JP7155281B2 - Cell information processing method - Google Patents

Description

The present invention relates to a cell information processing method. In particular, to analyze the interaction between different cell types that occurs when cell stimulation with drugs is given to cell populations cultured under conditions where different cell types are mixed, and the mechanism of action Regarding the method.

A large number of cell types exist in the body, coexist while being dependent on each other, and change due to mutual stimulation. In addition, when it changes, many genes change in one cell while having a dependent relationship, and the gene exerts its function in vivo through this dependent relationship. Therefore, in order to understand life phenomena and biological phenomena in the fields of drug discovery and regenerative medicine, it is important to analyze the mechanisms of interactions between cells or between cell groups and molecular behavior within cells. For example, when multiple types of cells are mixed and cultured, how are the cells mutually stimulated? By analyzing whether it changes over time, we can obtain clues for elucidating the mechanisms of life phenomena.

As a method that can use the above mechanism for analysis, Patent Document 1 describes "(1) a step of calculating a feature amount reflecting the similarity between each data from data indicating changes in the expression level of genes over time; (2) calculating the eigenvectors of the similarity matrix M for all inter-gene combinations from the calculated feature values; and (4) clustering each data based on the Boolean matrix N.” That is, based on the similarity of expression level changes over time, a plurality of genes are clustered. A grouping method is proposed (claim 9).

Also, US Pat. No. 6,200,000 discloses a method for clustering gene expression profiles, comprising: selecting one or more GO terms from a gene ontology (GO) tree; receiving a gene expression data set; classifying the set into groups according to the GO terms; first, clustering the gene expression data belonging to each group based on the similarity of the gene expression data; and second, seeding the results of the first clustering. clustering gene expression data sets using "a method of grouping a plurality of genes based on similarity of their expression data and similarity of the biological functions in which they are implicated. has been proposed (claim 1).

In addition, as a conventional method for analyzing drug efficacy and pharmacology, as shown in FIG. Then, a method of analyzing the drug efficacy using the averaged data of each cell population, or a method of giving a drug to a cell population containing multiple cell types extracted from an organ and culturing it, and measuring the drug efficacy of the averaged data of the cell population has been used.

However, even cell populations consisting of allogeneic cells may not show the dynamic behavior of individual cells over time (e.g., cell activation, cell inhibition, cell interaction, protein expression, protein secretion, cell proliferation, cell morphology, etc.). changes in cells, cell death, etc.), conventional methods have not been able to perform sufficiently high-precision analysis. Therefore, in recent years, in order to further improve the accuracy of analysis, cells that constitute a cell population consisting of only the same type of cells are converted to single cells, and analysis at the individual cell (single cell) level (single cell analysis) is performed. ing. Such single-cell analysis has made it possible to analyze individual cells, rather than analyzing averaged data of cell populations, so that cells and small dynamic behaviors that could not be detected by analysis with averaged data can be analyzed. changes can be detected. Such single-cell analysis is used, for example, to classify subspecies of cells present in cancer tissues by single-cell transcriptome analysis of cancer cells. Moreover, the method described in Patent Document 3 has been proposed as a method for evaluating cell activity using single-cell analysis.

Patent Document 3 discloses a method for evaluating cell activity, which comprises (a) placing a cell population containing different types of cells on a region, that is, placing different types of cells in each nanowell on a nanowell array plate. (b) assaying the dynamic behavior of the cell population as a function of time, i. (i.e., by image analysis) and measured over time; (c) identifying at least one cell to be analyzed from the cell population based on dynamic behavior; (d) identifying Characterizing the molecular profile of the identified at least one analyzed cell, i.e., performing mass spectrometry, genetic analysis, protein analysis, etc. at the single cell level for the identified at least one analyzed cell, transcriptional activity, trans Cryptomic profile, gene expression activity, genomic profile, protein expression activity, proteomic profile, protein interaction activity, cellular receptor expression activity, lipid profile, lipid activity, carbohydrate profile, microvesicle activity, glucose activity, metabolism and (e) correlating the information obtained from steps (b) and (d) (Claim 1). .
That is, the method described in Patent Document 1 evaluates cell activity of individual cell types in a cell population containing different types of cells.

JP 2010-157214 A U.S. Patent Application Publication No. 2009/0112480 Japanese Patent Publication No. 2018-514195

However, both of the methods described in Patent Documents 1 and 2 are methods for investigating intracellular networks, and cannot know interactions between cells and between cell types.

Moreover, the method described in Patent Document 3 is also a method only for evaluating the activity of each cell, so it is not possible to know interactions between cells and between cell types.
In addition, in the method described in Patent Document 3, humans observe the dynamic behavior of a cell population consisting of at least 100 to 200 cells by image analysis, and multiple types of cells and cells with different dynamic behavior To identify and select single cells to be analyzed from among them, and to analyze the mechanism of action of the dynamic behavior of single cells selected by humans and intercellular interactions, that is, each Humans arbitrarily judge the morphology of cell types and the dynamic changes of each cell over time to identify and select cells for analysis, so there is a problem that highly accurate analysis results cannot be obtained. rice field.
Further, as described as one embodiment of Patent Document 1, when immune cells are contained in the cell population to be analyzed, specifically, when immune cells are contained in a blood sample, the blood sample contains a drug There is a circumstance that the cells die in about 24 hours after giving In addition, the analysis method described in Patent Document 1 is to observe the drug-applied cell population over time and perform the analysis (that is, the observation of one sample is continued until the analysis is completed, Furthermore, since it is a method of analyzing while repeatedly selecting and culturing the cells to be analyzed), when the method described in Patent Document 3 is employed, all cells within at least 24 hours after giving the drug to the cell population There is a very tight time constraint to finish the analysis.

In addition, the method described in Patent Document 3 requires time and effort to identify cells and select target cells, and the calculation method and results of the evaluation by the analysis are complicated, so it is efficient and highly accurate. Moreover, it is not practical as an industrial method that requires rapid evaluation and analysis.
Furthermore, a method for identifying each cell type, which is performed in a cell population containing multiple different cell types by giving a drug to each cell population containing different types of cells and culturing them, and using single cell analysis technology, each cell At present, no methods have been reported to efficiently analyze the efficacy and pharmacology of species. Therefore, single-cell analysis will be used to investigate interactions between different cell types that occur when cell stimulation with drugs is applied to cell populations cultured in a mixture of different cell types, and their mechanisms of action. At present, no efficient analysis method has been reported.

Further, in the initial stage of searching for drugs using the efficacy and pharmacological screening method as shown in FIG. 8, interaction analysis with different cells and Since the analysis of the mechanism of drug efficacy and pharmacology is proceeding, in the initial stage of analysis when substances that can be used as drugs are extracted from a huge number of candidate substances, even if cell death is detected, for example, what kind of death will it be? Depending on the factors, it was possible that the cell death would not have been triggered and that it would not become a drug candidate.

In view of the above problems of the prior art, the present invention provides a cell information processing method for efficiently analyzing interactions between cell populations when different types of cell populations coexist. With the goal.
More specifically, the interaction between different cell types that occurs when a drug or cell stimulus is applied to a cell population cultured under conditions where different cell types coexist, and its mechanism of action. To provide a method capable of efficiently evaluating and analyzing with a simpler operation, with high precision and speed, and which can also be used on an industrial scale.

