microprocessor. In this configuration, a MAP can feed results directly to another MAP for pipelining and parallel execution of algorithms.

[0149] The above computer architectures and systems are examples only, and a wide variety of other computer, cell phone, and personal data assistant architectures and systems can be used in connection with example embodiments, including systems using any combination of general processors, co-processors, FPGAs and other programmable logic devices, system on chips (SOCs), application specific integrated circuits (ASICs), and other processing and logic elements. All or part of the computer system can be implemented in software or hardware. Any variety of data storage media can be used in connection with example embodiments, including random access memory, hard drives, flash memory, tape drives, disk arrays, Network Attached Storage (NAS) and other local or distributed data storage devices and systems.

[0150] In example embodiments, the computer subsystem of the present disclosure can be implemented using software modules executing on any of the above or other computer architectures and systems. In other embodiments, the functions of the system can be implemented partially or completely in firmware, programmable logic devices such as field programmable gate arrays (FPGAs), system on chips (SOCs), application specific integrated circuits (ASICs), or other processing and logic elements. For example, the Set Processor and Optimizer can be implemented with hardware acceleration through the use of a hardware accelerator card. Cell Surface Protein Repertoire:

[0151] The molecular barcoding and high-throughput sequencing as described above can be used for characterizing a cell surface protein repertoire of a single cell. A single cell can be a first cell or a second cell. The methods of the disclosure can be used for detecting the expression profile of a first cell that can interact with a second cell or a second cell that can interact with a first cell. The methods of the disclosure can be used for detecting the expression profile of cell surface proteins of a single cell, such as a first cell that can interact with a second cell or a second cell that can interact with a first cell. The method can comprise obtaining a biological sample containing a mixed population of cells from an individual, contacting said sample (which can be enriched) to the devices and systems such that a single cell is in a single well. The sample can be contacted with beads of the disclosure such that a single bead is in a single well with a single cell. The cell can be lysed. The bead can comprise a stochastic label that can bind to a specific location or gene in the cell and/or mRNAs of the cell. The molecules from the single cell associated with solid support can be subjected to the molecular biology methods of the disclosure, including reverse transcription, amplification, and sequencing.

[0152] The gene expression profile of the single cell can be divided into categories based on the presence of a signal sequence. One category can be cell surface proteins. The cell surface protein gene expression profile category can be generated based on the presence of a signal sequence in the mRNA, which can direct the subsequent peptide or protein to the cell membrane. This cell surface protein category can be indicative of the cell surface protein repertoire of the single cell.

Comparison of Cell Surface Protein Repertoires

[0153] Once a first cell surface protein repertoire and a second cell surface protein repertoire are identified, these cell surface protein repertoires can be compared and analyzed. For example, the presence or absence of cell surface proteins on the first cell can be compared with presence or absence of cell surface proteins with the second cell. This comparison can then be used to determine which cell surface proteins on the first cell can interact with the cell surface proteins on the second cell to create a cell surface protein interaction map for the first cell and the second cell. This map can be used to determine surface cell proteins on the first cell or on the second cell that are likely to contribute to a biological response as a result of an interaction between the first cell and the second cell. These surface cell proteins can be further categorized into surface cell proteins that are likely to contribute to an enhanced biological response, a dampened biological response, or an inhibited biological response. A biological response can therefore be altered by a high affinity reagent to one of these surface cell proteins. Furthermore, a biological response can be altered by multiple high affinity reagents to multiple surface cell proteins. Therefore, a surface cell protein can be identified as a target in which a high affinity reagent can lead to an enhanced, dampened, or inhibited response after an interaction of the first cell with the second cell in comparison to the response after an interaction of the first cell with the second cell in the absence of the high affinity reagent. Furthermore, multiple surface cell proteins can be identified as targets in which the presence of high affinity reagents can lead to an enhanced, dampened, or inhibited response after an interaction of the first cell with the second cell in comparison to the response after an interaction of the first cell with the second cell in the absence of the high affinity reagent.

Phage Display Selection of High Affinity Reagents

[0154] The present disclosure provides methods for phage display techniques to identify high affinity reagents to cell surface proteins. These cell surface proteins can be identified as described above. Phage display is a technique using bacteriophages to express a protein of interest by inserting a gene encoding the protein into the phage coat protein gene. This allows for the protein of interest to be expressed on the phage coat of the phage encoding the protein gene, thus keeping a connection between the genotype and phenotype of the protein. These

bacteriophages expressing a protein can then be screened for the ability of the expressed proteins that bind to other proteins, in the other proteins can be the proteins of interest. Multiple rounds of panning can be conducted to identify proteins with desirable properties. For example, filamentous phage displaying different proteins can be screened for binding to a target (e.g., cell surface protein identified as described above) to identify a protein that binds to the target with a high affinity (i.e. , identifying a protein displayed by the phage as a high affinity reagent). A high affinity reagent can be an antibody, single chain variable region (scFv), or any fragment thereof capable of binding to a protein of interest. For example, filamentous phage displaying different antibodies can be screened against a cell surface protein identified as described above to identify an antibody that can bind with very high affinity to the cell surface protein. High affinity can be a K_d of the high affinity reagent that is less than 10 nM, less than 1 nM, or less than 0.1 nM

