- Modern biological systems are composed of many different types of cells, and even cells belonging to the same type can exist in different molecular states. A tissue that appears uniform under a microscope may contain cells with very different patterns of gene expression, protein abundance, signaling activity, metabolism and regulatory states. Traditional omics technologies often measure many cells together, producing an average molecular profile. Although these measurements are extremely useful, averaging can hide important biological differences between individual cells. Single-cell multi-omics addresses this limitation by combining multiple molecular measurements at the level of individual cells or closely resolved cellular units.
- The fundamental concept is an extension of single-cell analysis and multi-omics integration. Instead of studying only the genome, transcriptome, proteome or epigenome separately, researchers attempt to connect multiple molecular layers within individual cells. Depending on the technology and experimental design, a single-cell experiment may measure RNA expression together with chromatin accessibility, DNA methylation, proteins, immune receptors or other molecular characteristics. These measurements can reveal not only which genes are expressed but also how regulatory states, transcriptional programs and cellular identities are connected.
- This approach is particularly important because cells within the same tissue are not molecularly identical. Development, differentiation, disease, immune activation, cancer progression and environmental responses can produce substantial cellular heterogeneity. A bulk transcriptomic experiment might indicate that a gene is moderately expressed in a tissue, but that average could represent strong expression in one cell population and little or no expression in another. Single-cell analysis separates these populations and makes their molecular differences visible.
- The distinction between bulk multi-omics and single-cell multi-omics is therefore important. Bulk multi-omics integrates different molecular layers but usually averages signals across large numbers of cells. Single-cell multi-omics attempts to preserve cellular resolution. This allows researchers to ask questions such as which cells express a particular gene, which cells have accessible regulatory regions, which cells contain a particular protein, which cell populations activate a signaling pathway, and how molecular states change during differentiation or disease.
- Single-cell transcriptomics, particularly single-cell RNA sequencing, is one of the most widely used approaches in this field. Instead of measuring the average RNA population of a tissue, it measures RNA molecules associated with individual cells. Computational analysis can then group cells according to their transcriptional profiles. These groups may correspond to known cell types, developmental states or previously unrecognized cellular populations.
- However, RNA expression alone does not explain the complete regulatory state of a cell. A gene can be expressed because its regulatory region is accessible, but RNA production can also depend on transcription factors, chromatin state and other regulatory mechanisms. This is why combining transcriptomics with single-cell epigenomics can provide additional information.
- Single-cell ATAC-seq, for example, measures regions of chromatin that are accessible to regulatory proteins. When chromatin accessibility and RNA expression are analyzed together, researchers can investigate relationships between regulatory regions and gene activity. This creates a connection between genome regulation and transcription at the level of individual cells.
- Single-cell multi-omics can therefore connect several layers of gene regulation. A regulatory DNA region may be accessible in a particular cell population, transcription factors may bind or influence that region, the associated gene may be transcribed, and the resulting protein may contribute to a specific cellular function. Connecting these layers helps researchers move from observing gene expression to investigating the regulatory mechanisms responsible for that expression.
- The proteomic layer provides another important dimension. RNA abundance does not necessarily predict protein abundance because translation efficiency, protein degradation, localization and post-translational regulation influence protein levels. Single-cell protein measurements can therefore complement single-cell transcriptomics by showing which proteins are actually present.
- Methods such as antibody-based single-cell protein profiling can measure selected surface or intracellular proteins together with RNA. This is particularly useful for identifying immune cell populations because cell-surface proteins can provide strong markers of cellular identity. Combining RNA and protein measurements can improve cell-type classification and reveal cases where transcript and protein abundance differ.
- Another important layer is single-cell phosphoproteomics and related measurements of protein modifications. Signaling pathways often operate through phosphorylation and other post-translational modifications rather than changes in protein abundance. A cell may contain the same amount of a kinase as another cell while having a very different signaling state because the kinase or its substrates are differently phosphorylated.
- This highlights a broader principle of single-cell multi-omics: molecular identity and molecular activity are not necessarily the same thing. A protein may be present but inactive, a gene may be transcribed but not efficiently translated, and a signaling pathway may be activated without major changes in RNA expression. Integrating molecular layers helps distinguish these states.
