WO2020254672A1

WO2020254672A1 - Spatial characterisation of target structures in a sample

Info

Publication number: WO2020254672A1
Application number: PCT/EP2020/067260
Authority: WO
Inventors: Jörn GLÖKLER; Steffen Hennig
Original assignee: Therycell Gmbh
Priority date: 2019-06-19
Filing date: 2020-06-19
Publication date: 2020-12-24

Abstract

The present invention relates to a method for determining a location of a target structure in a sample via nucleic-acid based barcoding.

Description

Spatial characterisation of target structures in a sample

This application claims the benefit of European Application EP19181359.1 filed 19 June 2019, incorporated herein by reference.

Background

Analytic samples can comprise a complex composition with respect to the type and number of molecules, and their spatial distribution. So far, the type and number of molecules in a sample can be characterised by a multitude of methods comprising mass-spectrometry, ELISA or electrophoresis. Especially nucleic acids can be easily analysed and enumerated by RT-PCR and sequencing. However, the distribution of molecules or functional groups in a solid or semi-solid sample is more difficult to be analysed and enumerated. This also holds true for disperse samples comprising cells, particles or complexes, each with a certain characteristic composition of molecules.

Analytical samples may comprise tissues, cells, viral particles, biofilms, natural and artificial particles or surfaces. A classical way to analyse planar samples is to employ a panel of specific antibodies and to probe the surface for specific binding. Three-dimensional samples can be transformed into a series of planar samples by means such as microtomy typically used for generating consecutive thin sections from tissue samples.

For instance, immunohistochemistry is a versatile method to detect the presence of target molecules in thin sections of clinical samples. Nucleic acids can be analysed by in situ hybridisation using labelled complementary nucleic acid probes instead of antibodies.

However, the complexity of such sample analyses is limited due to the number of different labels available to distinguish probes specifically. Fluorescent labels can be used to distinguish not much more than 4 to 8 probes due to intrinsic spectral overlap. Other means such as isotope labels can be employed to differentiate probes by mass-spectrometry. Yet the complexity of labels distinguishable by mass-spectrometry is not likely to exceed 40. Thus, only general presence of antigens, but not their detailed sequence information can be detected in a low complexity. New technologies allow identification and partial sequencing of cells in situ using thin sections such as formalin-fixed paraffin-embedded tissue (FFPE). For instance, Nanostring Technologies has developed methods for multiplexed detection and quantification of protein and gene expression in thin sections. NanoString’s GeoMx™ Digital Spatial Profiler combines standard immunofluorescence techniques with digital optical barcoding technology to perform highly multiplexed, spatially resolved profiling experiments. Yet the spatial isolation of barcodes relies on recovery by aspiration and parallel storage in microtiter plates, thus severely limiting the number of features in a sample to be analysed in parallel. Nanostring Technologies has also developed a method for enzyme-free sequencing method termed Hyb & Seq™. However, the Hyb & Seq™ technology cannot be directly applied to thin sections for in situ sequencing, but relies on a prior isolation and purification step of the nucleic acids to be sequenced. Yet another method is based on fluorescent in situ sequencing (FISSEQ) which allows the in situ identification of gene products and recent adaptations thereof. The current drawback of FISSEQ is the high level of enzymatic steps such as ligation, reverse transcription and rolling circle amplifications to be conducted before the actual sequencing steps. Due to the high background fluorescence of tissues and other natural samples, it is challenging to sequence molecules in high density and longer contigs by a process that requires less than several weeks.

Another platform directed at identifying mRNA in spatial resolution of planar samples is the so-called "spatial transcriptomics" method (ST) now commercialised by 10x Genomics, Inc. that allows a gridded barcoding of transcripts at a high resolution. However, the pre fabricated arrays with spatially arranged barcodes intrinsically have a limited resolution, especially due to the step in which the mRNAs are transferred from the sample are captured on the array. The inflexible design of the grid and the small size of the barcoded array makes it difficult to directly target the region of interest. In addition, any kind of barrier between the sample and the chip can lead to a transfer loss or at least strong bias with respect to the barcoded mRNA. There is a need for a method that enables a flexible means of identifying and enumerating the distribution any type of molecules on a substantially planar sample in a high resolution and throughput in a rapid fashion.

Summary of the invention

The invention relates to an optically-controlled in situ assembly of spatially-encoded nucleic acid barcodes on probes distributed on a substantially planar sample using an ultrafast crosslinking strategy.

A first aspect of the invention relates to a method for determining a location of a target structure, said method comprising the steps

a. providing a sample comprising the target structure;

b. contacting said sample with a binding molecule,

- wherein said binding molecule is able to specifically bind to said target structure, and

- wherein said binding molecule is coupled to a recorder nucleic acid

molecule, the recorder nucleic acid molecule being characterized by a recorder sequence that comprises a plurality of individual sequence tags (ISTs);

c. contacting said sample with a first barcode nucleic acid molecule that specifically hybridizes to a first 1ST comprised in said recorder sequence d. exposing a first area within said sample to electromagnetic radiation of a

specific wavelength,

- wherein said first barcode nucleic acid molecule under conditions of

electromagnetic radiation exposure of said specific wavelength is able to introduce a modification to said recorder nucleic acid molecule within said first 1ST if and when the barcode nucleic acid molecule is hybridized to its cognate section of the recorder sequence;

e. in a wash step, exposing the sample to conditions that allow for non-covalently hybridized nucleic acid sequences to separate, thereby removing said first barcode nucleic acid molecule from said recorder nucleic acid molecule;

f. repeating step c to e for each of the plurality of ISTs, thereby exposing a

different unique area of the sample to the electromagnetic radiation of the specific wavelength in each repetition;

g. extracting said recorder nucleic acid molecule from said sample;

h. determining whether, and which of, said ISTs of said recorder nucleic acid

molecule carry said modification, thereby determining the location of said target structure.

A second aspect of the invention relates to a kit comprising a binder molecule as described in aspect 1 , a recorder nucleic acid as described in aspect 1 , and at least one photoreactive barcode nucleic acid molecule as described in aspect 1.

Preferred embodiments are stated in the respective dependent claims and the following description.

Description of the figures

Fig. 1 shows a depiction of binary combinatorial barcode generation by skipping; a)

examples for simple combinatorial barcoding. R1-4 in the first row denote subunit attachment regions indicated by thick lines in an unmodified barcode template. Beneath are two examples wherein the region 1 or region 1 and 3 are modified by barcode subunit attachment and resulting binary value of the barcode is shown below; b) examples for combinatorial barcoding with proofreading. R1-4 denote subunit attachment regions indicated by thick lines in an unmodified barcode template, wherein each region comprises two separate attachment regions for redundancy. Beneath are two examples wherein the region 1 or region 1 and 3 are modified by barcode subunit attachment and resulting binary value of the barcode is shown below. The second example shows while only one of the two redundant regions of R3 is modified, the value can still be attributed as 1.

Fig. 2 shows the depiction of a linear combinatorial barcode generation with distinct ends. The first row shows an example using 4 barcode subunits (denoted S1-4) for each of the different cycles with inter-cycle compatible ends (denoted 1-5 for complementary sequences at overhangs). The encoding regions of the barcode subunits are indicated by thick lines. The second row shows the first attachment step wherein the first subunit is attached to the probe (P). Further attachment cycles lead to the assembly of a complete barcode comprising all 4 barcode subunits for proper spatial encoding of the probe. Abortive products missing a subunit due to incomplete attachment are indicated by crossed lines.

Fig. 3 shows configurations of barcode templates and attachment of subunits a) depicts a linear barcode in the first row and otherwise non-limiting examples for branched barcode configurations. P denotes the end bound or attached to the probe and the thin lines comprise an either direct or indirect linkage via linkers moieties to the attachment regions. The thick bars represent attachment regions (in this example R1-4) to which barcode subunits can be attached b) shows various means by which a barcode subunit (S) may be attached to the cognate attachment region in a barcode template. In the first row, the barcode subunit is substantially complementary to the attachment region (dashed lines indicate potential linkage to further regions). In the second row, the barcode subunit may comprise a central coding region (cr) which can be used for specific barcode identification. In the third row, the barcode subunit comprises a substantially free coding region which can be used for specific barcode identification. For simplicity, examples for redundancy (and proofreading) of attachment regions and/or barcode regions are not shown.

Fig. 4 shows the assembly-dependent editing of barcode template. Non-limiting

examples for sequence editing of the barcode template after assembly a) shows examples of the barcode template after combinatorial assembly with either attached barcode subunit or without thereof (e.g. by“skipping”) b) shows edited barcode templates i) represents the crosslinked complex in which substantially the whole subunit remains attached ii) is the crosslinked complex in which the barcode subunit is substantially removed iii) shows a barcode template in which the crosslinked base is removed, generating an abasic site iv) shows a barcode template in which the crosslinked base is edited from a previous C to U. v) shows a barcode template in which a base is specifically methylated vi) is an essentially unmodified barcode template, depending on a previous barcode subunit attachment or lack thereof.

Fig. 5 Flowchart of the spatial characterisation method for the identification of disease- specific antibodies and cognate antigens.

Fig. 6 Flowchart of the spatial and marker-based characterisation method for the

identification of disease-specific antibodies and cognate antigens.

Fig. 7 shows simplified graphic outline of the general spatial barcoding workflow. As a starting point, 4 ROIs are depicted as rectangles already comprising target structure-bound probes with recorder nucleic acid molecules (designated as x).

The sample is incubated with a first barcode subunit (1) that binds to the 1ST of the recorder molecules and one half of the ROIs is irradiated (indicated by wavy lines) to crosslink the barcode subunit 1 to the recorder molecules. After a washing step, only the crosslinked barcode remains bound at the recorder molecules of the previously irradiated ROIs. The sample is incubated with a second barcode subunit (2) that binds to the ISP of the recorder nucleic acid molecules and a partially overlapping half of the ROIs is irradiated (indicated by wavy lines) to crosslink the barcode subunit 2 to the recorder molecules. After a washing step, only the crosslinked barcode remains bound at the recorder molecules of the previously irradiated ROIs. As a result, target structure-bound probes in each ROI retain a different combination of barcode subunits crosslinked to the recorder nucleic acid molecules. Finally, the recorder nucleic acid molecules are extracted from the sample for analysis.

Fig. 8 shows simplified graphic outline of the single particle spatial barcoding workflow.

As a starting point, 4 ROIs are depicted as rectangles comprising particles (depicted as circles) and target nucleic acids (designated as x) along with probes with recorder nucleic acid molecules (shown as horizontal lines extending from the circles) bound to their cognate target structures in a previous step. The sample is incubated with a first barcode subunit (1) that binds to the 1ST of the recorder molecules and one half of the ROIs is irradiated (indicated by wavy lines) to crosslink the barcode subunit 1 to the recorder molecules. After a washing step, only the crosslinked barcode remains bound at the recorder molecules of the previously irradiated ROIs. The sample is incubated with a second barcode subunit (2) that binds to the ISP of the recorder nucleic acid molecules and a partially overlapping half of the ROIs is irradiated (indicated by wavy lines) to crosslink the barcode subunit 2 to the recorder molecules. After a washing step, only the crosslinked barcode remains bound at the recorder molecules of the previously irradiated ROIs. As a result, target structure-bound probes in each ROI retain a different combination of barcode subunits crosslinked to the recorder nucleic acid molecules. The particles are isolated from the sample comprising the barcoded recorder nucleic acid molecules and transferred singly into compartments (larger adjoining circles) comprising indices (i with a compartment-specific digit). The nucleic acids including the recorder molecules are extracted from the particles and linked to the compartment-specific indices. Finally, the nucleic acid molecules are extracted from the compartments for analysis.

Fig. 9 shows the method for linking spatial information to clonal indices based on

barcode subunits with additional functional features cross-linked to an ISP as a part of a recorder nucleic acid molecule bound to a particle. The coding region comprises a unique barcoding sequence (be) corresponding to the ISP for the barcode nucleic acid molecule and may comprise one (P1) or a second (P2) constant region common to other barcode nucleic acid molecules that can be used for primer binding and amplification and an optional label (F). After sorting of the particle into a compartment, the crosslinking is reversed to increase diffusion efficiency of the barcode subunit for the index linking step. The free barcode subunit can be contacted with a nucleic acid molecule comprising a

complementary sequence (PT) and a compartment-specific index sequence (I) that can be linked to the sequence of the barcoding sequence.