In the cell information processing method of the present invention, at least two or more predetermined containers seeded with target cell groups containing a plurality of different types of cells are prepared, and the target cell groups are stimulated at two or more time points after predetermined cell stimulation. , a cell recovery step of recovering all cells in one container, a single cell conversion step of converting all the cells collected at each time point into single cells, and single cell data from each single cell converted at each time point. All the cells collected at each time point are grouped into a plurality of cell populations having a common first cell characteristic based on the single cell data acquisition step to be acquired and the single cell data at each time point to form a two-dimensional plane Alternatively, a grouping step of plotting on a three-dimensional space and performing a process of identifying the cell type of each grouped cell population based on the second cell feature to obtain the clustering result at each time point; A cell change detection step of detecting a change over time in the cell population of the same cell type by comparing the clustering results in the above; By comparing the clustering results at the two or more time points along with the mechanism of action analysis step of analyzing the mechanism of action and the results of the mechanism of action analysis, the dependency relationship between clustering is analyzed, and different cell types and an interaction analysis step of analyzing interactions that occur between the cell populations.

The interactions occurring between the cell populations in the interaction analysis step preferably include interactions occurring between cell populations of the same cell type.
In the cell change detection step, by comparing the clustering results at each time point, changes over time in the number of cells constituting the cell population of the same cell type and changes in the first cell characteristics over time are detected. can also be detected.
The cell change detection step further detects changes over time in the number of cells constituting the cell population of the same cell type and changes in the first cell characteristics over time in the clustering results at each time point. good too.

The cell recovery step further includes preparing a target cell group containing the plurality of types of cells to be cultured without the predetermined cell stimulation, and at one or more time points, all the cells in the one predetermined container. is collected, and in the cell change detection step, by comparing the clustering results at each time point, changes over time in the cell population of the same cell tumor due to the presence or absence of the predetermined cell stimulation of the cell population are evaluated. You can also
The cell stimulation is preferably at least one selected from the group consisting of chemical stimulation and physical stimulation.
Preferably, the chemical stimulation is by addition of an agent that induces a biological response to cells.
The mechanism of action is preferably a specific biochemical reaction or interaction for exerting a biological phenomenon induced inside or outside the cell by the cell stimulation.

The above single-cell data consists of gene DNA sequence information (genome), epigenetic information that controls gene expression (DNA methylation, histone methylation, acetylation, phosphorylation), gene primary transcripts (mRNA, Translated RNA, microRNA, etc.) information (transcriptome), protein translation amount and modification information such as phosphorylation, oxidation, and saccharification, amino acid sequence information (proteome), metabolite information (metabolome), intracellular hydrogen ion concentration index (pH), intracellular ATP concentration, ion concentration (calcium, magnesium, potassium, sodium, etc.), and intracellular temperature.
The single cell data are preferably gene expression levels and gene DNA sequences.

In the grouping step, the first cell feature is preferably obtained by reducing the n-dimensional cell feature included in the single cell data to two or three dimensions.
The above dimensionality reduction methods are selected from the group consisting of Principal Component Analysis (PCA), Principal Component Analysis with Kernel (Kernel-PCA), Multidimensional Scaling (MDS), t-SNE, and Convolutional Neural Network (CNN). At least one is preferably selected.
In the grouping step, the first cell feature is preferably obtained by performing principal component analysis on the gene expression level and reducing the dimensions to two or three dimensions.
In the grouping step, the second cell feature is preferably at least one piece of cell information that enables identification of each cell type from cell function or cell state.

The above cell information includes gene DNA sequence information (genome), epigenetic information that controls gene expression (DNA methylation, histone methylation, acetylation, phosphorylation), gene primary transcripts (mRNA, Translated RNA, microRNA, etc.) information (transcriptome), protein translation amount and modification information such as phosphorylation, oxidation, and saccharification, amino acid sequence information (proteome), metabolite information (metabolome), intracellular hydrogen ion concentration index (pH), intracellular ATP concentration, ion concentration (calcium, magnesium, potassium, sodium, etc.), and intracellular temperature.
The cell function is preferably at least one selected from cell proliferation, repair, metabolism, and information exchange between cells.
The cell state is preferably at least one selected from gene expression status, protein expression status, and enzyme activity.

In the interaction analysis step, the interaction that occurs between the cell populations is a biological or physical action that occurs between cell populations of different cell types, or the number of cells due to the action, the first cell A change in characteristics is preferred.
Preliminarily creating time-series data obtained by acquiring values representing amounts or states of a plurality of biological substances from the single-cell data at a plurality of points in time for each biological substance, and changing the time-series data for each biological substance with time; Based on the similarity of the biological functions of each biological substance, the cells for which single-cell data was obtained are grouped into cell populations having a common first cell characteristic, and in the interaction analysis step, the plurality of generating a value representing the state of the cell population from one or more first cell characteristics contained in each of the plurality of cell populations for each of the time points; It is preferable to estimate the dependence of the state between cell populations from a data set consisting of time-series data obtained at a plurality of time points for each biological material.
The similarity of biological functions includes having a common gene ontology, belonging to a common canonical pathway, having common upstream factors, being involved in a common phenotype, and being involved in a common disease. It is preferably evaluated based on at least one selected from the group consisting of.

In the cell collection step and the single cell generation step, methods for collecting all cells and converting them into single cells include manual operation, flow cytometry, magnetic separation, laser capture microdissection, microfluidics, microdroplets, nanowells, Preferably, the method uses at least one selected from the group consisting of micropipette fine needle aspiration, laser tweezers, labeled arrays, surface plasmon response, and nanobiodevices.
In the above-described method of collecting whole cells and converting them into single cells, it is preferable to use at least one selected from the group consisting of fluorescent labeling, radioisotope labeling, antibody labeling, and magnetic labeling as the cell label.
The target cell group containing multiple types of cells is preferably cells obtained from at least one selected from the group consisting of biological tissue samples, blood samples, culture samples, and environmental samples.
The plurality of types of cells are preferably at least one selected from the group consisting of animal cells, plant cells, fungal cells and bacterial cells.

According to the present invention, it is possible to provide a cell information processing method for efficiently analyzing interactions between different cell populations when multiple types of cell populations coexist.
More specifically, identification of each cell type performed in a cell population in which different cell types exist, or drug efficacy and pharmacological screening can be evaluated and analyzed more easily, accurately and quickly. It is possible to provide a method that can be used on an industrial scale. Therefore, not only changes between cell populations of the same cell type but also changes between cell populations of different cell types (interactions between cell populations) can be efficiently and simultaneously confirmed. In addition, as a result, it is possible to analyze wider interactions between cells and between cell types (reciprocal networks) with high accuracy and ease.
In addition, since humans do not arbitrarily judge the dynamic behavior of cells over time, the identification of each cell type and the mechanism of action of drugs or cell stimulation on various cells should be analyzed with high accuracy. can be done.
In addition, target cells can be identified without taking time and effort such as image analysis.
In addition, even when performing drug efficacy screening for immune cells, analysis can be performed without particularly limiting the analysis time.
In addition, it is not necessary to apply a drug or cell stimulus to each sample cultured for each cell type for analysis. Therefore, it is possible to reduce the labor and expense of preparing and analyzing a large amount of samples.
In addition, even if the cells to be analyzed cannot be cultured with only one type of cell for cell growth and cell function maintenance, according to the present invention, a sample obtained by co-cultivating a plurality of types of cells Cells to be analyzed can be identified. In addition, it is possible to appropriately analyze the mechanism of action by co-culturing with cells other than the analysis target, for example, the mechanism of action of a sample drug or cell stimulation that requires co-culturing of analysis target cells and immune cells.