- Single-cell multi-omics is also closely connected to cellular identity. Cells can be classified using combinations of molecular features rather than relying on a single marker. Transcriptomic profiles can identify gene-expression programs, chromatin accessibility can reveal regulatory states, and protein measurements can identify characteristic cellular markers. Together, these measurements can provide a more detailed representation of cell identity.
- This is particularly useful in tissues containing closely related cell types. For example, immune cells can include many subpopulations with overlapping markers but distinct functional states. Single-cell multi-omics can identify these populations and investigate how their regulatory, transcriptional and protein states differ.
- An important application is the study of cellular differentiation. During development, cells transition from one state to another through coordinated changes in chromatin accessibility, transcription, protein expression and signaling. Single-cell datasets can capture different stages of this process within a population of cells.
- Computational methods can then attempt to arrange cells along developmental trajectories. Trajectory analysis and pseudotime analysis can estimate relationships between cellular states based on molecular similarity. These approaches can provide hypotheses about how one state may transition into another, although computational trajectories should not automatically be interpreted as direct evidence of a physical lineage unless supported by appropriate experiments.
- Single-cell multi-omics is particularly powerful for studying cellular heterogeneity in cancer. Tumors are not composed solely of genetically similar cancer cells. They can contain malignant subpopulations together with immune cells, endothelial cells, fibroblasts and other stromal populations. Even cancer cells carrying similar driver alterations can differ substantially in transcriptional state, signaling activity and metabolism.
- Single-cell genomics can identify genetic alterations within individual cells, while transcriptomics can characterize their gene-expression states. Epigenomic measurements can reveal differences in regulatory programs, and protein measurements can provide information about cellular signaling and phenotype. Integrating these layers can therefore reveal distinct cellular states within the same tumor.
- This cellular resolution is important for understanding tumor evolution and drug resistance. A treatment may eliminate one population while another population survives because of a different molecular state. Resistance may involve genetic changes, altered transcription, activation of alternative signaling pathways, changes in chromatin regulation or metabolic adaptation. Single-cell multi-omics provides a framework for investigating these mechanisms within individual cellular populations rather than averaging them across the tumor.
- Single-cell analysis is also transforming the study of the immune system. Immune responses involve many cell types and cellular states, including T cells, B cells, natural killer cells, macrophages, dendritic cells and other populations. These cells can undergo rapid changes in gene expression, signaling and metabolism following infection, vaccination, inflammation or therapeutic intervention.
- Single-cell multi-omics can connect immune-cell identity with receptor repertoires, gene expression, chromatin state and protein markers. In some approaches, the sequences of immune receptors can be linked to the molecular state of the same cell. This allows researchers to study not only which immune cells are present but also their molecular state and receptor characteristics.
- Another important application is single-cell genomics for studying genetic variation. In tissues composed of many cells, mutations may occur only in a subset of cells. Bulk sequencing can detect a variant but may provide limited information about which cells contain it. Single-cell sequencing can help determine the distribution of genetic alterations among cellular populations.
- When genetic information is combined with transcriptomic and epigenomic data, researchers can investigate how genomic differences are associated with cellular phenotypes. This is particularly useful in cancer, developmental biology, mosaicism and studies of clonal cell populations.
- Single-cell multi-omics can also help connect genetic variants with regulatory consequences. A variant may alter a regulatory element that is active only in a particular cell type. Bulk measurements may dilute this effect because the relevant regulatory region is inactive in most cells. Single-cell measurements can identify the specific cell population in which the variant has a molecular consequence.
- This principle is important for interpreting non-coding variants. A DNA variant located outside a protein-coding region may influence gene regulation rather than protein sequence. By integrating chromatin accessibility, transcription factor activity and gene expression at single-cell resolution, researchers can investigate whether the variant lies within a regulatory element active in a particular cell population.
- The connection with protein sequence analysis remains important for coding variants. When a variant changes an amino acid, sequence analysis can determine whether the affected residue occurs within a conserved protein domain, protein motif or other functionally important region. Protein structure prediction, structural comparison and protein-ligand interaction analysis can then provide information about possible molecular consequences.