[0155] Phage display techniques can be carried out using antibody fragments, such as the scFv domain, the Fab domain, or the Fv domain. Targets can be coated onto the bottom of microwell plates and antibody phage libraries can be incubated in each well. Multiple rounds of panning can be conducted, modulating stringency by varying parameters such as incubation time, incubation temperature, and multiple rounds of washing. Bound antibody phage can be selected and panning is repeated to identify the high affinity antibodies against cell- surface, membrane bound receptor targets. [0156] Subsequent to each round of panning, antibody phage can be eluted from the microwell plate and used in an enzyme linked immunosorbent assay (ELISA) to determine the level of enrichment occurring in each round of selection. For example, ELISA assays can be used to determine the level of binding of an antibody phage to the target as compared to a negative control, thereby quantifying a binding ratio. In general, binding ratios greater than 2 indicate high affinity antibodies. Phage display techniques of the present disclosure to identify high affinity antibodies or antibody fragments (Fab, scFv or Fv fragments) can be carried out using the techniques set forth by Miersch et al. in Scalable High Throughput Selection From Phage- displayed Synthetic Antibody Libraries (J Vis Exp. 2015; (95): 51492). High affinity antibodies of the present disclosure can have equilibrium dissociation constants (K_d) in the nanomolar or sub-nanomolar range. A high affinity antibody can a IQ for its target that is less than 10 nM, less than 1 nM, or less than 0.1 nM.

[0157] This technique can be used to generate a high affinity reagent that binds to a target as identified by the method above. The high affinity reagent can bind to a cell surface protein. This technique can be used to generate multiple high affinity reagents that bind to multiple targets. For example, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more targets can be identified as described by the methods above, and a high affinity reagent can be generated for each of these targets.

Altering Cellular Interactions with High Affinity Reagents

[0158] Once a high affinity reagent is identified for a cell surface protein, it can be used to alter the interaction between a first cell and second cell. For example, a high affinity reagent can bind to T-cell receptors (TCRs) on a CD4⁺ or CD8⁺ T cell, preventing antigen presenting cells (APCs) from stimulating T cells via major histocompatibility complex (MHC)-peptide. By occupying the TCR and preventing binding and stimulation by MHC-peptide, high affinity reagents can dampen the subsequent immune response. The following are additional non-limiting examples of cells and cellular interactions that are of interest to target with the compositions and methods of the present disclosure.

[0159] CD4+/MHC II. CD4⁺ T cells can drive the helper T cell response and function to aid CD8+ in cytolytic activity and aid in the maturation of B cells to plasma cells. Naive T cells are matured into CD4⁺ T cells by a number of different receptor/ligand interactions with antigen presenting cells (APCs), e.g. dendritic cells (DCs), macrophages, epithelial cells, B cells, monocytes, as well as cytokine stimulation. During T cell maturation, MHC II proteins on APCs presenting antigen-derived peptide fragments can bind the CD4 TCR to stimulate naive T cells. The present disclosure provides compositions, tools, and methods for identifying the sequence of the CD4 TCR using molecular barcoding techniques, as described below, and developing high affinity antibodies against the TCR using phage display techniques, as described below. High affinity antibodies against the CD4 TCR can bind to the Dl domain, the D2 domain, the D3 domain, the D4 domain, or any combination thereof. Blocking the CD4 TCR with a high affinity antibody of the present disclosure can result in dampening of the immune response by preventing binding of MHC-peptide expressing APCs.

[0160] CD8⁺/MHC I. CD8⁺ T cells, or cytotoxic T lymphocytes (CTLs), can function to eliminate infected cells presenting non-host antigenic peptides on MHC I. CD8+ T cells can eliminate infected cells through perforin/granzyme mediated cytolysis. Naive T cells are matured into CD8+ T cells by a number of different receptor/ligand interactions with antigen presenting cells (APCs), e.g. dendritic cells (DCs), macrophages, epithelial cells, B cells, monocytes, as well as cytokine stimulation. During T cell maturation, MHC I proteins on APCs presenting antigen- derived peptide fragments can bind the CD8 TCR to stimulate naive T cells. The present disclosure provides compositions, tools, and methods for identifying the sequence of the CD8 TCR using molecular barcoding techniques, as described below, and developing high affinity antibodies against the TCR using phage display techniques, as described below. High affinity antibodies against the CD8 TCR can bind to CD8aP TCRs (expressed by most CD8 T cells) or CD8aa TCRs (expressed by a subset of CD8 T cells, such as intraepithelial T lymphocytes located in the gut epithelium). Blocking the CD8 TCR with a high affinity antibody of the present disclosure can result in dampening of the immune response by preventing binding of MHC-peptide expressing APCs.