FIG. 1 is a flow chart for explaining the method for information processing of cells according to the present invention. FIG. 2A shows cells immediately after cell stimulation (time point 0). FIG. 2B shows cells 1 hour after cell stimulation (time point 1). FIG. 2C shows cells 6 hours after cell stimulation (time point 2). FIG. 3A shows the results of clustering according to FIG. 2A (time point 0). FIG. 3B shows the results of clustering according to FIG. 2B (time point 1). FIG. 3C shows the results of clustering according to FIG. 2C (time point 2). FIG. 4 is a diagram showing temporal changes in cell population based on FIG. 3B. FIG. 5 is a diagram showing temporal changes in cell population based on FIG. 3C. FIG. 6 is a diagram showing the results of measuring the amount of each component of nerve cells and glial cells over time. FIG. 7 is a diagram showing temporal changes in cell population based on FIGS. 3C and 6. FIG. FIG. 8 is a diagram for explaining a conventional drug screening method.

[Cell information processing method]
Below, the cell information processing method of the present invention will be described in detail.
Embodiment 1
Embodiment 1 is an analysis of the interaction between neurons and glial cells, that is, the mutual network of neurons and glial cells generated when a drug is added to a target cell group consisting of neurons and glial cells and cultured. Analysis of the formation process (change over time) will be explained as an example.
FIG. 1 is a flow chart showing a cell information processing method according to Embodiment 1 of the present invention.

<Cell collection step (S1)>
First, in this step, as shown in FIG. 2A, at least two predetermined containers 1 seeded with different types of cell groups (cell groups consisting of glial cells 2 and nerve cells 4) are prepared. Here, a cell group seeded in a predetermined container 1 is called a target cell group. Next, the target cell group seeded in the prepared container 1 is subjected to cell stimulation by adding a drug and cultured at two or more time points (for example, immediately after drug addition, after drug addition, 1 hour, 6 hours, 12 hours, 24 hours, 48 hours, 72 hours...) Harvest all cells in one container.

FIG. 2A shows cells immediately after cell stimulation (time point 0), FIG. 2B shows cells one hour after cell stimulation (time point 1), and FIG. 2C shows cells after cell stimulation , shows cells after 6 hours (time point 2). In this step, all cells in container 1 are collected at each time point.
Time 0 is the state before the shapes of astrocytes 2 and nerve cells 4 change, and at time 1, the shape of astrocytes 2 changes to reactive astrocytes 10 due to cell stimulation by addition of a drug. At time 2, the shape of the nerve cell 4 is changed due to the contact between the changed reactive astrocyte 10 and the nerve cell 4, and the shape of the nerve cell 4 is changed to include the nerve cell 12. FIG.

《Target cell group》
Here, the target cell group is a cell group seeded in a predetermined container 1 and means a cell group containing a plurality of different types of cells. More specifically, "a group of cells containing different types of cells", for example, a cell group composed of human iPS cells and mouse embryonic fibroblasts, from a plurality of cells of different cell types It means a cell group composed of cells, and may be a cell group composed of cells derived from the same species or a cell group composed of cells derived from a different species.
In this embodiment, glial cells and nerve cells are described as cells constituting the target cell group, but the types of cells contained in the target cell group are not particularly limited.

Examples of the "target cell group containing multiple types of cells" include biological tissue samples, blood samples, culture samples, and environmental samples.
Examples of the biological tissue samples include mouse brain tissue and human resected tumor tissue.
The blood sample is exemplified by a human blood sample.
Examples of the above-mentioned culture samples include co-culture samples of human iPS cells and fetal mouse fibroblasts.
Examples of the environmental samples include soil samples and water samples collected from submarine hydrothermal vents.

Examples of "cells" included in the "target cell group containing multiple cell types" include animal cells, plant cells, fungal cells, and bacterial cells.

Examples of the animal cells include cells of vertebrates, chordates (excluding vertebrates) or insects.
Examples of the vertebrates include mammals such as humans, chimpanzees, rhesus monkeys, dogs, pigs, mice, rats, Chinese hamsters, and guinea pigs, Xenopus, zebrafish, medaka, and tiger puffer.
The mammalian cells (mammalian cells) include, but are not limited to, tumor cells, hepatocytes, fibroblasts, stem cells, and immune cells.
The chordates (excluding vertebrates) are exemplified by sea squirts.
Examples of the insects include fruit flies, silkworms, tobacco hawksmoths, and honeybees.

Examples of the plant cells include angiosperm cells.
Examples of the angiosperms include Arabidopsis thaliana, rice, wheat, chamomile, Lotus japonicus, and tobacco.

The fungal cells are exemplified by fungal or yeast cells.
Examples of the mold include Neurospora crassa, Aspergillus oryzae, Aspergillus fumigatus, Aspergillus nidulans, Rhizopus oryzae, and Mucor circinelloides. exemplified.
Examples of the yeast include Saccharomyces cerevisiae, Schizosaccharomyces pombe, Candida albicans, Cryptococcus neoformans, and Trichosporon ovoides.

The bacterial cells include, for example, cells of Escherichia coli, Salmonella enterica, Clostridium difficile, or Bacillus anthracis.

Cell stimulation is not particularly limited as long as it is at least one selected from the group consisting of chemical stimulation (chemical substances, etc.) and physical stimulation (light, heat, pressure, etc.).

Examples of such chemical stimulation include addition of agents that induce a biological response to cells. The drug may be one that induces a biological reaction in cells by adding it to the biological sample.
Specific examples of the biological reaction include proliferation, cell death, differentiation, antigen-antibody reaction, secretion of growth factors, and the like.
The drugs mentioned above are not particularly limited as long as they are drugs intended to have a measurable effect on the structure and function of the body. Such agents include pharmaceuticals such as anticancer agents, growth factors, cytokines and low-molecular-weight drugs. Specific examples of growth factors include epidermal growth factor (EGF). Specific examples of are tumor necrosis factor (TNF-α), interleukin-1β (IL-1β), insulin, glucagon-like peptide-1 (GLP-1), imatinib, acetaminophen, adalimumab , and Nivolumab.

Note that the container 1 for containing the target cell group is not particularly limited as long as it is a cell culture container that can contain and culture the target cell group. Also, the culture medium to be used can be appropriately selected according to the cells and analysis method.
Each container contains a target cell group composed of a plurality of cell types at a predetermined ratio (for example, A cells, B cells, and C cells, and the number of each cell is A cells: B cells: C cells = 2: A 1:1 target cell population) is seeded. A target cell group containing a plurality of cell types is cultured for a predetermined period of time, and then a drug is added to provide cell stimulation.

<Single cell conversion step (S2)>
Next, in this step, the target cell group collected in the cell collecting step (S1) is converted into single cells (single-cell conversion).

There are no particular limitations on the method and equipment for converting a target cell group into single cells and collecting single cells, and known methods and equipment can be used. For example, known methods include manual, flow cytometry, magnetic separation, laser capture microdissection, microdroplet method, micropipette fine needle aspiration method, and surface plasmon resonance method, and known instruments include: Examples include microchannels, nanowells, laser tweezers, labeling arrays, and nanobiodevices.
Among these known methods, the microdroplet method, microchannel, nanowell, and flow cytometry are preferably used. This is because a large amount of cells can be separated and collected at high speed without requiring skill in single cell processing, and analysis accuracy can be improved.