Fig. 10 shows the method for linking spatial information to clonal indices based on

barcode subunits with coding region comprising a unique barcoding sequence (be) corresponding to the cross-linked ISP that forms a part of a recorder nucleic acid molecule bound to a particle. The unique barcoding sequence is hybridised to a complementary oligonucleotide comprising a reverse complementary sequence of the barcode (be’) with at least one additional primer binding site (P1) and an optional primer binding site (P2) before sorting of the particle into a separate compartment. Within the compartment, the complementary oligonucleotide can be detached by a denaturing step and annealed to a a nucleic acid molecule comprising a complementary sequnce (PT) and a compartment-specific index sequence (I) that can be linked to the sequence of the barcoding sequence.

Fig 11 Schematic representation of the method of the invention. Step a: A biological sample is provided. The sample may be composed of cells, virus particles, or other biological material. The sample is fixed in its position. Step b: The sample is contacted with binding molecules, e.g. antibodies. Each binding molecule is coupled to a recorder nucleic acid molecule. The recorder nucleic acid molecule comprises a sequence which is characteristic for the binding molecule and several individual sequence tags (ISTs). Step c: The sample is contacted with a first barcode nucleic acid molecule which is able to hybridize to the first 1ST. Thus, each first 1ST is non-covalently bound to a first barcode nucleic acid molecule.

Step d: A first region of interest is exposed to electromagnetic radiation. The electromagnetic radiation causes a covalent link via a light-sensitive nucleic acid between the first 1ST and the first barcode nucleic acid molecule. This happens only in the first region of interest. In all other regions, the first 1ST is non-covalently bound to the first barcode nucleic acid molecule. Step e (not shown): All non- covalently bound first barcode nucleic acid molecules are washed off. Step f: step c to e are repeated with different barcode nucleic acid molecules and different areas of electromagnetic radiation. This procedure creates recorder nucleic acid molecules, wherein the ISTs are covalently bound to a barcode nucleic acid molecule only if the corresponding region was exposed to electromagnetic radiation. The resulting recorder nucleic acid molecules may be analysed via any type of sequencing.

Fig. 12 Schematic representation of certain embodiments of the“particle method”. For the “particle method”, a sample is provided which has undergone the barcoding method of the invention. This sample is disintegrated and each entity (e.g. a cell or a virus particle) is bound to a particle. Each cell with a bound particle is segregated into one compartment. The particle is decorated with compartment nucleic acid molecules. Each barcode nucleic acid molecule (which is bound to a 1ST of the recorder nucleic acid molecule) is then linked to compartment nucleic acid molecule. Linking may be achieved by hybridization, as shown in Fig. 13. Each compartment nucleic acid molecule comprises a compartment-specific index sequence (which is specific for the particle it was bound to), a unique molecule identifier sequence (UMI, which is unique), and a primer binding site for amplification. Inside the compartment, a PCR is performed of the hybrid nucleic acid molecule composed of (a) the barcode nucleic acid molecule, and (b) the compartment nucleic acid molecule. After PCR, the amplified hybrid nucleic acid molecules may be pooled for sequencing.

Fig. 13 In one embodiment, the link between (a) the barcode nucleic acid molecule, and

(b) the compartment nucleic acid molecule is achieved via hybridization. The barcode nucleic acid molecule may have a poly-A tail and the compartment nucleic acid molecule may have a poly-T sequence at its 5’ end. Detailed description of the invention

Terms and definitions

A“nucleic acid" herein denotes a polymer comprising nucleosides. Said polymer may comprise naturally occurring nucleobases (i.e. adenosine, thymidine, guanosine, cytidine, uridine), nucleobase analogues (i.e. 2-aminopurine, 2-aminoadenosine, 2-thiothymidine, inosine, pyrrolpyrimidine, 3-methyladenosine, 5-methylcytidine, C5-bromouridine, C5- fluorouridine, C5-iodouridine, C5-propynyluridine, C5-propynylcytidine, C5-methylcytidine, 7- deazaadenosine, 7-deazaguanosine, 8-oxoadenosine, 8-oxoguanosine, 0(6)-methylguanine and 2-thiocytidine), chemically modified bases, biologically modifies bases (i.e., methylated bases), intercalated bases, abasic sites, ribose-sugars (RNA), 2'-deoxyribose sugars (DNA), terminal 3'-deoxyribose or 2',3'-dideoxyribose sugar, modified sugars (i.e., 2'-0-methylribose (2'-OMeRNA)), or modified phosphates (i.e., phosphorothioate and 5'-N-phosphoramidites). In addition, the backbone may be modified by synthetic sugars or analogues thereof. Non limiting examples are so-called“locked" (LNA),“unlocked" (UNA), 2‘-flouroribose (FNA), arabinose (ANA), 2’-flouroarabinose (FANA), hexose (HNA), threose (TNA), acyclic threose (aTNA), serinol (SNA), glycerol (GNA) sugar,“spiegelmers" (L-form sugars), triazol-sugars (synthesised by a click-reaction) or a peptidic backbone (PNA) or combinations thereof.

An“oligonucleotide” herein denotes nucleic acids with a length of 4 to 150 nucleotides. Oligonucleotides may comprise modifications such as labels or reactive groups for covalent immobilisation, crosslinking or derivatisation i.e. hydroxyl, phosphatitdyl, sulfonic ester, thiol, alkyne, azide, hydrazide, or EDO. Oligonucleotides may also comprise non-nucleosidic linking moieties such as polyethylene glycol (PEG), which can join subunits or regions of the oligonucleotide. Subunits of an oligonucleotide may also be joined by non-covalent interactions such as base pairing and/or haptens and their binding molecule (such as biotin and avidin) to become functionally linked in an oligonucleotide. Oligonucleotides may hybridise to complementary sequences under suitable conditions. Unblocked 3'-ends of oligonucleotides may serve as primers for polymerases. An oligonucleotide may also adopt a more complicated structure to act as an aptamer to selectively bind molecules that do not comprise a complementary sequence.

In the context of the present specification, a target structure is a molecule in a sample in a certain location of interest. A target structure may be selected from the list, but is not limited to, a protein, a peptide, an oligonucleotide, a carbohydrate, and a lipid. A synonym for target structure in the context of the present specification is a target molecule.

In the context of the present specification, a binding molecule is a molecule which is able to specifically bind to the target structure. A binding molecule may be selected from the list, but is not limited to, an antibody, an antibody-like molecule, an antibody fragment, an aptamer, and an oligonucleotide. A synonym for binding molecule in the context of the present specification is a probe.

In the context of the present specification, a recorder nucleic acid molecule is a single- stranded oligonucleotide comprising several individual sequence tags (ISTs). An 1ST is a unique nucleotide sequence which is able to specifically bind to and hybridize with a barcode nucleic acid molecule. Each 1ST has a reverse-complementary sequence to each barcode nucleic acid molecule. Either the barcode nucleic acid molecule or the recorder nucleic acid molecule carries at least one chemical group which is reactive under radiation exposure. A synonym for recorder nucleic acid molecule is a template. A synonym for barcode nucleic acid molecule is barcode subunit or simply barcode.

A“vinylcarbazole nucleoside” herein denotes a nucleoside comprising a vinylcarbazole group which can be photocrosslinked to an adjacent molecule. Preferably, the vinylcarbazole group is a 3-cyanovinylcarbazole, 3-methoxycarbonylcarbazole, 3-carboxyvinylcarbazole or 3- carbonylamidevinylcarbazole, or pyranocarbazole. Preferably, the vinyl-group of the vinylcarbazole nucleoside is capable to form a covalent bond with a diagonally opposite positioned nucleobase by photoirradiation. The nucleoside group preferably is comprised of a deoxyribose, a ribose, or a threose which can be incorporated into an oligonucleotide.

A first aspect of the invention relates to a method for determining a location of a target structure relative to a sample surface, the method comprising the steps

a. providing a sample comprising the target structure which is fixed in a position on the sample at least for the duration of the process described in the following;

b. contacting the sample with a binding molecule,

- wherein the binding molecule is able to specifically bind to the target structure under conditions prevailing throughout step b to f,

- wherein the binding molecule is coupled to a recorder nucleic acid molecule, which is a nucleic acid oligomer composed of DNA, RNA, nucleotide analogues or any combination thereof;

- wherein the recorder nucleic acid molecule is characterized by a recorder

sequence that comprises a plurality of individual sequence tags (ISTs); c. contacting the sample with a first barcode nucleic acid molecule;

wherein the first barcode nucleic acid molecule specifically hybridizes to a first 1ST comprised in the recorder sequence;

d. exposing a first area within the sample to electromagnetic radiation of a specific

wavelength, wherein the first barcode nucleic acid molecule under conditions of electromagnetic radiation exposure of the specific wavelength is able to introduce a modification to the recorder nucleic acid molecule within the first 1ST; e. in a wash step, exposing the sample to conditions that allow for non-covalently hybridized nucleic acid sequences to separate, thereby removing the first barcode nucleic acid molecule, if this first barcode molecule was not covalently bound as an effect of the electromagnetic radiation to its cognate recorder sequence, from the recorder nucleic acid molecule;

f. repeating step c to e for each of the plurality of ISTs with a corresponding (different) barcode nucleic acid molecule, and exposing a unique area of the sample to the electromagnetic radiation in each repetition, thereby creating an area of light exposure, and possibly, an area of modification of the recorder sequence to which the 1ST specific barcode is associated;

g. extracting the recorder nucleic acid molecule from the sample;

h. determining whether the ISTs of the recorder nucleic acid molecule carry the

modification, thereby determining the location of the target structure within the areas of electromagnetic radiation exposure.

Particularly, the recorder nucleic acid molecule serves as a template to which the barcoding nucleic acid molecules are attached, particularly covalently linked. Each barcoding nucleic acid molecule may be attached one after the other, thus forming a chain of barcoding nucleic acid molecules. Alternatively, the barcoding nucleic acid molecules may be attached at positions that are wider apart.

The skilled person in the art will understand that the wash step e can be omitted in the last round of repetition.

In certain embodiments, the ISTs are non-overlapping.

The wash step e may be performed via heating of the sample to the melting temperature of the IST/barcode nucleic acid molecule complex and/or by adding chemicals that disrupt nucleic acid double-strands.

In certain embodiments, the modification of the recorder nucleic acid molecule in step d is a covalent bond of the recorder nucleic acid molecule to the barcode nucleic acid molecule.

In certain embodiments, a plurality of different binding molecules with corresponding target structures are applied in step b. Each of the binding molecules is coupled to a recorder nucleic acid molecule as characterized above.

In certain embodiments, each recorder nucleic acid molecule comprises a sequence characteristic for the particular binding molecule it is bound to. When analysing the recorder nucleic acid sequence in step h, this sequence characteristic for the binding molecule is detected and thus, each recorder molecule can be traced back to its corresponding binding molecule, which itself is specific for a target structure. Therefore, it is possible to determine the location of the target structure within the unique areas of electromagnetic radiation exposure.

In certain embodiments, the electromagnetic radiation is light of a specific wavelength. In certain embodiments, the electromagnetic radiation is light of a wavelength of 300-600 nm. In certain embodiments, the electromagnetic radiation is light of a wavelength of 310-490 nm. In certain embodiments, the electromagnetic radiation is light of a wavelength of 340-420 nm.

In certain embodiments, the barcode nucleic acid molecule comprises a chemical entity selected from a pyranocarbazole group, a psoralen-coupled nucleoside, a diazirine-based nucleoside analogue, an aryl-azide-based nucleoside analogue, a 5-iodouracil, a 6- thioguanine, a 4-thiothymidine, a coumarin-modified thymidine, a 3- methoxycarbonylcarbazole nucleoside, a 3-carboxyvinylcarbazole nucleoside, a 3- carbonylamidevinylcarbazole nucleoside, and a 3-cyanovinylcarbazole nucleoside. In certain embodiments, the chemical entity is a nitrophenyl azide. In certain embodiments, the chemical entity is a vinylcarbazole nucleoside.

All the described chemical entities are able to cross-link with a nucleic acid base in close proximity under electromagnetic radiation exposure of a certain wavelength, and a suitable range of wavelengths can be determined by the skilled person without difficulty. Under electromagnetic radiation exposure, interstrand-crosslinks between the recorder nucleic acid molecule and the barcode nucleic acid molecule are formed.