[0161] γ:δ T cells. Another subset of T cells express the γ:δ TCR, which can directly bind antigen, can be bound by an MHC class lb molecule, or can be bound by a variety of other ligands. These T cells can be present in the tissues of the lymphoid system and can also be expressed by a subset of intraepithelial lymphocytes. The present disclosure provides

compositions, tools, and methods for identifying the sequence of the γ:δ TCR using molecular barcoding techniques, as described below, and developing high affinity antibodies against the TCR using phage display techniques, as described below. High affinity antibodies against the γ:δ TCR can bind to the receptors, thereby blocking binding by cognate ligands or soluble antigens and dampening the immune response.

[0162] CD28/CD80 or CD28/CD86. CD28 is a costimulatory molecule expressed at the surface of T cells. During T cell maturation, T cells can require multiple signals from APCs. For example, maturation can involve binding of MHC-peptide to the TCR (Signal 1) and binding of APC surface expressed proteins CD80 and CD86 to CD28 (Signal 2). Signal 2, co-stimulation, can be necessary to prevent tolerance in T cells only receiving Signal 1. CD28 is expressed by naive T cells and binds to either CD80 (B7.1) or CD86 (B7.2) expressed by APCs. The result of co-stimulation via CD28 can lead to downstream signaling (e.g. , phosphorylation or recruitment) that can result in secretion of cytokines, such as IL-2, that can promote differentiation and proliferation of T cells. The present disclosure provides compositions, tools, and methods for identifying the sequence of CD28 using molecular barcoding techniques, as described below, and developing high affinity antibodies against CD28 using phage display techniques, as described below. Blocking the CD28 co-stimulatory molecule with a high affinity antibody of the present disclosure can result in dampening of the immune response by preventing binding of CD80 or CD86 expressing APCs.

[0163] CTLA-4/CD80 or CTLA-4/CD86. The cytotoxic T lymphocyte associated (CTLA-4) molecule is a molecule expressed by lymphocytes that is related in structure to CD28 but, in contrast, can promote inhibition of the immune response. The result of co-stimulation of CTLA-4 expressing T cells via CD80 (B7.1) or CD86 (B7.2) expressing APCs, can be a dampening of the immune response. For example, decreased co-stimulation of CTLA-4 can lead to reduced T cell proliferation. Thus, blocking the CTLA-4 inhibitory molecule can be a strategy to enhance the immune response. The present disclosure provides compositions, tools, and methods for identifying the sequence of CTLA-4 using molecular barcoding techniques, as described below, and developing high affinity antibodies against CTLA-4 using phage display techniques, as described below. Blocking the CTLA-4 co-stimulatory molecule with a high affinity antibody of the present disclosure can result in dampening of the immune response by preventing binding of CD80 or CD86 expressing APCs. The compositions, tools, and methods of the present disclosure can also be used to develop high affinity reagents against other inhibitory molecules expressed by lymphocytes, such as programmed death- 1 (PD- 1) and B and T lymphocyte attenuator (BTLA). Enhancing the immune response by binding and blocking CTLA-4, PD- 1, or BTLA can be a method of promoting T cell activation against tumor cells in cancer immunotherapies.

[0164] LFA-l/ICAM-1 or LFA-l/ICAM-2. Cellular interactions that promote adhesion and prolong the length of time that two cells interact are of interest to the present disclosure. LFA- 1 expressed by T cells can transiently bind ICAM- 1 or ICAM-2 expressed by APCs, to bring the cells into close contact to promote formation of the immunological synapse and prolong the subsequent interaction time, thereby increasing the opportunity for MHC-peptide complexes and co-stimulatory ligands to bind the TCR and/or CD28 and active T cells. After binding of MHC- peptide complexes to the TCR, the affinity of LFA- 1 for ICAM- 1 or ICAM-2 can increase. The present disclosure provides compositions, tools, and methods for identifying the sequence of LFA- 1 using molecular barcoding techniques, as described below, and developing high affinity antibodies against LFA- 1 using phage display techniques, as described below. Blocking the LFA- 1 adhesion molecule with a high affinity antibody of the present disclosure can result in dampening of the immune response by preventing binding of ICAM-1 or ICAM-2 expressing APCs.