When collecting single cells, it is preferable to label each cell using fluorescent labeling, radioisotope (RI) labeling, antibody labeling, and magnetic labeling. This is because it can be used to identify the cell type of each cell population in the grouping step (S4) described later. In FIGS. 2A-2C, astrocytes 2 are labeled with a fluorescent label 6 and neurons 4 with a fluorescent label 8 . In the figure, only one astrocyte 2 and one nerve cell 4 are labeled with fluorescence, but all cells are labeled with fluorescence.
In particular, a combination of an antibody that binds to a protein expressed on the surface of a cell and a fluorescence, RI label, or magnetic label is preferred because the specificity of the antibody increases.

An automated solution for single-cell genome research such as C1 ^™ Single-Cell Auto Prep System (manufactured by Fluidigm) can also be used. This solution can automatically perform single cell isolation, cell labeling, cell lysis, and extraction of genomic DNA or total RNA in the single cell data acquisition step (S3) described later. For example, when using genomic DNA or total RNA to obtain single-cell data, work efficiency can be further improved.

<Single cell data acquisition step (S3)>
Next, in the single cell data acquisition step (S3), DNA sequence information of genes and gene expression levels are acquired as single cell data from each cell collected at each time point and made into a single cell. Single-cell data is acquired for each of all single-celled and collected single cells.
The “gene expression level” in the present invention is the amount of mRNA, which is the transcription product of a gene, and can be measured by examining the expression state of the gene by gene expression analysis. Alternatively, the amount of protein, which is a gene expression product, may be analyzed.

《Single cell data》
Note that single cell data means information indicating the function, property, and state of a single cell, and is not limited to the gene expression level and DNA sequence information described above. For example, DNA sequence information of genes (genome), epigenetic information that controls gene expression (DNA methylation, histone methylation, acetylation, phosphorylation), gene primary transcripts (mRNA, non-translated RNA, micro RNA, etc.) information (transcriptome), protein translation amount and modifications such as phosphorylation, oxidation, and saccharification, amino acid sequence information (proteome), metabolite information (metabolome), intracellular hydrogen ion concentration index (pH), cells Intracellular ATP concentration, ion concentration (calcium, magnesium, potassium, sodium, etc.), intracellular temperature, etc. may be acquired as single-cell data.
In addition, when collecting single cells, if each cell is labeled with a fluorescent substance, an antibody, or the like, the information thereof can also be obtained as single data.

<Grouping step (S4)>
Next, in the grouping step (S4), based on the single cell data (gene expression level data, DNA sequence and fluorescent labeling) obtained in the single cell data obtaining step, the cells from which the single cell data was obtained are grouped into a common first Grouped into cell populations having one cell characteristic, and further grouped by the first cell characteristic based on a second cell characteristic that can discriminate each cell type contained in the single cell data Identify the cell type of the cell population.
Specifically, using principal component analysis, n (n-dimensional) cell features contained in the gene expression level data are compressed to two or three dimensions that can be visualized. and plotted on a two-dimensional plane or three-dimensional space (that is, each principal component corresponds to the "first cell feature"). Furthermore, based on the base sequences and fluorescent labels (second cell characteristics) of the cells constituting each cell population, identification processing (clustering analysis) of the cell type of each grouped cell population is performed.
Here, it is preferable to search for an axis that maximizes the variance of data (random variables) as the principal component axis for the principal component analysis.

By this step, the cell characteristics of each cell population and the number of cells are associated and visualized, so that the relationship between the number of cells and the cell characteristics can be easily confirmed or detected at the same time with high accuracy.

The number of cell features used for grouping all collected cells into a plurality of cell populations is not particularly limited. One or more of the n cell features may be used to group cells.
In addition, in this embodiment, principal component analysis is used to visualize cells by dimensional reduction to two or three dimensions, and cells for which single cell data are obtained are grouped into cell populations having a common first cell characteristic. Dimensional reduction methods include, but are not limited to, principal component analysis (PCA), principal component analysis with kernel (Kernel-PCA), multidimensional scaling (MDS), t-SNE, Alternatively, a convolutional neural network (CNN) can be used.
In addition, from the single-cell data (information related to multiple biological substances) of each cell, values representing the amount or state of multiple biological substances are prepared in advance as time-series data obtained at multiple time points for each biological substance. , Based on the temporal change of the time-series data for each biomaterial and the similarity of the biological functions of each biomaterial, the cells for which single-cell data was acquired are grouped into cell populations having a common first cell characteristic. You can divide.
Here, the similarity of biological functions includes having a common gene ontology, belonging to a common canonical pathway, having common upstream factors, being involved in a common phenotype, and being involved in a common disease. It is preferably evaluated based on at least one selected from the group consisting of involvement.

FIG. 3A shows (principal components and) clustering results based on single-cell data obtained from the target cell group immediately after cell stimulation in FIG. 2A (time point 0), and FIG. Shows (principal components and) clustering results based on single-cell data obtained from the target cell group at time point 1), and FIG. (Principal components and) clustering results based on cell data.
Principal component analysis is performed on the single cell data (gene expression level data) obtained from the target cell group immediately after cell stimulation (time point 0) in FIG. and 16 (Fig. 3A), and based on the single-cell data (DNA sequence and fluorescent labels 6 and 8), the cell types that make up cell population 14 are astrocytes 2, making up cell population 16. The cell type is identified as neuronal 4.
Similarly, the single cell data (gene expression level data) obtained from the target cell group one hour after the cell stimulation in FIG. Cells were grouped into multiple cell populations 18, 20 and 22 (Fig. 3B), and based on single cell data (DNA sequences and fluorescent labels 6 and 8), the cell types that make up cell populations 18 and 20 were determined. Astrocytes 2 and the cell types that make up the cell population 22 are identified as nerve cells 4 .
Principal component analysis was performed on the single cell data (gene expression level data) obtained from the target cell group 6 hours after the cell stimulation in FIG. 2C (time point 2). is grouped into multiple cell populations 24 and 26 (Fig. 3C), and based on single cell data (DNA sequences and fluorescent labels 6 and 8), the cell type that constitutes the cell population 24 is astrocyte 2. , identify that the cell types that make up the cell population 26 are nerve cells 4 .

In this embodiment, the cell type of each cell population was identified using DNA sequences and fluorescent labels as the second cell characteristics, but the invention is not particularly limited to these. Distinguish each cell type from cell function (e.g., cell proliferation, repair, metabolism, and information exchange between cells) or cell state (e.g., gene expression status, protein expression status, and enzyme activity) Cellular information (included in single data), such as gene DNA sequence information (genome), epigenetic information that controls gene expression (DNA methylation, histone methylation, acetylation, phosphorylation) , gene primary transcript (mRNA, untranslated RNA, microRNA, etc.) information (transcriptome), protein translation amount and modifications such as phosphorylation, oxidation, and glycation, amino acid sequence information (proteome), metabolite information ( metabolome), intracellular hydrogen ion concentration exponent (pH), intracellular ATP concentration, ion concentration (calcium, magnesium, potassium, sodium, etc.), intracellular temperature, and the like can be used.
The cell information that can distinguish each cell type includes not only information such as the genes, proteins, and metabolites inherent in the cell, but also genes introduced from outside the cell (for example, immortalizing genes). , proteins and metabolites, and information on organic matter. Examples of immortalizing genes include the hTERT gene (human telomerase reverse transcriptase gene) and SV40T antigen (simian virus 40T antigen gene). In addition, when each cell is labeled with a fluorescent substance, an antibody, or the like when recovering single cells, the information thereof is also included.