The chemical entity provided for cross-linking the barcode molecule to its cognate recorder sequence is encorporated into the barcode molecule in a way so as to not interfere with the hybridization process. In particularly preferred embodiments, the chemical entity is provided in the form of a nucleoside analogue.

In certain embodiments, electromagnetic radiation exposure in step d is performed for < 60 sec. In certain embodiments, electromagnetic radiation exposure in step d is performed for £ 10 sec.

The electromagnetic radiation is light ranging from UV to IR, depending on the chemical entity comprised in the barcode nucleic acid molecule. In certain embodiments, different barcode modifications requiring different wavelengths are employed in subsequent steps in order to reduce false signals arising from lack of washing efficiency.

In certain embodiments, an enzyme is employed after step f or in step h to modify the 1ST when the 1ST is bound to a barcode nucleic acid molecule. In certain embodiments, the enzyme is a methyltransferase. In certain embodiments, the enzyme is a deaminase. These enzymes are able to discriminate between double-stranded and single-stranded nucleic acid strands. Only the double-stranded nucleic acid strands are modified, thereby the information a bound barcode nucleic acid molecule is transferred to the recorder nucleic acid molecule.

In certain embodiments, the barcode nucleic acid molecule is degraded in step h except for a cross-linked subunit of the barcode nucleic acid molecule. In certain embodiments, the barcode nucleic acid molecule is degraded in step h via enzymatic, chemical or

photochemical degradation leaving behind a covalent modification of each 1ST which was bound to a barcode nucleic acid molecule. In certain embodiments, the barcode nucleic acid molecule is composed of RNA nucleotides which are selectively cleaved. In certain embodiments, the barcode nucleic acid molecule has a phosphorothioate backbone which is selectively broken. In certain embodiments, DNA repair enzymes are used to deglycosylate cross-linked bases between the barcode nucleic acid molecule and the 1ST. The remaining part of the barcode nucleic acid molecule after degradation can be detected by methods described below.

In certain embodiments, nanopore-based sequencing, Single Molecule, Real-Time sequencing, sequencing by hybridization, or the sequencing by expansion method is employed in step h to sequence the recorder nucleic acid molecule. In certain embodiments, sequencing by hybridization is employed in step h to sequence the recorder nucleic acid molecule.

Methods for sequencing useful in the context of the present invention are described in the following patent documents, all of which are incorporated herein by reference:

WO2017201073, WO2018094385, WO2018138237, US2018142286, WO2019069371 ; US10385382B2; US9175338; US 7939259; US2010297644; US2017037456;

US2017159115; US2017314062; US2018044725; US2018087103; US2018334729;

US2019011424; US2019078075; US2019330254; US2020002754; US2020032247.

In certain embodiments, the binding molecule is an antibody or antibody fragment, an aptamer or an antibody-like molecule and the target structure comprises an epitope to which the binding molecule can specifically bind. An example of such binding-target pair is an extracellular or intracellular protein and an antibody or antibody fragment specific for this protein.

In certain embodiments, the binding molecule is a nucleic acid molecule and the target structure comprises a nucleic acid sequence that is able to hybridize with the binding molecule. Such binding-target pair may allow the localisation of viral or cellular mRNA in a sample.

In certain embodiments, the sample is a cell culture, a tissue slide, a viral particle, a biofilm, or a natural or artificial particle or surface. In certain embodiments, each barcode nucleic acid molecule comprises at least two of the chemical entities capable of cross-linking the barcode molecule to the recorder. Using several cross-linking chemical entities in each barcode nucleic acid molecule accounts for the non-complete cross-linking efficacy. Even if one chemical entity does not cross-link under electromagnetic radiation exposure, the information is not lost due to the other chemical entity. This mechanism may be called proof-reading.

In certain embodiments, step g comprises the following steps: i. segregating said sample into a plurality of compartments, wherein each

compartment comprises a plurality of compartment nucleic acid molecules, wherein each compartment nucleic acid molecule comprises:

a compartment-specific index sequence, wherein all compartment-specific index sequences of one compartment have the same sequence, and each compartment-specific index sequence of one compartment differs from any compartment-specific index sequence of another compartment;

optionally, a unique molecule identifier sequence (UMI), wherein each UMI differs from any other UMI of the same particle; and

optionally, a primer binding sequence, wherein said primer binding sequence is able to hybridize to a PCR primer under annealing conditions;

ii. linking a compartment nucleic acid molecule to each barcode nucleic acid molecule to yield a plurality of hybrid nucleic acid molecules;

iii. optionally, amplifying said plurality of hybrid nucleic acid molecules;

iv. pooling said plurality of hybrid nucleic acid molecules of all compartments.

In certain embodiments, the pooled plurality of hybrid nucleic acid molecules is subsequently sequenced.

The compartment-specific index sequence allows to determine in which compartment any barcode nucleic acid molecule was located.

The UMI allows to track back which sequences originate from the same molecule.

In certain embodiments, the segregation of the sample into a plurality of compartments is performed via magnetic beads. The magnetic beads are decorated with compartment nucleic acid molecules, which are linked to the recorder nucleic acid molecules. Each magnetic bead with its load is distributed to a single compartment via its magnetic property.

In certain embodiments, the segregation of the sample into a plurality of compartments is performed via emulsion droplets, which contain the compartment nucleic acid molecules. The segregation can be performed by limiting dilution which results in statistically less than one particle per droplet, thus ensuring a clonal distribution of both sample-derived particles.

Alternatively, the segregation is performed by active sorting of single particles into droplets or droplet size limitations that do not allow more than one particle to be present.

In certain embodiments, the primer binding sequence is a poly-A tail.

In certain embodiments, the step of amplifying the plurality of hybrid nucleic acid molecules is performed via PCR.

In certain embodiments, the step of amplifying the plurality of hybrid nucleic acid molecules is omitted and in step h of claim 1/the first aspect, single-molecule sequencing is performed.

In certain embodiments, step ii is performed via primer extension. In certain embodiments, the step (ii) of linking a compartment nucleic acid molecule to each barcode nucleic acid molecule to yield a plurality of hybrid nucleic acid molecules, is performed via ligation. In certain embodiments, step ii is performed via taqmentation. In certain embodiments, step ii is performed via hybridization.

In certain embodiments, before step i, the sample is segregated into single cells and the cells are pre-sorted. In certain embodiments, the cells are pre-sorted via FACS or affinity chromatography.

In certain embodiments, the cells are pre-sorted according to the modifications on the recorder nucleic acid molecule.

Detailed description of certain embodiments of the invention

Spatially resolved barcoding allows the identification and/or sequencing of molecules and nucleic acids in situ. Such barcoding may be useful for the identification and determination of specific cells and their distribution in a planar sample.

Immune cells of the adaptive immune system comprise antigen-specific receptors with unique sequences for their recognition. The B-cells encode B-cell receptors (BCRs) or secrete antibodies, whereas T-cells present T-cell receptors (TCRs) on their surface to recognise peptides presented on MHC-molecules of other cells. Both BCRs/antibodies and TCRs are composed of two chains with variable regions (CDRs) that are translated from separate mRNAs. In order to determine the function and reconstruct these immune receptors, it is necessary to obtain the sequence of both chains in a coupled fashion. A method that can spatially resolve the location of these mRNAs, preferably down to a single cell resolution, would allow such a coupled sequence determination.

Lymphoid structures are involved in the immune response to many diseases and are typically in direct vicinity to the diseased tissue. Lymphoid structures comprise regulatory networks of different types of lymphocytes. B-cells can contribute to the immune regulation by presenting antigens recognised by their antibodies to T-cells. A co-localisation of B- and T-cells in lymphoid structures can therefore directly correlate with their cognate antigen-specific interaction. Lymphoid structure-specific antibodies provide a fast track to the discovery of antigens for diagnostics and therapy of diverse diseases such as cancer and autoimmunity.

Spatial characterisation of cells

It has been discovered that disease-specific B-cells can selectively and efficiently present antigens directly or indirectly via secreted antibody and dendritic cell to a disease specific T- cell. Based on this discovery, a novel approach for the identification of disease-specific antigens has been developed by identifying disease-specific antibodies and/or B-cells and/or T-cells in situ. Disease-specific antibodies have several uses such as directly targeting of the diseased cells and/or tissue or identification of the respective disease-specific antigens, wherein said disease-specific antigens can be recognised by antibodies/BCRs and/or TCRs.

In order to identify the subset of disease-specific antibodies, other tissues and more distant lymphoid structures and/or blood may be obtained and only specifically enriched clones found in the diseased tissue are considered as disease-specific antibodies. Ideally, the spatial distribution of immune cells and/or diseased cells is characterised in a highly resolved manner (spatial profiling).

It may be sufficient to obtain a low level of information of the cells in situ that can serve as a spatial scaffold to be complemented with a more detailed information obtained by a less site- specific characterisation. It is important that the spatial information allows proper assignment of the detailed information. For instance, a set of probes may be sufficient to characterise the cells and their respective distribution pattern of biomolecules that can be used to assign more detailed information. Such probes may detect important immune cell markers comprising TCR, BCR, CD4, CD8, PD1 , FcR, and MHC. These markers may serve to determine the exact type and status of immune cells which is important to understand the respective function or lack thereof in the diseased tissue. More importantly, the spatial resolution may be high enough to detect immune cells in synapsis and their activational status. The latter may be detected by probes recognising the early signal transduction such as phosphorylation of the immune receptor ITAMs, or the later stages such as cytokine expression or related markers. Probes may recognise nucleic acids and/or the expressed antigens to obtain a distinct pattern preferably of each immune cell in the sample. In order to more correctly assign CDRs identified by sequencing, the probes for TCRs and/or BCRs used in the spatial characterisation may separately recognise the each of the V(D)J segments.

A detailed characterisation can comprise a single cell sequencing that allows the coupling the information of biomolecules, or a more scrambled information typically obtained from next generation sequencing of nucleic acids prepared from a sample. For instance lOxGenomics provides a microfluidic technology that allows sequencing of RNAs originating from single cells, thereby allowing a coupling of the information that is found on different mRNAs such as antibodies and T-cell receptor. Other platforms such as the C1™ system provided by

Fluidigm or BD Rhapsody™ by Beckton Dickinson are also available. Notably, it possible to use so-called SPLiT-seq to sequence the even mRNA of fixed cells in a coupled manner. Ideally, the resulting information allows the identification of a disease-specific immune cell and the reconstruction of the receptor recognising the disease-specific antigen.

In situ barcoding methods

The method of combinatorial barcoding disclosed herein is not limited to the determination of variable sequences of immune receptors in immune cells or antigens, but can be applied to any kind of target molecule including peptides and oligosaccharides that can be bound by probes. Samples comprising target molecules can comprise cells, tissues, viral particles, biofilms, natural and artificial particles or surfaces. The irradiation-controlled combinatorial barcoding is best applied to molecules in a sample that can be spatially addressed as separate features and preferably retain their position during the assembly procedure.

Preferably the sample is provided in a substantially planar form and the molecules are fixed in their relative position during the combinatorial barcoding process. The substantially planar surface to which a sample is attached may not be provided in a continuous form, but may comprise channels, nooks, creases and separate wells. For instance, a microtiter plate comprises an array of planar surfaces or features in separate wells in which one or more samples can be provided for in situ barcoding. Thus, a well can be considered as a feature that will be provided with a unique barcode sequence, or may be subdivided into a multitude of separate features in which each feature will be provided with a barcode reflecting the respective coordinates.

Homogeneous samples or suspensions can be transferred to a substantially planar surface and fixed in their coordinates for spatial combinatorial barcoding. The substantially planar surface may comprise ridges or wells to separate or guide the flow of reagents. If target molecules are encapsulated or located in cells, it is preferred that the respective barrier is permeated or removed in order to make said molecules accessible to probes. The probes are preferably highly specific for their cognate target molecule. Typical probes are antibodies and derivatives thereof, peptides, nucleic acid-binding proteins (e.g.

derivatives of transcription activator-like effector nuclease (TALEN) or Cas9-, or Cas13b- derivatives), other complex-forming proteins, haptens and/or their ligands, oligonucleotides such as aptamers, and preferably hybridisation probes and primers. It is preferred that probes comprise a nucleic acid moiety that can be used to generate a combinatorial barcode.