- Single-cell multi-omics adds another level by asking where and in which cells the altered protein is expressed. Thus, a genetic variant can potentially be followed from DNA sequence through protein structure and molecular interaction to cell-specific pathway activity and phenotype.
- Single-cell multi-omics also has an important spatial dimension. Cells do not function independently of their surroundings. Neighboring cells exchange signaling molecules, metabolites and other factors, while extracellular matrix and tissue architecture influence cellular behavior. Spatial omics preserves information about where molecular signals occur within tissues.
- Traditional single-cell sequencing often requires dissociation of tissue into individual cells, which can remove information about physical location. Spatial technologies address this limitation by measuring RNA, proteins or other molecular features while retaining spatial coordinates. Combining single-cell and spatial approaches can therefore connect molecular identity with tissue organization.
- This is particularly important in cancer, where the interaction between tumor cells and the surrounding tumor microenvironment can influence disease progression and therapeutic response. Spatial analysis can reveal whether specific immune cells are located near tumor cells, whether signaling programs are concentrated in particular regions, and whether metabolic or transcriptional states vary according to tissue location.
- Single-cell multi-omics can also be used to study cell-cell communication. If one cell population produces a signaling molecule and another expresses the corresponding receptor, researchers can generate hypotheses about communication between the two populations. Combining ligand-receptor analysis with transcriptomic, proteomic and spatial information can provide a more detailed picture of intercellular signaling.
- However, computationally inferred cell-cell communication is generally a hypothesis rather than direct proof of physical signaling. Experimental validation is required to establish whether a predicted interaction actually occurs and whether it produces the proposed biological effect.
- The enormous amount of information generated by single-cell multi-omics creates substantial computational challenges. Each cell can be represented by thousands of molecular features, and modern experiments may contain thousands, hundreds of thousands or even millions of cells. The resulting datasets are therefore extremely high-dimensional.
- Different molecular layers also have different characteristics. RNA sequencing produces count-based measurements, chromatin accessibility produces sparse accessibility profiles, protein measurements may have different dynamic ranges, and genomic measurements can contain allele-specific information. Integrating these heterogeneous datasets requires specialized statistical and computational methods.
- Dimensionality reduction methods such as principal component analysis, UMAP and related approaches are frequently used to visualize high-dimensional single-cell datasets. Clustering can identify groups of cells with similar molecular profiles. Integration algorithms can align datasets produced using different technologies or experiments.
- However, computational integration must be performed carefully. Technical differences between experiments can create apparent cellular populations that reflect batch effects rather than biology. Quality control, normalization and batch correction are therefore essential components of single-cell analysis.
- Another challenge is dropout or sparse detection. Because individual cells contain relatively small amounts of RNA and other molecular material, many molecules may not be detected even when they are biologically present. A zero measurement therefore does not always mean that the corresponding molecule is completely absent.
- Cell isolation can also alter cellular states. Tissue dissociation may induce stress responses, change gene expression or preferentially damage certain cell types. Consequently, single-cell measurements must be interpreted with awareness of the experimental procedure used to generate the dataset.
- The integration of single-cell datasets from different experiments introduces another challenge. Different technologies may measure different populations, use different molecular capture efficiencies or provide different levels of sensitivity. Data integration methods attempt to identify shared biological structure while minimizing technical variation, but excessive correction can potentially remove genuine biological differences.
- Single-cell multi-omics is also increasingly connected to artificial intelligence and machine learning. Machine-learning models can classify cell types, identify molecular states, predict regulatory relationships and integrate multiple data modalities. Deep learning and graph-based approaches can represent relationships among genes, proteins, regulatory elements and cells.
- Graph-based models are particularly relevant because biological systems can naturally be represented as networks. Genes can connect to regulatory regions, proteins can connect to interacting proteins, metabolites can connect to enzymes, and cells can connect through communication pathways. A multi-layer network can therefore represent relationships both within cells and between cells.
- Despite the power of these computational methods, biological interpretation remains essential. A computationally identified cell cluster is not automatically a new biological cell type, and a predicted regulatory interaction is not necessarily functional. Experimental validation remains necessary to distinguish biological mechanisms from statistical patterns.