<Cell change detection step (S5)>
Next, the cell change detection step (S5) is based on the clustering results obtained in the grouping step (S4). By comparing with the clustering results of the cells obtained, changes in the cell population of the same cell type over time (changes in real time or changes in pseudo time estimated from the changes) are detected. In addition, it is preferable to extract the cell characteristics that have changed over time, and to detect the amount of change over time as a numerical value.
Here, "time-dependent change in cell population" refers to the number of cells, cell characteristics (first cell characteristics), or other cell characteristics (i.e., DNA sequence information of genes (genome), gene expression Epigenetic information (DNA methylation, histone methylation, acetylation, phosphorylation), gene primary transcript (mRNA, untranslated RNA, microRNA, etc.) information (transcriptome), protein translation amount and phosphorylation , oxidation, glycation, etc., amino acid sequence information (proteome), metabolite information (metabolome), intracellular hydrogen ion concentration index (pH), intracellular ATP concentration, ion concentration (calcium, magnesium, potassium, sodium, etc.) , intracellular temperature, etc.), and changes in the biological material of the cell over time.

3A (time point 0) and FIG. 3B (time point 1), which are the clustering results obtained in the grouping step (S4), are compared, and the cell features of the cell population composed of cells of the same cell type (first cell feature Detect changes in principal components 1 and 2). Specifically, a comparison of cell population 14 identified as astrocyte 2 in FIG. 3A (time point 0) and cell populations 18 and 20 identified as astrocyte 2 in FIG. 3B (time point 1), and FIG. By comparing the cell characteristics of the cell population 16 identified as the nerve cell 4 in 3A (time point 0) with the cell population 22 identified as the nerve cell 4 in FIG. 3B (time point 1), the cell characteristics of each cell population Detects the presence or absence of changes over time.
Specifically, the cell characteristics of the cell population 14 identified as astrocytes 2 in FIG. 3A (time point 0) (primary components 1 and 2 as the first cell characteristics) and the astrocytes in FIG. 3B (time point 1) A comparison with the cell characteristics of the cell population 18 identified as 2 detected no changes over time, and a comparison with the cell characteristics of the cell population 20 identified as astrocytes in FIG. Detect that there has been a change over time. Also, the cell characteristics of the cell population 16 identified as the nerve cell 4 in FIG. 3A (time point 0) (main components 1 and 2 as the first cell characteristics) and the nerve cell 4 in FIG. 3B (time point 1) A comparison with the cell characteristics of the obtained cell population 22 detects no change over time.

In addition, the number of astrocyte 2 cells identified as constituting cell population 18 in FIG. 3B (time point 1) is compared to the number of astrocyte 2 cells identified as constituting cell population 14 in FIG. It is detected that the number of astrocyte 2 cells identified as constituting cell population 20 in FIG. Furthermore, the total number of glial cells that make up cell populations 18 and 20 in FIG. 3B (time point 1) is equal to the number of astrocyte 2 cells identified as making up cell population 14 in FIG. 3A (time point 0). also detect Furthermore, the number of nerve cells 4 identified as constituting cell population 22 in FIG. 3B (time point 1) is as follows: Detect no change.

From these detection results, it can be easily detected that changes in the cell characteristics and number of astrocytes 2, but not in nerve cells 4, are observed 1 hour after cell stimulation by addition of the drug. can do. Furthermore, from changes in the cell characteristics and number of astrocytes 2, it can be detected (estimated or confirmed) that the cell population 20 is composed of reactive astrocytes 10 in which the shape of the astrocytes 2 has changed. As a result, it is possible to easily detect (estimate or confirm) that the shape of astrocytes 2 changed from time 0 to time 1 and became reactive astrocytes 10 (solid arrow in FIG. 4).

Similarly, the clustering results obtained in the grouping step (S4) are compared with FIG. Principal components 1 and 2), which are cell characteristics of , detect the presence or absence of changes over time from time point 1 to time point 2.
That is, in FIG. 3C (time point 2), the cell population 18 composed of astrocytes 2 disappeared in FIG. 3B (time point 1), and the reactive astrocytes present in FIG. 3B (time point 1) Detect that only the cell population 20 detected as composed of 10 remains. In addition, in FIG. 3B (time point 1), the existing cell population 22 composed of nerve cells 4 disappears, and in FIG. 3C (time point 2), a cell population 26 identified as composed of nerve cells 4 newly Detect what is happening.
In addition, the cell characteristics of the cell population 20 detected to be composed of reactive astrocytes in FIG. Cellular characteristics of cell populations 24 identified as constituted are compared to detect no change over time. Similarly, the cell characteristics of the cell population 22 detected to be composed of neurons 4 in FIG. 3B (time point 1) (first cell characteristics, principal components 1 and 2), and the The cellular characteristics of the cell population 26 identified as composed of nerve cells 4 are compared to detect changes over time.

Furthermore, the number of reactive astrocytes 10 detected to constitute cell population 24 in FIG. 3B (time point 2) is the total number of cells constituting cell populations 20 and 18 in FIG. 3A (time point 1). and that the number of neurons making up cell population 26 in FIG. 3C (time point 2) is the same as the number of cells in cell population 22 in FIG. 3B (time point 1).
From these detection results, 6 hours after cell stimulation by addition of the drug, that is, 5 hours after time point 1, no change was observed in the cell characteristics and cell number of reactive astrocytes 10, but neuronal cells 4 can easily detect that there is a change. Furthermore, it is detected (estimated) that the cell population 26 is composed of the nerve cells 12 in which the shape of the nerve cell 4 is changed from the change in the cell characteristics and the number of cells of the cell population 22 detected to be composed of the nerve cells 4. or confirm). As a result, it is possible to easily detect (estimate or confirm) that the shape of nerve cell 4 has changed from time point 1 to time point 2 and that it has become nerve cell 12 . (Figure 5 solid line arrow).

Thus, in the present invention, since the cell characteristics of each cell population and the number of cells are associated and displayed, not only the presence or absence of cell death, but also changes in the number of cells and cell characteristics over time, and The relationship between numbers and cell characteristics can be easily confirmed or detected with high accuracy.

<Action mechanism analysis step (S6)>
Next, in the action mechanism analysis step (S6), based on the cell change detection results obtained in the cell change detection step (S5), the action mechanism of changes within or between cell populations of the same cell type is determined. To analyze. Here, the mechanism of action is a specific action for cell stimulation by a drug to exert its pharmacological effect, and a specific action found within a cell population of the same cell type or between cell populations means a biochemical reaction or interaction.

In the present embodiment, the mechanism of action of changes within or between cell populations refers to biological phenomena induced inside or outside cells by cell stimulation (e.g., proliferation, cell death, antigen-antibody reaction, growth factor secretion, etc.), more specifically, specific biochemical reactions or interactions for exhibiting biological phenomena induced inside and outside cells by cell stimulation (e.g., metabolism of biological substances, gene expression , energy metabolism, signal transduction, etc.).
In the present embodiment, each cell population obtained in the cell change detection step (S5) is analyzed from the cell change detection results obtained in the cell change detection step (S5) by analyzing the mechanism of action as described above. Factors (eg, biomaterials, genes, etc.) involved in changes in the number of cells and changes in cell characteristics are extracted.