[0165] CD2-CD58. Another set of adhesion molecules of interest include CD2 expressed by T cells and its interaction with CD58 expressing APCs. This interaction can function similarly to LFA-1/ICAM interactions, as described above. CD2 binding to CD58 (LFA-3) can result in bringing cells into close contact to promote formation of the immunological synapse and prolong subsequent interaction time, thereby increasing the opportunity for MHC-peptide complexes and co-stimulatory ligands to bind the TCR and/or CD28 and active T cells. The present disclosure provides compositions, tools, and methods for identifying the sequence of CD2 using molecular barcoding techniques, as described below, and developing high affinity antibodies against CD2 using phage display techniques, as described below. Blocking the CD2 adhesion molecule with a high affinity antibody of the present disclosure can result in dampening of the immune response by preventing binding of CD58 APCs.

[0166] Cytokine Receptors. Naive and activated T cells can express various cytokine receptors, which when bound by secreted cytokines can function to drive T cell proliferation and differentiation. For example, T cells can express receptors for IL-6, IL-12, IL-23, IL-4, and IFN- γ. Upon activation, these receptors can have increased affinity for secreted cytokines and binding of subsets of these receptors can drive differentiation of lymphocytes. The present disclosure provides compositions, tools, and methods for identifying the sequence of any cytokine receptor using molecular barcoding techniques, as described below, and developing high affinity antibodies against the cytokine receptor of interest using phage display techniques, as described below. Blocking the cytokine receptor molecule with a high affinity antibody of the present disclosure can result in dampening of the immune response by preventing binding of cytokines secreted by paracrine or autocrine pathways.

[0167] Furthermore, multiple high affinity reagents can be identified for multiple cell surface proteins that are present on a first cell or a second cell. These high affinity reagents can be combined into a composition that can be used to alter the interaction between the first cell and the second cell. For example, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more targets can be identified as described by the methods above. High affinity reagents can be generated for each target, which can be combined into a single composition. This composition can be used to alter the interaction between the first cell and the second cell. This composition can be more effective at altering the interaction due to the targeting of multiple cell surface proteins. Additionally, a specific combination of high affinity reagents can be used to fine tune the degree of altering of the interaction in order to achieve a desired threshold of a biological response. For example, a specific combination of targets can be chosen from the cell surface protein interaction map of a first cell and a second cell. The combination of targets can include some targets that when bound by a high affinity reagent will result in enhanced cell signaling and some targets that when bound by a high affinity reagent will result in dampened cell signaling. The combination of these high affinity reagents can then lead to an intermediate level of altering of the cell signaling that occurs as a result of the interaction between the first cell and the second cell, and thus an intermediate level of a biological response.

[0168] High affinity reagents identified using the methods of the present disclosure can be used to alter cell signaling resulting from the interaction of a first cell with a second cell by binding to a cell surface protein. This binding can prevent binding and stimulation by a soluble ligand or a cognate ligand expressed by another cell. High affinity reagents identified herein can include antibodies or antibody fragments, which bind to membrane-bound cell surface receptors. High affinity antibodies can be administered in vitro to study the altering of cellular interactions. High affinity antibodies can also be administered to a subject in need thereof as a therapeutic. A subject in need thereof can have a condition such as a viral or bacterial infection, an autoimmune disease, cancer, or any condition associated with dysregulated cellular interactions.

[0169] The cellular interaction can be between any two cells. For example, the interaction can be between any antigen presenting cell, such as a dendritic cell, a macrophage, an epithelial cell, or a B cell, and a naive T cell. The interaction can also be between a secreted, soluble factor and a membrane-bound cell surface receptor. For example, the interaction can be between a cytokine (e.g. IL-6, IL-12, IL-23, IL-4, and IFN-γ) and the corresponding cytokine receptor.

[0170] High affinity reagents can be administered via different routes of administration including intravenous administration, subcutaneous administration, intramuscular administration, mucosal administration, oral administration, or nasal administration. Interruption of cellular interactions with high affinity reagents of the present disclosure can result in dampening the immune response, enhancing the immune response, or driving T cell responses against tumor cells.

Pharmaceutical Compositions

[0171] The compositions and methods described herein can be considered useful as

pharmaceutical compositions for administration to a subject in need thereof. Pharmaceutical compositions can comprise at least the compositions described herein and one or more pharmaceutically acceptable carriers, diluents, excipients, stabilizers, dispersing agents, suspending agents, and/or thickening agents. A therapeutic composition of a high affinity reagent as described herein can be prepared for storage by mixing a high affinity reagent having the desired degree of purity with a pharmaceutically acceptable carrier, excipient, and/or stabilizer. This formulation can be in the form of a lyophilized formulation or aqueous solution. An acceptable carrier, excipient, and/or stabilizer can be nontoxic to a recipient at the dosage and concentration employed. An acceptable carrier, excipient, and/or stabilizer can be a buffer such as phosphate, citrate, and other organic acids; an antioxidant including ascorbic acid and methionine; a preservative, (e.g. , octadecyldimethylbenzyl ammonium chloride; hexamethonium chloride; benzalkonium chloride, benzethonium chloride; phenol, butyl or benzyl alcohol; alkyl parabens such as methyl or propyl paraben; catechol; resorcinol; cyclohexanol; 3-pentanol; and m-cresol); a low-molecular- weight (e.g. , less than about 10 residues) polypeptide; a protein, such as serum albumin, gelatin, or immunoglobulin; a hydrophilic polymer such as