A more special embodiment is in situ sequencing without using highly specific probes. For instance, it is possible to use degenerate primers or polyT oligonucleotides to randomly target sequences or mRNA. The hybridisation or strand invasion may be enhanced by additional factors such as single-strand binding proteins (e.g. E. coli SSB) or recombinases (e.g. RecA). It is also possible to achieve direct covalent attachment to the ends of nucleic acids by ligases or by tagging of continuous nucleic acids by using transposases. So-called tagmentation can be used to efficiently tag entire genomes with defined sequences that can be barcoded. Thus, it is possible to barcode nucleic acids of individual cells, organelles, chromosomes, viral particles and plasmids in a defined spatial manner for later assembly of co-localised sequences in an ordered manner. A more specific application is the spatial reconstruction of genome organisation in eukaryotic nuclei or prokaryotic nucleoids.

Means for site-specifically attaching barcodes to a probe preferably involve excitation by a light source that can be highly focused. Preferably, photo-cross-linking is performed by irradiation at wavelengths longer than 300 nm in order to preserve the integrity and structure of nucleic acids. The attachment can be performed photochemically by direct cross-linking of probes with a barcode or subunits thereof, or indirectly by site-specific deprotection of the probes and subsequent chemical or enzymatic attachment. However, it is preferred to perform direct attachment in a single step by photo-cross-linking without the requirement of enzymes. The most preferred type of photocrosslinking employs a group that is mainly reactive during excitation by irradiation, thus providing the high degree of spatiotemporal control required for correct barcode generation. Other photocontrolled attachment methods that involve deprotection lead to a reactive group or substrate that can result in off-target attachments for instance by dissociation of barcode subunits from the cognate target after activation before the enzymatic or chemical attachment step.

Barcode subunits preferably comprise oligonucleotides, peptides or proteins. Many proteins are known that bind to nucleic acids with high sequence specificity and affinity. Especially TALENS can be engineered to bind to virtually any kind of sequence, even recognising certain base modifications. Alternatively, Cas9 variants have been generated which bind nucleic acids dependent on their guide RNAs. In addition, Cas13b can be engineered in a guide-RNA dependent manner to specifically target RNA sequences. Several sequence- specific single-stranded DNA binding proteins are known. For example, KH domains are known that recognise short ssDNA sequences in a sequence-specific manner. Proteins can be site-specifically modified (e.g. by introduction of a cysteine in the ligand binding interface) to comprise a reactive group that can be further derivatised with a photo-cross-linker for irradiation-dependent cross-linking to its bound target.

The non-enzymatic attachment can be performed by making interstrand cross-links (cross links between two separate nucleic acids or between a nucleic acid and a protein or peptide) or ligation by cross-linking neighbouring bases. It has been surprisingly discovered that interstrand cross-links (ICL) is more rapid and efficient than ligation and therefore preferred when performing non-enzymatic attachment. Preferably, the attachment step is fast, requiring less than 10 minutes for attachment of more than 50% of a given subunit. More preferably, the attachment step is very fast, requiring less than 1 minute for attachment of more than 50% of a given subunit. Most preferably, the attachment step is ultrafast, requiring less than 10 seconds for attachment of more than 50% of a given subunit. Efficient means of interstrand photo-cross-linking are psoralen-coupled oligonucleotides, diazirine-based nucleoside analogues, aryl-azide-based (preferably nitrophenyl azides) nucleoside analogues, 5-iodouracil, 6-thioguanine, 4-thiothymidine, coumarin-modified thymidine, 3- cyanovinylcarbazole nucleoside and the like. A coumarin-modified thymidine (e.g. linked by a triazol) efficiently induces interstrand crosslink (ICL) formation upon photoirradiation at 350 nm while the cross-linking can be reversed by irradiation at 254 nm (2). When 3- cyanovinylcarbazole nucleoside (^CNVK) is incorporated into an oligonucleotide, very rapid photo cross-linking to the complementary strand can be induced at one wavelength and rapid reversal of the cross-link is possible at a second wavelength. Neither wavelength has the potential to cause significant DNA damage. Irradiation of a duplex containing a single incorporation of ^CNVK at 366 nm leads to an almost complete cross-linking to thymine base in 1 second. Especially ^CNVK can be used to perform a reversible photocrosslinking in the case it is required to anchor a probe to the target while performing multiple other reactions such as washing steps on the sample. Several other carbazole derivatives such as 3- methoxycarbonylcarbazole nucleoside (^OMeVK), 3-carboxyvinylcarbazole nucleoside (^OHVK), and 3-carbonylamidevinylcarbazole nucleoside (^NH2VK) are known with different properties with respect to crosslinking speed, nucleobase selectivity and reversibility. As both primer and barcode are synthetic oligonucleotides, even completely different photo-cross-linking reagents can be used for attachment. Such compatible reactive groups are known from photochemistry. Yet another modification of 3-cyanovinylcarbazole nucleoside (^CNVK) is known in which the ribose sugar is replaced with a threose moiety (^CNVD). The photoreactivity of ^CNVD is ca. 1.8-fold (toward T) and eightfold (toward C) greater than that of ^CNVK. In the case of ^CNVD, since the difference of the photoreactivity between T and C is closer than that of ^CNVK, the freedom for designing the sequences for the photo-cross-linking reaction is increased.

For enzymatic attachment, a barcode can be hybridised to a probe 3’-end that is protected by a photocleavable group which can be copied by a polymerase only if deprotected by excitation. In order to assure both a highly selective and rapid photo-cross-linking, it is preferred that the reactive groups are in close proximity with each other when properly assembled by binding. For example, a reactive group such as the vinyl group of ^CNVK has a certain degree of freedom in rotation and diffusion and can therefore be positioned within an occupational sphere for all its possible conformations. However, the radius of such an occupational sphere differs strongly between the free and unbound form and the bound form in a hybridised duplex with adjacent and surrounding nucleotides. Preferably, the

occupational radius of the reactive group in the bound complex such as a double stranded nucleic acid is less than 5 nm, more preferably less than 2 nm and most preferably less than 1 nm and comprises the target moiety such as cytosine, 5-methyl-cytosine, thymidine or uracil in the opposite strand.

Combinatorial barcode assembly

The barcode can be assembled by repeated attachment cycles using defined subunits. This allows a more combinatorial barcoding of the probe for high throughput distinction of molecules in samples or cells. In order to fully exploit the power of combinatorial barcode assembly it is required to employ a highly efficient means of subunit attachment. It is possible to spatially barcode all primers in a sample in a combinatorial manner by strict coordinates based on a grid encompassing a large part of the substantially planar sample. However, it is preferred to only select the coordinates of the previously identified desired features differentially to reduce background and conserve the number of subunits required to distinguish features in samples. In addition, a strictly grid-based systematic barcoding automatically means that neighbouring features could be falsely targeted due to focussing errors during the photoactivation which can lead to overlapping barcodes. In order to generate barcodes in situ economically, it is preferred to first characterise the sample. For example, whole regions of interest can be described for obtaining information on certain tissue types in a sample. If single-cell information is to be obtained, it is preferred to select the cells of interest and irradiate only these for combinatorial barcoding and not the surrounding area. Depending on the presence and distribution of the target molecule a sub- cellular structure may be targeted. For instance, membrane proteins can be targeted by illuminating the cell membrane, mRNA can be found in most parts of the cytoplasm and DNA in the nucleus. Thus, for cells it is preferred to employ high resolution optics to obtain information from the cell of interest will low contamination from untargeted regions. If the sample is a thin section, it is possible to use generic staining methods and microscopy to identify cells and their sub-cellular structures.

In some cases, it is preferred to perform immunohistochemistry in order to identify the cells of interest and their coordinates in a given sample. Such initial characterisation of a sample can reduce the number of features to be barcoded. The amount of different barcode subunits per attachment cycle is mostly determined by the efficiency of the barcode attachment. If the barcode attachment is highly efficient, then more subsequent barcoding assembly cycles can be made. In addition, the more different barcode subunits are used per assembly cycle, the more separate barcode subunit incubation, attachment and washing steps have to be performed. This means that a properly hybridised probe or primer could be lost in too many recurrent steps. Therefore, it is preferred to optimise combinatorial barcode assembly by employing a similar amount of different barcode subunits per cycle as the amount of total cycles. In addition, it is preferred to initially cross-link the probe to its target molecule or other molecules in direct proximity to preserve its spatial orientation during the entire barcoding procedure. The cross-link of the probe can be performed chemically or photochemically and may be reversible. Preferably, the cross-linking method of the probe is compatible with the barcoding procedure.

An amount of n cells are combinatorially barcoded in s steps (for each different subunit per cycle) and c cycles: n=s^c, whereas the total amount of required steps for assembly is only s*c. Thus, to keep the amount of cycles and the total amount of required steps low, it is preferred to use a greater amount of different barcode subunits per cycle if the hybridisation and washing steps are very fast. The combinatorial barcoding method is very powerful: a set of 10 barcode subunits per cycle and 10 steps in just 3 cycles (a total of 30 steps) can assemble a distinct barcode for up to 1000 features. However, the number of subunits per cycle can vary. For instance, a set of 8 subunits can be used in 2 cycles and 4 in one cycle to barcode 1024 features in a total of 3 cycles.

Another important consideration is the total irradiation time required for each step. This largely depends on the duration d, required for activation or crosslinking and the number of features to be processed per cycle. Parallelisation of photocrosslinking is highly desirable. If more than one features can be addressed in parallel, the time required for subunit attachment can be reduced considerably. For instance, laser beams can be split and thereby multiplied for parallel and independent selective irradiation of features. Alternatively, several individual light sources are employed to selectively address more than one feature in parallel. Yet a much higher degree of parallelisation can be achieved by using a digital micromirror device (DMD) that allows simultaneous irradiation of millions of features in parallel, thus reducing an attachment step for all of these features to mere seconds. The total processing time t comprises the duration d_h of hybridisation and washing steps required for each subunit in a cycle: t=(d*n+d_h)*c. The number of features to be processed per cycle can be reduced if a proofreading method is employed that can distinguish between an inefficient subunit attachment and an intended absence of attachment. For instance, a binary information can be encoded in a barcode with just one subunit per cycle if two or more attachment sites are present for such a barcode template. Thus, if only one attachment site comprises a subunit in a barcode, it still counts as a positive signal instead of a complete absence of attachment.

Alternatively, two or more different subunits are offered in one attachment step to introduce redundancy for proofreading. Therefore, instead of processing all features in one cycle, half of them can be“skipped” in a binary setting by attaching only one subunit in a given region (see Fig. 1). This can strongly reduce the time required to assemble a barcode. In addition, skipping reduces the total amount of exposure of features to potentially damaging irradiation. It is also preferred to reduce potential damage by photooxidation by providing antioxidants (also often termed as“antifade” reagents in microscopy) in the buffer during irradiation steps. Preferred are thiol-based antioxidants (e.g. DTT), more preferred are phosphine-based antioxidants (e.g. TCEP) and most preferred is ascorbic acid-based antioxidants (e.g. vitamin C). However, the skipping effect is less pronounced the more subunits are attached per cycle. If subunits are expensive, a binary setting of skipping (0) and the use of only one subunit (1) per cycle is most economic for barcoding (exemplified in Fig. 1). For instance, only 10 different subunits can be used to barcode up to 1024 features in this manner. If many different subunits are employed in barcoding, it is preferred to introduce modifications such as Super T (5-hydroxybutynyl-2’-deoxyuridine) that compensate for the weaker AT hydrogen bonds in sequences to ensure more specific hybridisation to occur even with short oligonucleotides. See examples to calculate total processing times in seconds for a combination of 1024 features, 60 seconds of subunit hybridisation and washing steps and 1 second of attachment time (in a serial manner) per feature in Table 1 using various amounts of subunits per cycle or skipping of attachment.