FIG. 6 is considered to relate to interactions between astrocytes 2 (and reactive astrocytes 10) and nerve cells 4 and 12 in the single cell data acquisition step (S3), along with the aforementioned DNA sequences and fluorescent labels. The component amounts of components A to Z are also obtained as single cell data, and all cells are analyzed along a pseudotime estimated based on the similarity of expression profiles after dimensionality reduction by principal component analysis. It is a plotted figure. By arranging cells in this way, pseudo temporal cell changes and gene expression changes can be found.
Figures 6 (1) to (4) show changes in components A to C and Z of astrocytes 2 (and reactive astrocytes 10), and Figures 6 (5) to (8) show changes in neurons 4 and 12 component AC and Z variations are shown.

The components considered to be related to the interaction between astrocytes 2 (and reactive astrocytes 10) and nerve cells 4 and 12 used in this step are not particularly limited. In the single cell data obtaining step (S3), obtainable single cell data can be used.

In this embodiment, from FIG. 6 (1) to (4), cell population 18 (astrocyte 2) to cell population 20 ( The cell characteristics of reactive astrocytes 10) and changes in the number of cells over time (solid arrows in FIG. 4) are the detection results that the B component, whose component amount changes from time 0 to time 1, is related. Earn.
6 (5) to (8), the cell population 28 related to the time points 1 and 2 detected in the cell change detection step (S5) described above, that is, the cell population 22 (nerve cell 4) to the cell population 26 (nerve cell 12) cell characteristics and cell number over time (solid arrow in Fig. 5) are related to the Z component and B component, whose component amounts change from time 1 to time 2. Get detection results.

In this embodiment, it is used to analyze the process of mutual network formation (change over time) between glial cells and neurons, but is not limited to this, and fungi such as Candida albicans and oral cavity It can also be used to analyze the process of infection with mucosal epithelial cells.

For example, gene expression profile analysis and molecular target analysis can be used as the mechanism of action analysis.
For example, tensor decomposition can be used to analyze gene expression profiles.

An example of molecular target analysis will be described.
First, we identify genes whose expression is altered by drugs (cell stimulation) and whose alteration can be expected to be consistent with disease. Here, although the compound (drug) actually binds to the protein, what is measured is the mRNA expression level. It is unlikely that the binding of a compound (drug) to a protein changes the amount of mRNA in the gene that encodes that protein. ) is assumed to be absent.
The effect on other gene expression profiles of the protein to which the compound actually binds is expected to be similar to knocking out the gene that encodes that protein. Therefore, the target gene is estimated by referring to the gene expression profile when the gene is comprehensively knocked out.

In order to find the optimal indications for cell stimulation from multiple disease or case type candidates, it is possible to obtain the order of effects by simultaneously performing efficacy determination and action mechanism analysis on multiple types of cells. . Optimal indications or uses can be found based on a comparison or ranking of their effects. Here, the determination of the effect of cell stimulation is performed by comparing, for example, data without cell stimulation (drug) treatment and time-dependent data with cell stimulation (drug) treatment, and cells derived from the same cell population are biological data. Alternatively, by comparing cell characteristics and cell numbers, if there is a change, it can be determined that there was an effect.

<Interaction Analysis Step (S7)>
Next, in the interaction analysis step (S7), interactions occurring between cell populations of different cell types are analyzed based on the results of the action mechanism analysis obtained in the cell change detection step (S6). As a result, it is possible to analyze interactions occurring not only between cell populations of the same cell type but also between cell populations of different cell types. can.

The interaction occurring between cell populations is a biological or physical action occurring between different types of cell populations, or a change caused by those actions.
The biological or physical action includes, for example, transfer of humoral factors, cell adhesion, pulsation, or weak current, and the changes caused by the action include, for example, morphological change, differentiation, proliferation, cell death, migration, increase in surface antigens, or release of humoral factors.
The humoral factors are, for example, hormones, growth factors, cytokines, exosomes, metabolites, or inorganic substances such as ions.

When, in the grouping step (S4), values representing the amounts or states of a plurality of biosubstances are obtained at a plurality of time points for each biosubstance from the single-cell data of each cell (information related to a plurality of biosubstances) Series data are prepared in advance, and cells from which single-cell data are acquired based on time changes in time-series data for each biological substance and similarities in biological functions of each biological substance are classified into common first cell characteristics. When grouped into cell populations having, interaction analysis can be performed using the method described in International Publication 2018/150878A1.
In the present invention, in the interaction analysis step (S7), a value representing the state of the cell population is generated from one or more first cell characteristics contained in each of the plurality of cell populations for each of the plurality of time points. Then, the generated values representing the states of multiple cell populations at multiple time points are obtained from a data set consisting of time-series data obtained at multiple time points for each biological material, and the dependency relationships between the state of the cell populations are calculated. A biological or physical effect that occurs between cell populations of different cell types by extracting dependencies between different cell populations in the inter-group (inter-cell population) dependencies that are presumed , or their effects (interactions occurring between cell populations of different cell types) can be estimated.

For example, the state dependency between cell populations can be estimated as follows.
Cells are identified from the sequence data, and genes with similar temporal change patterns of gene expression levels are identified, for example, the state value transitions at 7 adjacent 2 time points increased in light of a certain threshold. It is determined which of the three of unchanged and decreased, and each is classified into 3 ⁷ =2187 groups (cell populations).
Genes with similar biological functions are grouped using Functional Annotation Clustering of the public web tool DAVID (https://david.ncifcrf.gov/) with similar gene ontology. grouping based on similarity of temporal change and similarity of biological function, and temporal dependencies between state values in each group (cell population), e.g., in the light of a Bayesian network model, or , by linking groups (cell populations) according to chronological or biological relationships.

In the present embodiment, the results of the action mechanism analysis obtained in the cell change detection step (S5) described above, that is, the cell characteristics of glial cells and the number of glial cells from time 0 to time 1 (Fig. 4 The solid line arrow) indicates that the B component, whose amount of component changes from time 0 to time 1, is related, and the time-dependent changes in the cell characteristics and number of neurons from time 1 to time 2 ( Solid arrow in FIG. 5) indicates that the Z component and B component, whose component amounts are changing from time 1 to time 2, are related. A cell population composed of glial cells at time point 1 based on changes in components A to C and Z according to and changes in neuronal components A to C and Z shown in FIGS. 6 (5) to (8) The biological or physical action (stimulation) that occurs between 20 and the cell population 22 composed of nerve cells 4, or their action (stimulation), i.e., the interaction that occurs between cell populations of different cell types. From the time point 0 to the time point 2, first, the B component of the glial cells increased due to the cell stimulation by the addition of the drug, and the cell characteristics of the glial cells changed (solid arrows in Figs. ), that is, the astrocytes 2 are changed into reactive astrocytes 10, and the changed reactive astrocytes 10 give biological or physical stimulation to the nerve cells 4 (thick arrows in FIG. 7). Subsequently, in response to the decrease in the component C of the nerve cell 4 stimulated by the reactive astrocytes 10, the Z component of the nerve cell 4 increases, thereby changing the cell characteristics of the nerve cell (Fig. 5). solid line arrows), ie, the change of nerve cell 4 to nerve cell 12, an intercellular network between glial cells and nerve cells can be presumed.