- Single-cell multi-omics is also increasingly important in drug discovery. Conventional drug screening often measures the average response of a population of cells. Single-cell analysis can reveal whether a drug affects all cells similarly or whether distinct cellular populations respond differently.
- A drug may strongly inhibit a target pathway in one cell population while having little effect in another because of differences in target expression, signaling state, genetic background or cellular environment. Single-cell multi-omics can therefore help identify responsive and non-responsive cell populations.
- This information can be combined with drug-target interactions, polypharmacology, network pharmacology, molecular docking, molecular dynamics and structure-based drug design. Structural approaches describe how a drug interacts with its molecular targets, while single-cell multi-omics can reveal how those molecular interactions affect different cell populations.
- Single-cell measurements can also contribute to understanding combination therapies. Two drugs may affect different cellular populations or different molecular pathways. A combination may therefore produce a broader response than either treatment alone, although the biological effects and mechanisms must be experimentally established.
- Another important direction is the integration of single-cell multi-omics with time-series experiments. Cells can change state dynamically in response to developmental signals, infection, environmental stress or drug treatment. Measuring cells at multiple time points can reveal how molecular states evolve.
- A time-resolved single-cell experiment can therefore move beyond a static classification of cell types and investigate cellular transitions. Combining temporal measurements with perturbation experiments can provide stronger evidence about causal relationships between molecular events.
- Perturbation biology is particularly important because observational single-cell data primarily reveal associations. Researchers can experimentally alter genes, signaling pathways or environmental conditions and then use single-cell multi-omics to measure the resulting molecular changes. This creates a powerful cycle in which computational analysis generates hypotheses and experiments test them.
- For example, inhibition of a signaling protein may alter phosphorylation, transcription, chromatin accessibility and metabolism in different cell populations. Single-cell multi-omics can determine whether all cells respond similarly or whether specific populations activate compensatory pathways.
- This approach contributes to a more predictive form of systems biology. Instead of simply cataloguing cell populations, researchers can investigate how cellular systems respond to controlled perturbations and construct models describing those responses.
- The relationship between single-cell multi-omics and systems biology is therefore particularly important. Systems biology seeks to understand biological systems as interconnected networks, while single-cell multi-omics provides detailed molecular measurements from the individual components of those systems.
- A tissue can be viewed as a collection of interacting cells, each containing interconnected molecular networks. Within each cell, genes, proteins and metabolites form regulatory and biochemical networks. Between cells, signaling molecules and physical interactions create additional networks. Single-cell multi-omics can therefore connect intracellular molecular networks with intercellular communication.
- The broader framework can be represented as: DNA → chromatin → RNA → protein → protein modification → metabolic state → cellular pathway → cellular state → cell-cell interaction → tissue phenotype
- Single-cell multi-omics attempts to measure several points along this chain while preserving information about individual cells.
- This also extends the molecular-to-system hierarchy developed throughout this article series: Sequence → family → domain → motif → structure → interaction → pathway → network → multi-omics → single-cell state → tissue organization → phenotype
- At the sequence level, researchers can study individual amino acids and conserved regions. At the structural level, they can investigate three-dimensional molecular organization and interactions. At the systems level, they can examine pathways and networks. Single-cell multi-omics adds cellular resolution to this framework, revealing how molecular networks differ between individual cells and how those cells interact within tissues.
- The future of this field is likely to involve increasingly integrated measurements. Researchers are working toward combining genomics, epigenomics, transcriptomics, proteomics, metabolomics, spatial information and functional measurements within the same biological systems. Such approaches may provide increasingly detailed maps of cellular states and transitions.
- Nevertheless, more molecular measurements do not automatically produce better biological understanding. Experimental design, appropriate controls, statistical rigor, biological validation and careful interpretation remain essential. The goal of single-cell multi-omics is not simply to measure more molecules but to understand how molecular mechanisms generate cellular diversity and biological function.
- Single-cell multi-omics therefore represents an important bridge between multi-omics integration and cellular systems biology. It transforms an averaged molecular description into a cell-resolved view of biological systems, allowing researchers to investigate cellular heterogeneity, regulatory states, developmental trajectories, disease mechanisms, drug responses and cell-cell interactions.