Proofreading itself can also help in resolving issues with overlapping features and barcoding. If the template for barcoding is too large for the probe portion to bind to its cognate target, it is possible to attach the barcode template to the probe or primer after initial in situ binding to the target as an adapter. Preferably, the barcode template adapter is attached covalently to the probe or primer by a ligation reaction in order to allow a seamless and coupled readout of the information encoded by the probe or elongated primer and the barcode. Preferably, said ligation is carried out chemically or enzymatically for the entire sample before barcoding in order to minimise photooxidation or other damage by too many photoactivated attachment reactions. Table 1 : Calculation examples for serial barcoding of features and total processing time

From these examples it is obvious that a high-throughput labelling of 1000 features is extremely time consuming, if not impossible due to sample and probe instability, if the attachment method is anything slower than ultrafast if attachment is performed by serial irradiation of features. Therefore, it is preferred to perform very fast attachment methods if the number of features exceeds 50 and ultrafast methods if the number of features is higher than 200. A higher number of subunits per cycle help reducing the required amount of cycles, but at the same time increase the requirement for hybridisation and washing steps. If barcode subunits are assembled in a linear fashion by attaching one subunit at the end of another, the total efficiency of barcoding follows a compound interest calculation. If accidentally no subunit was attached during linear barcode assembly in a molecule, this would result in a deletion of information and it is not possible to determine at what point this has occurred. In order to circumvent such deletions in the final product it is preferred to employ different subunits in each barcode assembly cycle that have attachable ends that are only compatible between cycles. This means that some molecules will be lost due to incomplete assembly of a linear barcode (see Fig. 2). Thus, if the attachment of the barcode subunits efficiency is low, it is preferred to use less attachment cycles at the expense of performing more steps per cycle. In order to further reduce the problem of lost barcode information during linear assembly, it is preferred to assemble a branched and/or templated barcode. Each branch of the barcode template preferably comprises at least one or more specific region that allows only one subunit to be specifically attached per cycle (see Fig. 3). More than one specific region may be present for each barcode subunit in order to introduce redundancy for proofreading.

It is preferred to use more than one photo-cross-linker in an attachment step, each of which can be triggered selectively at different wave lengths. For example, a multispectral photo- cross-linking can employ pyranocarbazole groups (^PCX) that efficiently for an ICL by irradiation at 400 and/or 450 nm in combination with ^CNVK which is separately and selectively excited at 366 nm (3). Thus, 2 different barcode subunits can be attached simultaneously and independently in one step, effectively halving the required time and the potentially damaging exposure to near UV radiation for combinatorial barcoding.

If the assembled barcode is to be determined by hybridisation using a method such as nCounter, it is preferred to assemble the barcode in a nanostring-like fashion that allows optical resolution of the individual subunits. If the barcode was assembled by a reversible crosslinker such as ^CNVK, it is possible to selectively remove a potentially interfering barcode after its readout before sequencing of the elongated primer. Alternatively, a cleavable linker can be used to remove the attached barcode or other potentially interfering moieties. If the barcode is to be decoded by hybridisation steps, it is preferred to employ hybridisation probes and barcode subunits that do not strongly interact with naturally occurring

polynucleotides such as RNA and DNA. It is preferred to use oligonucleotides with non standard base modifications also known as“Self-Avoiding Molecular Recognition Systems” (SAMRS). It is also possible to use non-natural backbone modifications such as spiegelmers that only form duplexes with their respective antisense. Ultimately, it is possible to combine base and backbone modifications to exclude any background arising from interactions with natural polynucleotides.

Washing steps are performed to remove any residual barcode subunits that are not attached to the template before repeating addition of new barcode subunits in another cycle. The washing steps can also be performed in a less stringent manner in order to remove any barcode subunits that are not bound or hybridised to the template before the attachment step. This may result in less off-target attachment of subunits within the sample. The washing steps may be performed depending on the nature of the barcode subunit. For instance, by heating the sample below the melting temperature can destabilise unspecifically bound subunits for removal. It is also possible to use chemical denaturants to destabilise unspecifically bound subunits. Another means to increase specific hybridisation of barcode subunits to their cognate sequence tags is to add competitive blocking oligonucleotides, preferably lacking photoactivatable groups, that hybridise to non-cognate sequence tags of the recorder nucleic acid molecule. For non-covalently attached subunits more stringent conditions can be applied. Nucleic acids are best denatured with a denaturant such as formamide. Typically, formamide washes (e.g. 50% formamide in water) may be performed to remove any unbound linker and barcode, followed by a water wash before applying the next subunit in a hybridisation buffer. Proteins are more effectively denatured by urea or guanidinium hydrochloride. However, it is preferred to remove proteinaceous barcode subunits by competitive elution using oligonucleotides and/or heparin.

After barcoding, the probes and/or primers in the sample must be retrieved for analysis and/or editing. Many commercial nucleic acid isolation kits are available and can be chosen depending on the nature of the sample to optimally retrieve barcoded probes and/or primers. It is preferred that probe, primers and/or barcode subunits comprise haptens such as biotin and/or digoxigenin to effectively separate the barcoded nucleic acids using affinity chromatography or direct immobilisation to a solid phase for analysis.

Combinatorial sequence editing

A special case of using a barcode template involves assembly-dependent sequence editing. For instance, the barcode precursor or template comprises attachment regions that can be modified by or after assembly. Thus, the information of the subunit attachment status is transferred to the template strand (see Fig. 4). Such information transfer is especially advantageous for sequencing methods that do not require amplification steps. A great number of nucleic acid modifying enzymes can discriminate between single-stranded and double stranded nucleic acids. Preferably, the modifying enzymes recognise only double- stranded regions and do not modify single-stranded regions. Typical examples for such enzymes are methyltransferases or repair enzymes. By specifically generating an

attachment-dependent modification, the generic template strand is edited resulting in a unique barcode. Alternatively, the nucleotides of template strand are specifically masked due to subunit attachment and only become accessible to altering enzymes or reagents in other regions due to exposure in a single-stranded context. It is preferred to keep attachment- dependent modifications restricted to the barcode region in order to preserve the information encoded in the probe and/or elongated primer. Especially the deamination of cytosine to uracil and/or 5-methyl-cytosine (5mC) to T is a useful template editing method due to its simple and effective detection by most sequencing methods. Enzymatic deamination of C in a single-stranded region can be performed with many deaminases, however AID and APOBEC3G are preferred because they cannot deaminate 5mC to T. Thus, single-stranded regions other than the barcode can be protected from deamination by using 5mC or by adopting a double-stranded conformation. Alternatively, deamination of single-stranded regions may be performed chemically by using reagents such as sodium bisulphite to convert cytosine residues to uracil residues in single-stranded DNA, under conditions whereby 5mC remains non-reactive.

Methyltransferases mostly recognise double-stranded DNA for modification, but require a specific target sequence. However, some methylases such as Human DNA (cytosine- 5) Methyltransferase (Dnmtl), CpG Methyltransferase (M.Sssl) and GpC Methyltransferase (M.CviPI) require only a sequence of two bases for cytosine methylation. Such short recognition sequences make the sequence design of attachment regions more flexible while also offering the simple generation of multiple modification sites per attachment site for proofreading purposes. A non-limiting list of modifying enzymes other than

methyltransferases and deaminases for attachment-dependent template editing comprises 5- hydroxymethyluridine DNA kinase, and repair enzymes such as alkyladenine DNA

glycosylase, methylcytosine dioxygenases of the TET family, and enzymes of the Dnd complex for DNA phosphorothioation.

Alternatively, a crosslinked barcode subunit is removed leaving behind a“scar” as a modification of the template that can subsequently be detected by hybridisation or sequencing. Such removal may be achieved by using enzymes selectively degrading the subunit. Yet, it is also possible to chemically or photochemically remove most of the subunit by using a cleavable linker or labile nucleotides. For instance, RNA-bases can be selectively cleaved under alkaline conditions and a phosphorothioate backbone can be selectively broken by reagents such as peracetic acid. Alternatively, DNA repair enzymes can be used to deglycosylate cross-linked bases, leaving behind an abasic site in the template.

Preferably, the DNA repair enzyme does not have a lyase activity or the backbone is modified to protect the edited template from strand cleavage. The abasic site is normally unstable under alkaline conditions also forming reactive aldehydes and may therefore be kept under neutral conditions until analysis. Alternatively, the abasic site can be further stabilised by reaction with amine groups forming covalent amine adducts. Especially nanopore-based sequencing methods are useful to detect abasic sites or adducts thereof in a nucleic acid. Another more specific type of ICL is represented by a proteinaceous barcode subunit covalently linked to the template. In this case, the protein itself can be removed efficiently and selectively by digestion with a protease such as proteinase K. The remaining crosslinked amino acid on the template can also be recognised by nanopore-based sequencing.

Bulky scars, or even entire subunits cross-linked (nucleic acid and/or protein barcode subunits) on the template can be useful for barcode distinction in less stringent sequencing methods. For instance, the sequencing by expansion (SBX™) method currently developed by Stratos Genomics employs nanopores that do not directly distinguish nucleobases, but rather the linker properties between them. If so-called Xpandomers are used in the elongation of the primer and/or probe, it is possible to obtain the whole sequence information along with the barcode identification.

It is preferred to analyse the barcode and its modifications by a single-molecule sequencing method because many modifications such as methylations are lost during copying or amplification and the exact copy number of respective probes or elongated primers in a feature can be determined. If amplification is performed before sequencing, a recoding modification such as C U or 5mC T is preferred for editing of the template. In addition, a Unique Molecular Identifier (such as a short stretch of random nucleotides) can be added before the amplification step to keep track of the copy number and detect potential mutations introduced by said amplification.

Currently, many different base modifications can be directly distinguished by single-molecule sequencing methods such as Single Molecule, Real-Time (SMRT®) sequencing by Pacific Biosciences and nanopore DNA sequencers such as the MinlON of Oxford Nanopore Technologies.

A preferred editing method employs reversible photo-cross-linking of subunits to a template that directly causes a recoding modification for editing without the use of any enzymes in the combinatorial barcoding process. More specifically, a reversible photocrosslinker such as a carbazole derivative can directly alter the coding properties of the linked nucleobase. The cross-linking of ^CNVK to thymine in the opposite strand is ultrafast, but does not alter the base after reversion of the crosslink. However, crosslinking of ^CNVK to cytosine in the opposite strand can generate an uracil instead by selective deamination albeit the cross-linking with cytosine is slower than with thymine. This can be compensated by using 5mC which is almost as fast in cross-linking as T. In addition, the flanking base composition can also influence the cross-linking speed and efficiency. More specifically, it has been discovered that a 5’-positioned 2-aminopurine (2aP) or inosine to ^CNVK can accelerate the cross-linking of ^CNVK to cytosine (C). It is assumed that the restricted motion of C in the double-stranded DNA caused by base pairing with an opposite nucleobase suppressed the photo-cross- linking reaction of ^CNVK and C. An irradiation at 366 nm, 1600 mW/cm² is sufficient to crosslink more than 50% of ^CNVK (next to 2aP) to C in only 1 second and close to 100% in 3 seconds.

Editing of carbazole-crosslinked 5mC or C can be performed by heating of the adduct and subsequent photosplitting at 312 nm (120 mW/cm² at 37°C). The fastest conversion with a yield of >85% uracil is observed with the photoadduct of ^OHVK at 70°C after 6 h. The sequence context may also affect the speed and efficiency of the editing of C to U. For example, an inosine directly 5’ of ^OHVK is fastest at 38°C, taking 7 days to a yield of more than 65% uracil. However, the deamination reaction is even more accelerated if the 5’-end of ^CNVK is terminated by a phosphate (p^CNVK) or phosphorothioate with a yield of almost 100% uracil formation after 24h at 37°C. The reaction was further enhanced in the presence of 100 mM sodium ions. Thus, it is preferred to perform the conversion of the photoadduct of cytosine with p^CNVK at a temperature between 50° and 100°C in the presence of >50 mM NaCI. Yet it may be possible to further accelerate the conversion reaction by other means such as microwave irradiation or use of other nucleophilic modifications at the 5’ end. As annealing and crosslinking kinetics of a 5’ p^CNVK-modified subunit are not as fast as an internally modified subunit, it is preferred to additionally employ at least one additional internal 5’ inosine-^CNVK modification with a T in the opposite strand for efficient crosslinking. This ensures an ultrafast covalent attachment of barcode subunits and allows a final polishing step in which any unreacted 5’ p^CNVK modifications can be crosslinked by irradiation at 366 nm. Alternatively, the crosslinking is performed by an internal (optional 5’-inosine)^CNVK modification and cleavage of the 5’ group is performed to generate a 5’ p^CNVK end before the deamination step. Such cleavage may be achieved enzymatically by digesting the 5’ end by an exonuclease such as T7 exonuclease or by an inosine-specific repair enzyme such as endonuclease V, or chemically by using a scissile modification such as an RNA base or phosphorothioate 5’ of the ^CNVK modification. The same considerations hold true for the more C-reactive threose modified nucleotide analogue ^CNVD.