According to the method of Embodiment 1, single-cell analysis is performed on all cells that constitute a target cell group in which a plurality of different cell types are present, and the cell type of each cell is identified. Compared to the conventional method of artificially identifying the cell type of each cell from a target cell group in which cell types exist, the identification of each cell type and the mechanism of action of drugs or cell stimulation on various cells can be analyzed with higher accuracy. be able to.
In addition, since the identification of each cell type and the mechanism of action of drugs or cell stimulation on each cell type can be analyzed with high accuracy, not only interactions between the same cell type but also interactions between different cell types can be analyzed. It is highly accurate and can be easily analyzed. In addition, as a result, it is possible to analyze wider interactions between cells and between cell types (reciprocal networks) with high accuracy and ease.
In addition, the identification and selection of cell types does not require time and effort such as image analysis, and even when performing drug efficacy screening for immune cells, observation of one sample does not continue until the analysis is completed. In particular, the analysis can be performed without limiting the analysis time.
In addition, since single-cell analysis is used, the number of time points for observing (collecting) cells can be reduced compared to the case of artificially confirming changes in cells and selecting target cells.

In addition, since the method of Embodiment 1 can visualize the state of cells at each time point by the grouping step (S4), changes in each cell (cell population) can be easily detected in the cell change detection step (S5). can be confirmed. As a result, in the action mechanism analysis step (S6), it is possible to easily extract cell types, genes, etc. that require further analysis.

Modification of Embodiment 1 In the above-described Embodiment 1, in the cell collection step (S1), cells collected immediately after cell stimulation and cells cultured for a predetermined period of time after cell stimulation are collected, and single cells of these cells are collected. The analysis is performed based on the comparison of cell data, but is not limited to this. In the cell collection step (S1), a target cell group to which no cell stimulation is applied and a target cell group to which cell stimulation is applied are prepared respectively. After culturing for a predetermined period of time, cells may be harvested and screening analysis may be performed based on comparison of single-cell data of those cells.

Specifically, in the cell collection step (S1), among the plurality of containers 1 prepared, the target cell group seeded in some containers 1 is cultured without giving cell stimulation, and the remaining The target cell group seeded in the container 1 of Embodiment 1 is subjected to cell stimulation, and all cells in one container are collected at two or more time points. is similar to

In this way, by further preparing and analyzing a control sample, it is possible to confirm whether the effect is due to the drug (cell stimulation) or other factors such as culture conditions.
More specifically, in the mechanism of action analysis step (S6), the mechanism of action of the drug (cell stimulation) within individual cell populations or between cell populations can be analyzed, and furthermore, the molecular target of cell stimulation can be analyzed. It will also be possible to analyze and identify indications for treatment.
Here, examples of indications for which treatment is assumed include, for example, diseases or symptoms that are expected to be improved by cell stimulation, or diseases or symptoms related to molecular targets.

EXAMPLES The present invention will be described in more detail with reference to Examples below, but the present invention is not limited to these Examples.

[Example 1]
Human iPS cells and mouse embryo-derived fibroblasts were mixed at 1:1 (cell number ratio), and ROCK (Rho-associated coiled-coil forming kinase / Rho binding kinase) inhibitor Y-27632 (10 μL; Fujifilm Wako Jun Yakusha) were added and seeded in 6-well plates at 1000 cells/well/200 μL, 3000 cells/well/200 μL, and 9000 cells/well/200 μL.
Cells were harvested at time of seeding (0 hours), 6 hours, 12 hours, 24 hours, 48 hours, 72 hours, 96 hours and 120 hours after seeding.
The cell suspension was diluted to 1000 cells/μL, and single cells were captured using the C1 system (manufactured by Fluidigm). Cell lysis, mRNA reverse transcription and cDNA pre-amplification were then performed using ^{the SMARTer®} Ultra® Low RNA kit.
The resulting cDNA was collected and concentrations higher than 0.05 ng/μL were selected for library preparation. Library preparation was performed using Nextera ^(R) XT DNA sample preparation kit (manufactured by Illumina).

The prepared library was sequenced with a next-generation sequencer (HiSeq2500 system, Illumina) using 2×100 bp terminal reads. Gene expression levels for each cell were calculated from the obtained data, time point samples were collected for each drug, and clustering analysis was performed using principal component analysis.
The cells are identified as human iPS cells or mouse fetal fibroblasts from the sequence data, and genes with similar temporal change patterns of gene expression levels, specifically, 7 It is determined which of the three states, that is, the transition of the state value at two adjacent points in time is increased, remained unchanged, or decreased in light of a certain threshold, and each group is divided into 2187 (=3 to the 7th power) groups. classified into Genes with similar biological functions were grouped using Functional Annotation Clustering of the public web tool DAVID (https://david.ncifcrf.gov/) with similar gene ontology. Grouping based on similarity of temporal changes and similarity of biological functions resulted in 7305 groups. A Bayesian network model was used to estimate the temporal dependencies between the state values of the 7305 groups. was able to capture the chronological mechanism.

[Example 2]
Primary human neurons, immortalized human microglia transfected with SV40T antigen, and immortalized astrocytes transfected with the hTERT gene were mixed at a ratio of 2:1:1 (cell number ratio), seeded in a 6-well plate, and cultured for 24 hours. did.
After confirming engraftment, saline, TNF-α (tumor necrosis factor α), IL-1β (interleukin 1β), IL-1ra (interleukin 1 receptor antagonist), IL-12 (interleukin 12), IFNγ (interferon gamma), IRF1 (interferon regulatory factor 1), IRF2 (interferon regulatory factor 2), IRF3 (interferon regulatory factor 3), IRF4 (interferon regulatory factor 4), IRF5 (interferon regulatory factor 5), IRF6 (interferon regulatory factor) Factor 6), IRF7 (interferon regulatory factor 7), IRF8 (interferon regulatory factor 8), IRF9 (interferon regulatory factor 9), and LPS (lipopolysaccharide) were added to each well.
Cells were harvested by trypsinization at the time of addition (0 hours), 6 hours, 12 hours, 24 hours, and 48 hours after addition.
The cell suspension was diluted to 1000 cells/μL, and single cells were captured using the C1 system (manufactured by Fluidigm).

Target primers 500-600 and 750-870 of the TP53 gene were added to a cDNA preparation kit (SMARTer ^(R) Ultra ^(R) Low RNA kit, manufactured by Clontech) to lyse cells, reverse-transcribe mRNA, and pre-amplify cDNA. did
The resulting cDNA was collected and concentrations higher than 0.05 ng/μL were selected for library preparation. Library preparation was performed using Nextera ^(R) XT DNA sample preparation kit (manufactured by Illumina).
The prepared library was sequenced with a next-generation sequencer (HiSeq2500 system, Illumina) using 2×100 bp terminal reads.
Gene expression levels for each cell were calculated from the obtained data, time point samples were collected for each drug, and clustering analysis was performed using principal component analysis.

Cells in which the SV40 T antigen gene was detected were classified as immortalized microglial clusters, cells in which the hTERT gene was detected as immortalized astrocyte clusters, and cells in which neither the SV40 T antigen gene nor the hTERT gene were detected as human primary neurons. Each was grouped into cell clusters.
For each cell type, we found changes in the number of cells and gene expression levels over time, and obtained temporal change patterns of genes. By grouping and patterning the temporal variation patterns of each gene in each cell and calculating the temporal dependence, the dependence between cells was obtained.