polyvinylpyrrolidone; an amino acid such as glycine, glutamine, asparagine, histidine, arginine, or lysine; a monosaccharide, a disaccharide, and other carbohydrates including glucose, mannose, or dextrin; a chelating agent such as EDTA; a sugar such as sucrose, mannitol, trehalose or sorbitol; a salt-forming counter- ion such as sodium; a metal complex (e.g. , Zn- protein complexes); and/or a non- ionic surfactant such as TWEEN™, PLURONICS™ or polyethylene glycol (PEG).

[0172] High affinity reagents can be antibodies. The antibodies can be in a pharmaceutical composition. A pharmaceutical composition of antibodies can be lyophilized (See, e.g., U.S. Pat. No. 6,267,958). Antibody compositions can be aqueous antibody (See, e.g., U.S. Pat. No.

6, 171,586 and WO06/044908).

[0173] The composition herein can also contain more than one active compound as necessary for the particular indication being treated. The active compounds can have complementary activities that do not adversely affect each other. For example, the composition can comprise a

chemotherapeutic agent, cytotoxic agent, cytokine, growth-inhibitory agent, anti-hormonal agent, anti-angiogenic agent, and/or cardioprotectant. Such molecules can be present in combination in amounts that are effective for the purpose intended. Active ingredients can be entrapped in microcapsules (e.g., hydro xymethylcellulose or gelatin-microcapsules and poly- (methylmethacylate). Active ingredients can be entrapped in microcapsules in colloidal drug delivery systems (e.g., liposomes, albumin microspheres, microemulsions, nano-particles and nanocapsules) by coacervation techniques or by interfacial polymerization (e.g.,

hydro xymethylcellulose or gelatin microcapsules and poly-(methylmethacylate) microcapsules, respectively), in colloidal drug-delivery systems (e.g., liposomes, albumin microspheres, microemulsions, nano-particles and nanocapsules) or in macroemulsions.

[0174] The formulation herein can also contain more than one active ingredient as necessary for the particular indication being treated (e.g., cancer). [0175] Sustained-release preparations can be prepared. Examples of sustained-release preparations can include semipermeable matrices of solid hydrophobic polymers that can contain the antibody, in which the matrices can be in the form of shaped articles (e.g. , films or microcapsules). Examples of sustained-release matrices include polyesters, hydrogels (e.g. , poly(2-hydroxyethyl-methacrylate), or poly(vinylalcohol)), polylactides, copolymers of L- glutamic acid and γ ethyl-L-glutamate, non-degradable ethylene- vinyl acetate, degradable lactic acid-glycolic acid copolymers such as the LUPRON DEPO™ (i.e. , injectable microspheres composed of lactic acid-glycolic acid copolymer and leuprolide acetate), and poly-D-( - )-3- hydroxybutyric acid.

[0176] The pharmaceutical composition to be used for in vivo administration can be generally sterile (e.g., by filtration through sterile filtration membranes). Sterilization can be accomplished by filtration through sterile filtration.

Methods of Treatment

[0177] A high affinity reagent can be administered to a patient in a therapeutically effective amount (i.e. , an amount that has the desired therapeutic effect). A high affinity reagent can be used in vivo or ex vivo. A high affinity reagent can be administered parenterally. The dose and dosage regimen can depend upon the severity of the diagnosis and the characteristics of the particular high affinity reagent used (e.g., its therapeutic index, the patient, and the patient's history). A high affinity reagent can be administered continuously over a specified period of time. A high affinity reagent can be administered intravenously. A high affinity reagent can be administered cutaneously. A high affinity reagent can be administered subcutaneously. A high affinity reagent can be administered intraperitoneally. A high affinity reagent can be

administered at the site of a tumor or affliction. The administration can be made during the course of adjunct therapy such as combined cycles of radiation, chemotherapeutic treatment, or administration of tumor necrosis factor, interferon or other cyto-protective or immunomodulatory agent.

[0178] For parenteral administration, the high affinity reagent can be formulated in a unit dosage injectable form (e.g. , letter solution, suspension, emulsion) in association with a pharmaceutically acceptable parenteral vehicle. Such vehicles can be inherently nontoxic and no n- therapeutic. Examples of such vehicles can be water, saline, Ringer' s solution, dextrose solution, and 5% human serum albumin. Nonaqueous vehicles such as fixed oils and ethyl oleate can also be used. Liposomes can be used as carriers. The vehicle can contain minor amounts of additives such as substances that enhance isotonicity and chemical stability (e.g. , buffers and preservatives). [0179] A high affinity reagent pharmaceutical composition can be used in therapy that can be formulated and with dosages that can be established in a fashion consistent with good medical practice taking into account the disorder to be treated, the condition of the individual patient, the site of delivery of the composition, the method of administration and other factors known to practitioners. A high affinity reagent pharmaceutical composition can be prepared according to the description of preparation described herein.