It is preferred that the combinatorial barcoding process can be monitored by fluorescence microscopy. Ideally, the sample is visualised for attached barcodes after a washing step before application of new barcode subunits. If barcode subunits are labelled or the ICL itself is fluorescent, the efficiency of the combinatorial barcoding at desired features and/or ROI can be determined. The evaluated degree of barcoding per feature can be used for quality control and/or to correct any variations in the number of barcoded probes obtained from respective features.

Method for combinatorial barcoding of multiple features in a substantially planar sample.

1) provide a substantially planar sample of which the coordinates of desired features are given

2) provide probe(s) and/or primer(s) to said substantially planar sample

3) incubate probes under conditions that specific binding occurs to target molecules

4) provide a specific barcode or subunit thereof

5) site-specifically attach the barcode or the subunit thereof to target bound probe at coordinates of one feature

6) remove non-covalently attached barcode or subunit thereof

7) repeat steps 4)-6) until a sufficient number of desired features are barcoded 8) retrieve barcoded probes from sample

Optionally, step 3 can comprise additional steps

a. cross-link probe(s) and or primer(s)

and/or

b. attach barcoding region to probe(s) preferably by ligation, wherein the

barcoding region optionally comprises a UMI

Method for combinatorial editing of probe(s) bound to multiple features in a substantially planar sample.

2) provide probe(s) and/or primer(s) to said substantially planar sample

3) incubate probe(s) under conditions that specific binding occurs to target molecules

4) provide specific barcode or subunit thereof

5) site-specifically attach barcode or subunit thereof to target bound probe(s) at

coordinates of one feature

6) remove non-covalently attached barcode or subunit thereof

7) repeat steps 4)-6) until a sufficient number of desired features are processed

8) retrieve subunit-modified probe(s) from sample

9) edit subunit-modified region of probe(s)

Optionally, step 3 can comprise additional steps

a. crosslink probe(s) and or primer(s)

and/or

b. attach barcoding region to probe(s) preferably by ligation, wherein the

barcoding region optionally comprises a UMI

• Sequential elongation and sampling

1) provide a substantially planar sample of which the coordinates of desired immune cells are known

2) provide primers to said substantially planar sample

3) hybridise primers under conditions that specific binding occurs to target sequences

4) site-specifically deprotect primers from coordinates of one immune cell

5) elongate deprotected primers with polymerase

6) elute elongated primers and store sample in a separate container.

7) repeat steps 2)-6) until a sufficient amount of desired immune cells are processed • Parallel elongation and sequential sampling

2) provide primers to said substantially planar sample

4) elongate primers with polymerase

5) site-specifically elute elongated primers and store sample in a separate container.

6) repeat step 5) until a sufficient amount of desired immune cells are processed

• Parallel elongation and combinatorial barcoding

2) provide primers to said substantially planar sample

4) elongate primers with polymerase

5) provide a specific barcode or subunit thereof

6) site-specifically attach barcode or subunit thereof to elongated primers at coordinates of one or more immune cell

7) repeat step 5)-6) until a sufficient amount of desired immune cells are barcoded

8) retrieve barcoded and elongated primers from sample

• Parallel elongation and combinatorial sequence editing

2) provide primers to said substantially planar sample

4) elongate primers with polymerase

5) provide a specific barcode or subunit thereof

7) repeat step 5)-6) until elongated primers of a sufficient amount of desired immune cells are modified by subunits

8) retrieve combinatorially modified elongated primers from sample

9) edit subunit-modified region of elongated primers

Spatial barcode linking to particles and/or cells

A planar sample may comprises a plurality of particles such as cells, or fragments thereof such as organelles (preferably nuclei) and/or viruses. Particles provide a physical linkage (covalent or non-covalent) of all its components such as the presence of target structures and/or sequence information of proteins and/or nucleic acids. Therefore, if a spatial barcode is linked to a particle, said barcode becomes also linked to its components. A planar sample can be disintegrated to release the particles that can be handled individually. Thereby the defined physical orientation of the particles is lost while the spatial barcode previously linked to the particle still carries this information. For analysis of its components, particles have to be at least partially disintegrated in order to release bound probes, target structures and/or nucleic acids. In order to still retain the linkage of its components, this disintegration step is performed in a compartment which provides a physical barrier that prevents mixing with other particles. This can be achieved by dispensing only one particle into a given compartment.

This can be performed in parallel for processing a plurality of sample-derived particles. While it is possible to analyse the target components (probes, spatial barcode and nucleic acids) of a particle including the spatial barcode separately from each compartment, it is preferred to pool the content from all compartments and analyse these in one step. This requires a linking of these target components to an index sequence that is unique for each of the

compartments and therefore to the target components released from the single particle. This index sequence must be linked to the target components before the content of the

compartments is pooled. By massive parallel sequencing, the pooled indexed target components can be analysed. Sequences comprising the same compartment index indicate the origin these sequences from a common particle. The sequences of the indexed spatial barcodes finally disclose the coordinates of the particles within the original planar sample.

Cells can comprise highly individual information. For instance immune cells such as T-cells or B-cells comprise immune receptors with variable sequences that are encoded on separate nucleic acids. Each B-cell or T-cell clone comprises a different set of variable sequences that determines its specificity. Dissolution of the cells would cause the clonal linkage of these encoding nucleic acids to be lost. If particles such as cells are to be characterised in a clonal fashion, the spatial barcode information can be linked to multiple particle components after transfer into separate compartments. A comparison of the general workflow versus the single particle approach is outlined in Figs. 7 and 8. This allows the retrieval of clonal sequence information in combination with location even if the size of the separate features addressed in spatial barcoding exceeds the size of particles.

Intact particles of natural origin such as living cells normally do not display nucleic acids on their surface that can be hybridised with a nucleic acid probe as part of the recorder molecule. Therefore, recorder nucleic acid molecules are preferably conjugated to antigen binding molecules such as antibodies or aptamers to decorate only target particles or less specific binders such as amine reactive groups or membrane-specific moieties (i.e. lignoceric acid) to decorate particles in a less discriminate manner. Identifier sequences can form an integral part of a recorder nucleic acid or be conjugated to other antigen binding molecules in order to indicate the presence of recognised antigens on, at, or in the particle. Such identifier sequences comprise a sequence specific to the antigen binding molecule, at least one or more constant regions common to other nucleic acid molecules in the sample that can be used for primer binding and amplification and may optionally comprise a UMI to quantify the bound antigens. A useful constant region common to other nucleic acids in the sample is a repeat of up to 28 adenines as many target molecules of interest are mRNA which are usually copied by using an oligo-dT primer with a length between 18 and 28 bases.

As several recorder nucleic acid molecules can be coupled to particles, it may not be required to use redundant barcode binding sites (ISTs). Dead cells or porous particles can be decorated with recorder sequences internally. The decorated particles are retrieved after the combinatorial barcoding of the substantially planar sample comprising said particles. Single particles can subsequently be distributed into separate compartments for linking their clonal information with spatial information. Optionally, particles can be pre-selected by employing one or more specifically labelled spatial barcodes to enrich only those particles from a region of interest. It is possible to deposit single particles into a compartment by performing limiting dilution which is directed to have less than one particle per compartment. Alternatively, particles can be handled by flow cytometry (FACS) or image-based cell selection and acoustic dispensing (see cellenOne by cellenion SASU, France) that allows single particle dispensing into multiple compartments such as microtiter wells. These single-particle/cell dispensing methods can be directly coupled with a region-specific pre-selection step based on fluorescent labelling of the respective barcode fragments on recorder molecules. A higher throughput in single-particle/cell analysis can be achieved by microfluidic platforms that generate emulsions with additional particles comprising individual index sequences such as Next GEM Technology (10x Genomics). Once single particles are isolated in a compartment, the spatial information can be linked to clonal particle information. Linking of such information can be achieved physically by combining spatial barcode sequences with particle-derived sequences and/or by linking both spatial barcode and particle-derived sequences with an index sequence that is unique for the compartment. Index sequences may optionally comprise a UMI. The most effective means of linking a multitude of sequences in parallel is performing assembly by overlap extension or template switching (for instance in a reverse transcription step). This is preferably performed in combination with PCR. Linking can also be achieved by ligation or tagmentation. The compartment-specific sequence can be applied in a highly ordered manner, preferably if the compartments are arranged in an arrayed format which allows addition of defined index sequences. These indices may be linked to one end of the sequence or both ends in a combinatorial fashion. Yet, it is also possible to link random indices as long as they are unique to each of the compartments comprising a single particle with spatial barcode. This is typically employed in microfluidic platforms such as Next GEM Technology which adds clonally linked unique index sequences in the form of beads. It is preferred to employ barcode nucleic acid molecules in the format as shown in Fig. 3 b) with an additional coding region which is not entirely hybridised to the recorder sequence. This coding region comprises a unique barcoding sequence corresponding to the ISP for the barcode nucleic acid molecule and may additionally comprise one or more constant region(s) common to other barcode nucleic acid molecules that can be used for primer binding and amplification. The unique barcoding sequence can be hybridised to a complementary oligonucleotide for labelling, addition of further functional sequences such as primer binding sites, and/or affinity- based enrichment. Fig. 9 discloses the use of a coding region comprising at least one additional primer binding site that can be used to link a compartment- specific index sequence, preferably after detaching the barcode nucleic acid molecule from its cognate ISP on the recorder nucleic acid molecule. Yet it is preferred to skip the reversal of crosslinks for detachment of the barcode nucleic acid molecule. This can be achieved by specifically hybridising a complementary oligonucleotide with at least one additional primer binding site before the single particle dispensing step. This hybridised oligonucleotide can be subsequently linked to a compartment-specific index sequence in order to transfer the information of the barcode nucleic acid molecule as depicted in Fig. 10. After linking of indices to the particle-derived sequences and spatial barcode information, the nucleic acid products can be pooled from all compartments for sequencing and subsequent analysis.

Co-localisation of single immune cells

Direct interaction of immune cells such as B-cells presenting cognate antigens to T-cells may be detected by in situ barcoding as well. The B-cell and T-cell clones can be identified on the basis of their specific variable sequences of their respective receptors as outlined in Fig. 5. The activational status of the T-cell by specific interaction with the presented antigen can be additionally probed for markers by short-lived signal transduction pathways such as phosphorylation of the TCR-associated ITAMs, for instance using antibodies as probes as outlined in Fig. 6. Depending on the resolution of the barcoded features, the directly interacting immune cells can be identified on the basis of variable receptor genes (both TCR and BCR) comprising the same barcode, or a combination of barcodes for each of the different immune receptors from neighbouring features.

Composition for combinatorial barcoding

A composition for combinatorial barcoding of probes bound to targets in a sample preferably comprises a composition of barcode subunits, a composition of barcode template and/or probes and a composition buffers for hybridisation, crosslinking and washing steps, and optionally opaque slides. The slides are preferably plastic microscope slides which are optically opaque and substantially non-fluorescent.

The barcode subunits preferably comprise at least one reactive group that can be used for ICL to a template upon photoactivation. The different barcode subunits preferably do not substantially hybridise to non-cognate templates under stringent conditions in order to allow a correct assembly of the barcode. Preferably, the barcode subunits comprise at least two reactive groups (photocrosslinker) that can be used for ICL to a template upon

photoactivation. Optionally, the barcode subunits comprise an additional fluorescent label for monitoring and/or quality control of the barcoding process that is excitable at wavelengths longer than the photocrosslinkers in order not to quench the radiation required for ICL formation.

The barcode template may be directly linked to a probe or may be linked to a probe in a specific manner in a separate step in situ by chemical or enzymatic ligation. It is preferred that the barcode template comprises a UMI. The barcode template comprises at least two regions for specific hybridisation to different substantially complementary barcode subunits. Preferably, the barcode template comprises at least 4, more preferably at least 10 and most preferably at least 20 different hybridisation regions for substantially complementary barcode subunits. The barcode template may be branched or linear.