Although various embodiments and examples of the cell information processing method of the present invention have been described in detail above, the present invention is not limited to these embodiments and examples, and does not depart from the gist of the present invention. Of course, various improvements or changes may be made within the scope.

Claims (24)

At least two or more predetermined containers seeded with a target cell group containing a plurality of different types of cells are prepared, and after predetermined cell stimulation for the target cell group, all cells in one container at two or more time points a cell recovery step of recovering
A single cell conversion step of converting all cells collected at each time point into single cells;
A single-cell data acquisition step of acquiring single-cell data from each single-celled cell at each time point;
Based on the single cell data at each time point, all cells collected at each time point are grouped into a plurality of cell populations having a common first cell characteristic and plotted on a two-dimensional plane or three-dimensional space; And, based on the second cell characteristics, a grouping step of performing a process of identifying the cell type of each grouped cell population and obtaining a clustering result at each time point;
A cell change detection step of detecting changes over time in the cell population of the same cell type by comparing the clustering results at each time point;
A mechanism-of-action analysis step of analyzing the mechanism of action of changes over time in the cell population of the same cell type based on the detection results;
By comparing the clustering results at the two or more time points together with the results of the mechanism of action analysis, the dependencies between clusterings are analyzed, and interactions that occur between the cell populations of different cell types are analyzed. and an interaction analysis step. 2. The cell information processing method according to claim 1, wherein the interactions occurring between the cell populations in the interaction analysis step include interactions occurring between the cell populations of the same cell type. In the cell change detection step, by comparing the clustering results at each time point, changes over time in the number of cells constituting the cell population of the same cell type and changes in the first cell characteristics over time The cell information processing method according to claim 1 or 2, wherein the is detected. The cell change detection step further detects changes over time in the number of cells constituting the cell population of the same cell type and changes in the first cell characteristics over time in the clustering results at each time point. The cell information processing method according to claim 3. The cell recovery step further includes preparing a target cell group containing the plurality of types of cells to be cultured without applying the predetermined cell stimulation, and at one or more time points, all cells in one of the predetermined containers collect the
2. The cell change detection step evaluates changes over time in the cell population of the same cell tumor due to the presence or absence of the predetermined cell stimulation of the cell population by comparing the clustering results at each of the time points. 5. The cell information processing method according to any one of -4. The cell information processing method according to any one of claims 1 to 5, wherein the cell stimulation is at least one selected from the group consisting of chemical stimulation and physical stimulation. 7. The cell information processing method according to claim 6, wherein said chemical stimulation is by addition of a drug that induces a biological response to cells. The action mechanism according to any one of claims 1 to 7, wherein the cell stimulation is a specific biochemical reaction or interaction for exerting a biological phenomenon induced inside or outside the cell. A method for processing cellular information as described. The single cell data is information on biological substances that indicate the functions, properties, and states of single cells, and includes DNA sequence information (genome) of genes, epigenetic information that controls gene expression (DNA methylation, histone methylation, acetylation, acetylation, phosphorylation), primary gene transcripts (mRNA, untranslated RNA, microRNA, etc.) information (transcriptome), protein translation amount and modification information such as phosphorylation, oxidation, saccharification, etc., amino acid sequence selected from the group consisting of information (proteome), metabolite information (metabolome), intracellular hydrogen ion concentration index (pH), intracellular ATP concentration, ion concentration (calcium, magnesium, potassium, sodium, etc.), intracellular temperature The cell information processing method according to any one of claims 1 to 8, which is at least one. 10. The cell information processing method according to claim 9, wherein the single cell data are gene expression levels and gene DNA sequences. 11. The method according to any one of claims 1 to 10, wherein in the grouping step, the first cell feature is obtained by reducing n-dimensional cell features included in the single cell data to two or three dimensions. A method for processing cellular information as described. The dimensionality reduction method is selected from the group consisting of Principal Component Analysis (PCA), Principal Component Analysis with Kernel (Kernel-PCA), Multidimensional Scaling (MDS), t-SNE, and Convolutional Neural Network (CNN). The cell information processing method according to claim 11, wherein at least one is selected. 13. Any one of claims 10 to 12, wherein in the grouping step, the first cell feature is obtained by performing principal component analysis on the gene expression level and reducing the dimensions to two or three dimensions. The method for information processing of cells according to the item. 14. Any one of claims 1 to 13, wherein in the grouping step, the second cell feature is at least one piece of cell information that enables identification of each cell type from cell function or cell state. The method for processing cell information according to . The cellular information includes gene DNA sequence information (genome), epigenetic information that controls gene expression (DNA methylation, histone methylation, acetylation, phosphorylation), gene primary transcripts (mRNA, non- Translated RNA, microRNA, etc.) information (transcriptome), protein translation amount and modification information such as phosphorylation, oxidation, and saccharification, amino acid sequence information (proteome), metabolite information (metabolome), intracellular hydrogen ion concentration index (pH), intracellular ATP concentration, ion concentration (calcium, magnesium, potassium, sodium, etc.), at least one selected from the group consisting of intracellular temperature. The cell information processing method according to claim 14 or 15, wherein the cell function is at least one selected from cell proliferation, repair, metabolism, and information exchange between cells. The cell information processing method according to claim 14 or 15, wherein the state of the cell is at least one selected from gene expression status, protein expression status, and enzyme activity. In the interaction analysis step, the interaction that occurs between the cell populations is a biological or physical action that occurs between the cell populations of the different cell types, or the number of cells due to the action, the first The cell information processing method according to any one of claims 1 to 17, which is a change in cell characteristics. Time-series data obtained from the single-cell data, values representing amounts or states of a plurality of biomaterials obtained at a plurality of points in time for each of the biomaterials are created in advance, and the time-series data for each of the biomaterials is changed over time. grouping the cells from which the single-cell data was obtained into cell populations having the common first cell characteristic based on similarities in biological function of each biological material;
In the interaction analysis step, for each of the plurality of time points, a value representing the state of the cell population is generated from one or more first cell characteristics contained in each of the plurality of cell populations, and generated , estimating the dependence of the state between cell populations from a data set consisting of time-series data obtained at a plurality of time points for each biological material, with values representing the state of a plurality of cell populations at multiple time points. 19. The cell information processing method according to any one of 1 to 18. The similarity of biological function includes having a common gene ontology, belonging to a common canonical pathway, having common upstream factors, involving a common phenotype, and involving a common disease. The cell information processing method according to claim 19, wherein the evaluation is based on at least one selected from the group consisting of: In the cell collection step and the single-cell conversion step, the methods for collecting all cells and converting them into single cells are manual, flow cytometry, magnetic separation, laser capture microdissection, microfluidics, microdroplets, nanowells, A method using at least one selected from the group consisting of micropipette fine needle aspiration, laser tweezers, labeled arrays, surface plasmon response, and nanobiodevices, cell information according to any one of claims 1 to 20 Processing method. 22. The method according to claim 21, wherein in the method of collecting the whole cells and converting them into single cells, at least one selected from the group consisting of fluorescent labeling, radioisotope labeling, antibody labeling, and magnetic labeling is used as the cell label. Cell information processing method. Any of claims 1 to 22, wherein the target cell group containing multiple types of cells is obtained from at least one selected from the group consisting of biological tissue samples, blood samples, culture samples, and environmental samples. 1. The cell information processing method according to 1. 24. The cell information processing method according to claim 23, wherein said plurality of types of cells are at least one selected from the group consisting of animal cells, plant cells, fungal cells and bacterial cells.

Images