Diseases and Conditions

[0180] The high affinity reagents described herein can be useful for the treatment of a cancer or tumor. In certain embodiments, the cancer comprises breast, heart, lung, small intestine, colon, spleen, kidney, bladder, head, neck, ovarian, prostate, brain, pancreatic, skin, bone, bone marrow, blood, thymus, uterine, testicular and liver tumors. In certain embodiments, tumors which can be treated with the antibodies of the invention comprise adenoma, adenocarcinoma, angiosarcoma, astrocytoma, epithelial carcinoma, germinoma, glioblastoma, glioma, hemangioendothelioma, hemangio sarcoma, hematoma, hepatoblastoma, leukemia, lymphoma, medulloblastoma, melanoma, neuroblastoma, osteosarcoma, retinoblastoma, rhabdomyosarcoma, sarcoma and/or teratoma. In certain embodiments, the tumor/cancer is selected from the group of acral lentiginous melanoma, actinic keratosis, adenocarcinoma, adenoid cystic carcinoma, adenomas, adenosarcoma, adenosquamous carcinoma, astrocytic tumors, Bartholin gland carcinoma, basal cell carcinoma, bronchial gland carcinoma, capillary carcinoid, carcinoma, carcinosarcoma, cholangiocarcinoma, chondrosarcoma, cystadenoma, endodermal sinus tumor, endometrial hyperplasia, endometrial stromal sarcoma, endometrioid adenocarcinoma, ependymal sarcoma, Swing's sarcoma, focal nodular hyperplasia, gastronoma, germ line tumors, glioblastoma, glucagonoma, hemangioblastoma, hemangioendothelioma, hemangioma, hepatic adenoma, hepatic adenomatosis, hepatocellular carcinoma, insulinite, intraepithelial neoplasia,

intraepithelial squamous cell neoplasia, invasive squamous cell carcinoma, large cell carcinoma, liposarcoma, lung carcinoma, lymphoblastic leukemia, lymphocytic leukemia, leiomyosarcoma, melanoma, malignant melanoma, malignant mesothelial tumor, nerve sheath tumor,

medulloblastoma, medulloepithelioma, mesothelioma, mucoepidermoid carcinoma, myeloid leukemia, multiple myeloma, neuroblastoma, neuroepithelial adenocarcinoma, nodular melanoma, osteosarcoma, ovarian carcinoma, papillary serous adenocarcinoma, pituitary tumors, plasmacytoma, pseudosarcoma, prostate carcinoma, pulmonary blastoma, renal cell carcinoma, retinoblastoma, rhabdomyosarcoma, sarcoma, serous carcinoma, squamous cell carcinoma, small cell carcinoma, soft tissue carcinoma, somatostatin secreting tumor, squamous carcinoma, squamous cell carcinoma, undifferentiated carcinoma, uveal melanoma, verrucous carcinoma, vagina/vulva carcinoma, vipoma, and Wilm's tumor. In certain embodiments, the tumor/cancer to be treated with one or more antibodies of the invention comprise brain cancer, head and neck cancer, colorectal carcinoma, acute myeloid leukemia, pre-B-cell acute lymphoblastic leukemia, bladder cancer, astrocytoma, preferably grade II, III or IV astrocytoma, glioblastoma,

glioblastoma multiforme, small cell cancer, and non-small cell cancer, preferably non-small cell lung cancer, lung adenocarcinoma, metastatic melanoma, androgen-independent metastatic prostate cancer, androgen-dependent metastatic prostate cancer, prostate adenocarcinoma, and breast cancer, preferably breast ductal cancer, and/or breast carcinoma.

[0181] The high affinity reagents described herein can be useful for the treatment of autoimmune diseases. An autoimmune disease can be Addison's disease, Agammaglobulinemia, Alopecia areata, Amyloidosis, Ankylosing spondylitis, Anti-GBM/Anti-TBM nephritis, Antiphospholipid syndrome (APS), Autoimmune hepatitis, Autoimmune inner ear disease (AIED), Axonal & neuronal neuropathy (AMAN), Behcet's disease, Bullous pemphigoid, Castleman disease (CD), Celiac disease, Chagas disease, Chronic inflammatory demyelinating polyneuropathy (CIDP), Chronic recurrent multifocal osteomyelitis (CRMO), Churg-Strauss, Cicatricial

pemphigoid/benign mucosal pemphigoid, Cogan's syndrome, Cold agglutinin disease,

Congenital heart block, Coxsackie myocarditis, CREST syndrome, Crohn's disease, Dermatitis herpetiformis, Dermatomyositis, Devic's disease (neuromyelitis optica), Discoid lupus,