A probe comprises a target specific binding region and preferably a nucleic acid moiety. A probe may comprise a barcode template or a region for specific attachment thereof.

Preferably, the region of attachment is covalently linkable to the barcode template after in situ binding. The probe may comprise an elongatable terminus and act as a primer for polymerases and/or ligases.

A composition of buffers suitable for combinatorial barcoding of probes bound to targets in a sample preferably comprises buffers for hybridisation and or binding, crosslinking and washing steps, wherein

• said crosslinking buffer comprises an antioxidant;

• and said washing buffer comprises a denaturant.

A composition of barcode subunits suitable for combinatorial barcoding of probes bound to targets in a sample preferably comprises at least 2 subunits with substantially different binding properties to specific barcode template regions, wherein each of said comprising subunits comprise at least one photocontrollable group capable of covalently forming an ICL.

A composition of barcode subunits suitable for combinatorial barcoding of probes bound to targets in a sample preferably comprises at least 4 subunits with substantially different binding properties to specific barcode template regions, wherein each of said comprising subunits comprise at least one photocontrollable group capable of covalently forming an ICL; and wherein said formed ICL can be used to edit the cognate barcode template region.

A composition of barcode subunits suitable for combinatorial barcoding of probes bound to targets in a sample preferably comprises at least 2 subunits with substantially different binding properties to specific barcode template regions, wherein each of said comprising subunits comprise at least one photocontrollable group, preferably ^CNVK, or a derivative thereof, capable of covalently forming at least one ICL; and wherein said formed ICL can be used to edit at least one base in the cognate barcode template region.

A composition of barcode subunits suitable for combinatorial barcoding of probes bound to targets in a sample preferably comprises at least 2 subunits with substantially different binding properties to specific barcode template regions, wherein each of said comprising subunits comprise at least one photocontrollable group, preferably ^CNVK, or a derivative thereof, capable of covalently forming at least one ICL; wherein said ^CNVK, or a derivative thereof is positioned at the 5’-end comprising a nucleophilic group, preferably a phosphate or phosphorothioate; and wherein said formed ICL can be used to edit at least one base in the cognate barcode template region.

A composition of barcode subunits suitable for combinatorial barcoding of probes bound to targets in a sample preferably comprises at least 2 subunits with substantially different binding properties to specific barcode template regions, wherein each of said comprising subunits comprise at least two photocontrollable groups, preferably ^CNVK, or a derivative thereof, capable of covalently forming at least two ICL; wherein one of said at least two ^CNVK, or a derivative thereof, is positioned at the 5’-end comprising a nucleophilic group, preferably a phosphate or phosphorothioate; and wherein at least one of said formed ICLs can be used to edit at least one base in the cognate barcode template region.

Apparatus for combinatorial barcoding

An apparatus for combinatorial barcoding of probes bound to targets in a substantially planar sample preferably comprises at least one light source, at least one digital micromirror device (DMD), valves and containers for buffers, barcode subunits, a programmable controlling unit, and means for heating and cooling of the sample.

A DMD may comprise at least 100 individually controllable mirrors. Preferably, a DMD may comprise 1000 mirrors or more than 1 ,000,000 mirrors. More preferably, a DMD may comprise more than 800,000 mirrors and most preferably more than 2,000,000 mirrors and even more than 4,000,000 mirrors. It is preferred that the intrinsic scattering of the DMD is low in order to achieve a very high contrast ratio (full-on:full-off (FO:FO) system contrast ratio). Preferably the contrast ratio if the DMD is greater than 300 at the preferred wave length, more preferably greater than 500 and most preferably greater than 1 ,000. It is preferred that the remaining background is removed by a filter that subtracts low power irradiation. It is preferred that scattering from the irradiated sample itself is reduced by mounting the substantially planar sample on a surface in the apparatus which is opaque and/or non-reflective.

As a DMD has a spatially fixed array of mirrors, the corresponding illuminated area is subdivided in independently controlled pixels in a similar grid-like fashion. Thus, it is preferred to treat the expressions“pixel” and“feature” in an interchangeable manner when using a DMD.

Preferably, said apparatus also comprises a microscopic imaging system to identify regions of interest (ROI) and/or individual target features in the sample and guiding and/or monitoring the combinatorial barcoding process by fluorescence microscopy. Said fluorescence monitoring is either performed by excitation at the crosslinking wavelength, or when additional labels are present in the subunits, at their respective wavelengths.

In order to be able to obtain a microscopic image of larger samples and/or to illuminate larger surfaces for combinatorial attachment by ICL, it is preferred to mount the sample on a motorised stage which can be controlled by a microstepping motor.

At least one light source preferably directly emits light in the range of 300-500 nm or is filtered to said range in order to enable ICL of by photocrosslinking. More preferably, the range is limited to 350-380 nm for most ^CNVK and 400-450 nm. Preferably, at least one light source can emit an energy of at least 1 ,500 mW/cm² for ^CNVK. In the case of ^PCX it is most preferred that at least one light source can apply 8,000 mW/cm² at 400 nm and/or at 450 nm an energy of at least 1 ,000 mW/cm².

The pixel size of for the DMD preferably is less than 100 pm², more preferably less than 10 pm² and most preferably less than 5 pm². Thus, with a very large array of mirrors, a given area can be combinatorially barcoded with a high resolution.

The apparatus preferably allows a parallel handling of several samples provided in a typical microscope slide format. It is possible to perform combinatorial barcoding on several such slides in parallel in the apparatus, yet it is preferred to perform a robotic handling of slides in order to increase throughput and minimise human interaction. For example, a“hotel” can be used to store several slides under conditions that preserve the integrity of samples until they are processed. A robotic hand can take a slide from the hotel and place it in the processing chamber. The processing chamber allows liquid handling of the slide by at least one inlet and one outlet. The apparatus also comprises valves, hoses and/or microchannels and containers to enable the controlled incubation of the sample to different buffers during the combinatorial barcoding process. Preferably, the apparatus comprises at least 3 containers for buffers such as cross-linking buffer, washing buffer and hybridisation buffer, and a waste container, and at least 2, more preferably at least 4 additional containers for different barcode subunits. In addition, the apparatus preferably comprises a heating and cooling unit to control the temperature of at least the processing chamber and the buffer chambers, and optionally the hotel and the barcode subunit containers. The heating and cooling should be enabled in the range of 30-90° at least for heating of the processing chamber and general cooling in the range of 0<10°C. The apparatus may also comprise an additional unit to control the humidity in the processing chamber if it is open during the barcoding procedure.

The apparatus comprises a programmable controlling unit that minimises human interaction required for combinatorial barcoding. The controlling unit preferably controls the processes of performing microscopy to identify and select features and/or ROI for combinatorial barcoding, the process of combinatorial barcoding comprising the iterative steps for buffer changes and the selective combinatorial hybridisation with barcode subunit and feature illumination by DMD for attachment thereof. The controlling unit may also move the stage comprising the processing chamber and its sample in order to allow processing of different areas of the sample. Finally, the processing unit also controls the exchange of samples between hotel, processing chamber and waste container for a higher throughput.

Examples

Example 1 : templated spatial barcodinq of target molecules

Well-specific oligonucleotides (characterized by a sequence specific for the microtitre plate well in which they are located) are labelled by templated spatial barcoding which can be used to trace back the location after template conversion and sequencing:

At first 7 different biotinylated oligos (biowell#1-7) with blocked 3’ ends are attached to 7 different wells in a streptavidin-coted microtitre plate (medium capacity from Biotez, Berlin, Germany). The attachment is performed according to the manufacturer’s protocol. 2 pmol of biotinylated oligos are applied per well. In a next step, 2 pmol the recorder nucleic acid oligo (bc2template) with duplicate ISTs for barcoding oligo binding is added to the wells that hybridises to the 3’ region of the biotinylated oligos. The hybridisation was carried out in 40 pi 1x NEB 2 Buffer at room temperature for 10 min and washed once with 40mI 1x NEB 2 Buffer, before adding 1x NEB 2 Buffer comprising 330 mM of each dNTP. Using Klenow- fragment exo- polymerase (NEB, Ipswich, MA), the hybridised 3’ end of the recorder nucleic acid oligo is elongated to copy the distinct sequence of the immobilised biotinylated oligos. A total of 1 U Klenow exo- was added to each reaction well to start the reaction. The microtiterplate was covered with saran wrap and incubated for 5 min at 37°C. The wells are washed twice with 50 pi crosslink buffer (100 mM NaCI, 50 mM sodium cacodylate, pH 7.6).

_• Sequential Barcoding

For conditional crosslinking, 5’-phosphorylated barcode CNVK oligos comprising CNVK in sequences for specific binding to different target regions of the template oligo and a region for barcode oligo hybridisation are used. The barcode CNVK oligos are applied to the beads in the wells in 50 mI crosslink buffer. The crosslinking is performed by irradiation at 365 nm for 30 seconds using a high power LED array (Kagel et al. , Photonics 2019,

https://doi . org/10.3390/photonics6010017).

Each of the 3 different crosslink barcode CNVK oligos (P04-CNVKbc#1-3) are sequentially added to all wells, hybridised, crosslinked and washed twice. In order to differentially crosslink barcode CNVK oligos to recorder templates, a specific irradiation pattern is used for each the wells:

_• Deamination

The deamination of crosslinked thymidines is performed in sealed microtiter wells by incubation at 90 °C for 6 h in crosslink buffer. The photo-crosslinks were cleaved afterwards in the same buffer by irradiation (312 nm, 121.3 mW cm2, 37 °C) for 2 minutes.

_• Sequencing

The deaminated recorder nucleic acid oligos were eluted from the wells under denaturing conditions and subjected to PCR with lllumina standard p5 and p7 primers. The amplified products were pooled and subjected to lllumina sequencing.

_• Evaluation

A total of 10.000 high quality reads were used for the evaluation. The efficiency of C U conversion was determined by correlation of the defined well-specific sequence copied by extension of the recorder nucleic acid template and the respective reads that identified a T in the respective crosslink-site. The average efficiency of C U conversion was found to be in the range of 80 % and no conversion was found in regions that were not cross-linked by the respective barcode subunits. By combining the two crosslink sites per 1ST, an overall correct identification of 96 % was found, leaving a total of 4 % lacking any conversion. Therefore, if more reads with specific sequences correlate with determined coordinates, a very high confidence for the correct barcode can be achieved.

Example 2: templated spatial barcoding of single particles

In order to demonstrate the feasibility of spacial barcoding for single-particle or cell sequencing, a recorder nucleic acid molecule was assembled on a magnetic particle comprising a well-specific sequence. Each single barcoded particle is then transferred into a reaction vessel comprising index primers. The index primers are used to amplify each of the barcode subunits and a bead-specific sequence, thereby providing indexed sequences in amplicons that correlate with the well-specific sequence of the single particle:

• Oligo binding to streptavidin beads

A microtitre plate is used for the spatial barcoding experiment. 7 separate wells are each dispensed with 10 pg of Dynabeads® M-270 Streptavidin (Thermo Fisher). For binding, each well is added with a biotinylated particle recorder nucleic acid oligo (and a different biotinylated oligo with a well-specific sequence (bioparticle#1-7). The binding step is carried out according to manufacturer’s recommendation using 50 pi 1x B&W Buffer comprising 10 pmol of both biotinylated oligos per well. The beads with bound oligos are retained by a magnet manifold during 2 washing steps with 50 mI 1x B&W Buffer.

• Sequential Barcoding

The beads in microtitre wells are resuspended in 50 mI crosslink buffer (100 mM NaCI, 50 mM sodium cacodylate, pH 7.6).

For conditional crosslinking, barcode CNVK oligos comprising CNVK in sequences for specific binding to different ISTs of the recorder nucleic acid oligo and a region for barcode oligo hybridisation are used. The barcode CNVK oligos are applied to the beads in the wells in 50 mI crosslink buffer. The crosslinking is performed by irradiation at 365 nm for 10 seconds using a high power LED array (Kagel et al., Photonics 2019,

https://doi . org/10.3390/photonics6010017).