Dressier' s syndrome, Endometriosis, Eosinophilic esophagitis (EoE), Eosinophilic fasciitis, Erythema nodosum, Essential mixed cryoglobulinemia, Evans syndrome, Fibromyalgia,

Fibrosing alveolitis, Giant cell arteritis (temporal arteritis), Giant cell myocarditis,

Glomerulonephritis, Goodpasture's syndrome, Granulomatosis with Polyangiitis, Graves' disease, Guillain-Barre syndrome, Hashimoto's thyroiditis, Hemolytic anemia, Henoch- Schonlein purpura (HSP), Herpes gestationis or pemphigoid gestationis (PG),

Hypogammaglobulinemia, IgA Nephropathy, IgG4-related sclerosing disease, Inclusion body myositis (IBM), Interstitial cystitis (IC), Juvenile arthritis, Juvenile diabetes (Type 1 diabetes), Juvenile myositis (JM), Kawasaki disease, Lambert-Eaton syndrome, Leukocytoclastic vasculitis, Lichen planus, Lichen sclerosus, Ligneous conjunctivitis, Linear IgA disease (LAD), Lupus, Lyme disease chronic, Meniere's disease, Microscopic polyangiitis (MP A), Mixed connective tissue disease (MCTD), Mooren's ulcer, Mucha-Habermann disease, Multiple sclerosis (MS), Myasthenia gravis, Myositis, Narcolepsy, Neuromyelitis optica, Neutropenia, Ocular cicatricial pemphigoid, Optic neuritis, Palindromic rheumatism (PR), PANDAS (Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcus), Paraneoplastic cerebellar

degeneration (PCD), Paroxysmal nocturnal hemoglobinuria (PNH), Parry Romberg syndrome, Pars planitis (peripheral uveitis), Parsonnage-Turner syndrome, Pemphigus, Peripheral neuropathy, Perivenous encephalomyelitis, Pernicious anemia (PA), POEMS syndrome

(polyneuropathy, organomegaly, endocrinopathy, monoclonal gammopathy, skin changes), Polyarteritis nodosa, Polymyalgia rheumatica, Polymyositis, Postmyocardial infarction syndrome, Postpericardiotomy syndrome, Primary biliary cirrhosis, Primary sclerosing cholangitis, Progesterone dermatitis, Psoriasis, Psoriatic arthritis, Pure red cell aplasia (PRCA), Pyoderma gangrenosum, Raynaud's phenomenon, Reactive Arthritis, Reflex sympathetic dystrophy, Reiter's syndrome, Relapsing polychondritis, Restless legs syndrome (RLS), Retroperitoneal fibrosis, Rheumatic fever, Rheumatoid arthritis (RA), Sarcoidosis, Schmidt syndrome, Scleritis, Scleroderma, Sjogren's syndrome, Sperm & testicular autoimmunity, Stiff person syndrome (SPS), Subacute bacterial endocarditis (SBE), Susac's syndrome, Sympathetic ophthalmia (SO), Takayasu's arteritis, Temporal arteritis/Giant cell arteritis, Thrombocytopenic purpura (TTP), Tolosa-Hunt syndrome (THS), Transverse myelitis, Type 1 diabetes, Ulcerative colitis (UC), Undifferentiated connective tissue disease (UCTD), Uveitis, Vasculitis, Vitiligo, or Wegener's granulomatosis (now termed Granulomatosis with Polyangiitis (GPA)).

[0182] The high affinity reagents described herein can be useful for the treatment of diseases involving dysfunctional cell interactions.

[0183] The following examples are included to further describe some aspects of the present disclosure, and should not be used to limit the scope of the invention.

EXAMPLE 1

Development and Use of High Affinity Antibodies to Alter Cellular Interactions

[0184] This example describes the development and use of high affinity antibodies to interrupt cellular interactions. A schematic of the development and use of high affinity antibodies to alter cellular interactions is shown in FIG. 1. A membrane-bound, cell surface receptor of interest is identified and sequenced using the molecular barcoding techniques of the present disclosure. Soluble cell surface receptor is immobilized in cell culture plates. Antibody phage libraries are generated and screened against immobilized cell surface receptors using the methods of the present disclosure. Multiple rounds of panning are carried out by varying incubation time, incubation temperature, and number of washes. Enrichment of antibody is determined via ELISA of eluted antibody-phage. Phage display techniques identify high affinity antibodies with a dissociation constant in the nanomolar or sub-nanomolar range.

[0185] High affinity antibodies are administered in vitro to evaluate altering of cellular interactions. High affinity antibodies are administered to a subject in need thereof. The subject has a condition, such as a viral infection, a bacterial infection, or cancer. The high affinity antibodies bind the cell surface receptor of interest, thereby preventing binding of soluble factors or cognate receptors on other cells. Altering of cellular interactions leads to dampening of the immune response, enhancement of the immune response, or enhancement of anti-tumor cell T cell responses.

EXAMPLE 2

Use of High Affinity Antibodies to Treat Cancer

[0186] This example describes the development and use of high affinity antibodies to interrupt cellular interactions for a subject with cancer. A tumor sample is resected from a subject with cancer. The gene expression profile of single cells from the tumor sample is produced using molecular barcoding and high-throughput sequencing. The cell type of the single cells is identified, e.g., cancer cell, T cell, dendritic cell, etc. based on their gene expression profile. The cell surface protein gene expression profile is generated for each single cell. For cell types that interact, cell surface protein interaction maps are generated, and cell surface proteins are identified as targets for producing high affinity reagents that inhibit tumor cell growth. Phage display is used to generate high affinity reagents for these target cell surface proteins. These high affinity reagents are combined into a single composition and administered to the subject with cancer. The growth of the cancer in the subject is inhibited by the high affinity reagent composition.

[0187] While preferred embodiments of the present invention have been shown and described herein, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the invention. It should be understood that various alternatives to the embodiments of the invention described herein can be employed in practicing the invention. It is intended that the following claims define the scope of the invention and that methods and structures within the scope of these claims and their equivalents be covered thereby.

Claims

CLAIMS WHAT IS CLAIMED IS:

1. A method of producing a high affinity reagent to a target comprising:

a) identifying a target involved in an interaction of a first cell with a second cell by comparing a first cell surface protein repertoire of the first cell with a second cell surface protein repertoire of the second cell; and

b) generating a high affinity reagent that binds the target.

2. A method of altering an interaction between a first cell and a second cell, the method comprising:

a) identifying a target involved in an interaction of a first cell with a second cell by comparing a first cell surface protein repertoire of the first cell with a second cell surface protein repertoire of the second cell;

b) generating a high affinity reagent against the target; and

c) administering the high affinity reagent to the first cell and the second cell.

3. The method of claims 1-2, wherein the first cell surface protein repertoire and the second cell surface protein repertoire are identified by molecular barcoding and high-throughput sequencing.

4. The method of claims 1-3, wherein a cell surface protein in the first cell surface protein repertoire is identified by a membrane signal sequence.

5. The method of claims 1-4, wherein a cell surface protein in the second cell surface protein repertoire is identified by a membrane signal sequence.

6. The method of claims 1-5, wherein the high affinity reagent is generated using phage display.

7. The method of claims 1-6, wherein the high affinity reagent is an antibody.

8. The method of claims 1-7, wherein the high affinity reagent is a small molecule.

9. The method of claims 1-8, wherein a IQ of the high affinity reagent is less than 10 nM, less than 1 nM, or less than 0.1 nM.

10. The method of claims 1-9, wherein the target is a receptor.

11. The method of claims 1-9, wherein the target is a ligand of a receptor.

12. The method of claims 1-11, wherein the target is expressed on the first cell.

13. The method of claims 1-11, wherein the target is expressed on the second cell.

14. The method of claims 1-9, wherein the target is secreted by the first cell or the second cell.

15. The method of claims 1-14, wherein the first cell is an immune cell or a tumor cell.

16. The method of claims 1-15, wherein the second cell is an immune cell or a tumor cell.

17. The method of claims 1-16, wherein the first cell is a T cell.

18. The method of claims 1-17, wherein the second cell is an antigen presenting cell.

19. The method of claims 2-18, wherein the altering of the interaction comprises enhancing a biological response.

20. The method of claims 2-18, wherein the altering of the interaction comprises inhibiting a biological response.

21. The method of claims 2-18, wherein the altering of the interaction comprises dampening a biological response.

22. The method of claims 2-18, wherein a biological response is an immune response, an anti-tumor response, cell proliferation, or cell apoptosis.

23. A composition comprising a high affinity reagent, wherein the high affinity reagent is produced by the methods of claims 1, and 3-18.

24. The composition of claim 23 further comprising 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or more high affinity reagents, wherein each high affinity reagent binds to a different target.

25. A pharmaceutical composition comprising the composition of claims 23-24 and a pharmaceutically acceptable carrier.

26. A method of treating a subject in need thereof, comprising administering a therapeutic dose of the composition of claims 23-24 or the pharmaceutical composition of claim 25.

27. The method of claim 26, wherein the subject has cancer, an autoimmune disease, or dysfunctional cell signaling.

28. The method of claims 26-27, wherein the composition or pharmaceutical composition is a administered intravenously, cutaneously, subcutaneously, or injected at a site of affliction.