Each of the 3 different crosslink barcode CNVK oligos are sequentially added to all wells, hybridised, crosslinked and washed twice. In order to differentially crosslink barcode CNVK oligos to recorder nucleic acid oligos, a specific irradiation pattern is used for each the wells:

_• Single bead sorting

The beads are pooled from all wells and hybridised to barcode oligos. Briefly, the pooled beads are resuspended in 50 pi 1x B&W Buffer comprising 100 pmol of all barcode oligos (bc-amp#1-3). After hybridisation, the beads are washed twice with 50 mI 1x B&W Buffer and sorted by BD FACSMelody™ cell sorter (Becton Dickinson Inc.).

_• Random sorting

In a first experiment, the pooled beads are sorted randomly without fluorescence to deposit single magnetic beads into wells.

Region-Selective sorting

In a second experiment, only beads comprising 488-nm blue fluorescence (with a FITC-5’ modification in bc-amp#1) are deposited directly into 96 PCR reaction vessels comprising 5 mI of 1 x PCR buffer.

_• Indexing of single beads by PCR

A combination of index primers that correlate with the coordinates of the 96 wells are added in aliquots of 5 mI in 1 x PCR buffer each. For this purpose, the primers IDT for lllumina- DNA/RNA UD Indexes were used (UDP0001 to UDP0096).

The PCR reaction vessels are subsequently supplied with 10 m I of PCR buffer comprising dNTP and polymerase and transferred to a thermocycler for PCR. A total of 25 cycles were performed.

_• Sequencing

Of all wells, 5 mI are used for pooling and purification before sequencing by an lllumina MiSeq sequencer.

A total of 100.000 high quality reads of the random sorting experiment are analysed for correlation of barcode fragments (designating the coordinates of beads in the wells during the sequential barcoding) with the well-specific sequence (a direct label for the beads in the wells before and during barcoding) bearing the same index combinations (single bead well coordinates). The final evaluation of the reads found more than 98 % of correlation of barcode fragments and well-specific sequences, thus corroborating the efficiency and specificity of the templated spatial barcoding of single particles.

For the region-selective sorting experiment, another 100.000 high quality reads were used for evaluation. It was determined that more than 97 % of the reads comprised the correct well-specific sequences associated (wells 4-7) with the region that is encoded by the fluorescently labelled barcode fragment (bc-amp#1). Of these, more than 98 % displayed the correct combination of well-specific sequences and the respective well-specific sequences. This demonstrates the feasibility of selecting particles by region and obtaining detailed spatial and clonal information by using recorder nucleic acids for templated barcoding in combination with indexes for single particles.

Oligonucleotide sequences

biowell#1-7:

#1 : Biotin-5’-

TCGTCGGCAGCGTCAGGTACATATACTTCCACCGAAAAAAAAAAAAAAAAAAAAAAAAAAAA (SEQ ID NO 001)

#2: Biotin-5’-

TCGTCGGCAGCGTCGTGGATACACACTTCACATCAAAAAAAAAAAAAAAAAAAAAAAAAAAA (SEQ ID NO 002)

#3: Biotin-5’-

TCGTCGGCAGCGTCTTAGATCGGATTACACGCTCAAAAAAAAAAAAAAAAAAAAAAAAAAAA (SEQ ID NO 003)

#4: Biotin-5’-

TCGTCGGCAGCGTCTTAACGGGTCGTAACACGGCAAAAAAAAAAAAAAAAAAAAAAAAAAAA (SEQ ID NO 004)

#5: Biotin-5’-

TCGTCGGCAGCGTCTCGAGCTGTGCATATAGTGCAAAAAAAAAAAAAAAAAAAAAAAAAAAA (SEQ ID NO 005)

#6: Biotin-5’-

TCGTCGGCAGCGTCGCGAACTGTGTCTTTATCCGAAAAAAAAAAAAAAAAAAAAAAAAAAAA (SEQ ID NO 006)

#7: Biotin-5’-

TCGTCGGCAGCGTCGCGCATAGTAGTAATGAAGCAAAAAAAAAAAAAAAAAAAAAAAAAAAA (SEQ ID NO 007) bc2tem plate:

5’-

GTCTCGTGGGCTCGGTAGTACGCATAGTACGCAATCATCGCAATCATCGCACTTGACGCACTTGACGC

_{TTTTTTTTTTTTTTTTTTTTTTTTTTTT} (SEQ ID NO 008)

P04-CN VK-bc#1 -3 (P04 = phosphorylation; K = CNVK)

#1 : P04-5’-KGT ACT A

#2: P04-5’-KGATGAT

#3: P04-5’-KGTCAAG

bioparticle#1-7:

#1 : Bi0tin-5’-TCGTCGGCAGCGTCAGGTACATATACTTCCACCGCCGAGCCCACGAGAC (SEQ ID NO 009)

#2: Bi0tin-5’-TCGTCGGCAGCGTCGTGGATACACACTTCACATCCCGAGCCCACGAGAC (SEQ ID NO 010)

#3: Biotin-5’-TCGTCGGCAGCGTCTTAGATCGGATTACACGCTCCCGAGCCCACGAGAC (SEQ ID NO 011)

#4: Biotin-5’-TCGTCGGCAGCGTCTTAACGGGTCGTAACACGGCCCGAGCCCACGAGAC (SEQ ID NO 012)

#5: Bi0tin-5’-TCGTCGGCAGCGTCTCGAGCTGTGCATATAGTGCCCGAGCCCACGAGAC (SEQ ID NO 013)

#6: Biotin-5’-TCGTCGGCAGCGTCGCGAACTGTGTCTTTATCCGCCGAGCCCACGAGAC (SEQ ID NO 014)

#7: Biotin-5’-TCGTCGGCAGCGTCGCGCATAGTAGTAATGAAGCCCGAGCCCACGAGAC (SEQ ID NO 015)

biobctemplate:

Bi0tin-5’-AAACGGGTGCAAAATACGGTGTCAAATGCAGGTAGG (SEQ ID NO 016)

CNVK-bc-ext#1-3 (K = CNVK)

#1 : GCCTTAGATAAAGTTACGCGAATGCA CCGT (SEQ ID NO 017)

#2: CTCACTAGGTATGTGTAAGGAAGACA CGTA (SEQ ID NO 018)

#3: CGTGCTGTCAAGTATTGTCGAACCTA CTGCA (SEQ ID NO 019)

bc-amp#1-3 #1 : FITC-TCGTCGGCAGCGTCCGCGTAACTTTATCTAAGGCCCGAGCCCACGAGAC (SEQ ID NO 020)

#2: TCGTCGGCAGCGTCCCTTACACATACCTAGTGAGCCGAGCCCACGAGAC (SEQ ID NO 021) #3: TCGTCGGCAGCGTCCGACAATACTTGACAGCACGCCGAGCCCACGAGAC (SEQ ID NO 022)

Claims

1. A method for determining a location of a target structure, said method comprising the steps

a. providing a sample comprising the target structure;

b. contacting said sample with a binding molecule,

- wherein said binding molecule is coupled to a recorder nucleic acid

molecule characterized by a recorder sequence that comprises a plurality of individual sequence tags (ISTs);

c. contacting said sample with a first barcode nucleic acid molecule;

- wherein said first barcode nucleic acid molecule specifically hybridizes to a first 1ST comprised in said recorder sequence;

d. exposing a first area within said sample to electromagnetic radiation of a

specific wavelength,

- wherein said first barcode nucleic acid molecule under conditions of

electromagnetic radiation exposure of said specific wavelength is able to introduce a modification to said recorder nucleic acid molecule within said first 1ST;

f. repeating step c to e for each of the plurality of ISTs, thereby exposing a different unique area of the sample to the electromagnetic radiation of the specific wavelength in each repetition;

g. extracting said recorder nucleic acid molecule from said sample;

h. determining whether said ISTs of said recorder nucleic acid molecule carry said modification, thereby determining the location of said target structure.

2. The method according to claim 1 , wherein said modification of said recorder nucleic acid molecule in step d is a covalent bond of said recorder nucleic acid molecule to said barcode nucleic acid molecule.

3. The method according to any one of claims 1 or 2, wherein in step b, a plurality of different binding molecules are contacted with the sample, each binding molecule capable to specifically bind to a different target structure.

4. The method according to claim 3, wherein each recorder nucleic acid molecule

comprises a sequence characteristic for said binding molecule.

5. The method according to any one of the preceding claims, wherein said

electromagnetic radiation is light of a specific wavelength, particularly light of a wavelength of 300-600 nm, particularly 310-490 nm, more particularly 340-420 nm.

6. The method according to any one of the preceding claims, wherein said barcode nucleic acid molecule comprises a chemical entity selected from a pyranocarbazole group, a psoralen-coupled nucleoside, a diazirine-based nucleoside analogue, an aryl-azide-based nucleoside analogue, a 5-iodouracil, a 6-thioguanine, a 4- thiothymidine, a coumarin-modified thymidine, a 3-methoxycarbonylcarbazole nucleoside, a 3-carboxyvinylcarbazole nucleoside, a 3-carbonylamidevinylcarbazole nucleoside, and a 3-cyanovinylcarbazole nucleoside, particularly a vinylcarbazole nucleoside.

7. The method according to any one of the preceding claims, wherein electromagnetic radiation exposure in step d is performed for < 60 sec, particularly for £ 10 sec.

8. The method according to any one of the preceding claims, wherein after step f or in step h, an enzyme, particularly a methyltransferase or a deaminase, is employed to modify the 1ST when bound to a barcode nucleic acid molecule.

9. The method according to any one of claims 1 to 7, wherein in step h, said barcode nucleic acid molecule is degraded except for a cross-linked subunit of the barcode nucleic acid molecule, particularly via enzymatic, chemical or photochemical degradation, leaving behind a covalent modification of each 1ST which was bound to a barcode nucleic acid molecule, particularly

the barcode nucleic acid molecule is composed of RNA nucleotides which are selectively cleaved; or

the barcode nucleic acid molecule has a phosphorothioate backbone which is selectively broken; or

DNA repair enzymes are used to deglycosylate cross-linked bases between the barcode nucleic acid molecule and the 1ST.

10. The method according to any one of the preceding claims, wherein in step h,

nanopore-based sequencing, Single Molecule Real-Time (SMRT) sequencing, sequencing by hybridization, or the sequencing by expansion method is employed to sequence said recorder nucleic acid molecule,

particularly wherein sequencing by hybridization is employed to sequence said recorder nucleic acid molecule.

11. The method according to any one of the preceding claims, wherein said binding molecule is an antibody or antibody fragment, an aptamer or an antibody-like molecule and said target structure comprises an epitope.

12. The method according to any one of claims 1 to 10, wherein said binding molecule is a nucleic acid molecule and said target structure comprises a nucleic acid sequence being able to hybridize with said binding molecule.

13. The method according to any one of the preceding claims, wherein said sample is a cell culture, a tissue slide, a viral particle, a biofilm, or a natural or artificial particle or surface.

14. The method according to claim 6, wherein each barcode nucleic acid molecule

comprises at least two of said chemical entities.

15. The method according to any one of the preceding claims, wherein step g comprises the following steps:

i. segregating said sample into a plurality of compartments, wherein each compartment comprises a plurality of compartment nucleic acid molecules, wherein each compartment nucleic acid molecule comprises a. a compartment-specific index sequence, wherein all compartment- specific index sequences of one compartment have the same sequence, and each compartment-specific index sequence of one compartment differs from any compartment-specific index sequence of another compartment;

b. optionally, a unique molecule identifier sequence (UMI), wherein each UMI differs from any other UMI of the same particle; and c. optionally, a primer binding sequence, wherein said primer binding sequence is able to hybridize to a PCR primer under annealing conditions;

ii. linking a compartment nucleic acid molecule to each barcode nucleic acid molecule to yield a plurality of hybrid nucleic acid molecules; iii. optionally, amplifying said plurality of hybrid nucleic acid molecules;

16. The method according to claim 15, wherein step iii is performed via PCR.

17. The method according to claim 15, wherein step iii is omitted and in step h of claim 1 , single-molecule sequencing is performed.

18. The method according to claim 15 to 17, wherein step ii is performed via primer

extension, ligation, taqmentation or hybridization.

19. The method according to any one of claims 15 to 18, wherein before step i, the

sample is segregated into single cells and the cells are pre-sorted, particularly via FACS or affinity chromatography.

20. A kit comprising a binder molecule as described in claim 1 , a recorder nucleic acid as described in claim 1 , and at least one photoreactive barcode nucleic acid molecule as described in claim 1.