Guest Author: Jennifer MacArthur
Single cell RNA sequencing can help connect datasets for multiple dimensions to reveal a rich picture of cell phenotypes.
Our cells are bustling with activity: transcribing thousands of genes, translating and modifying millions of proteins, responding to signals, transporting molecules, controlling growth, and working in cooperation with the trillions of other cells that make up our human bodies. RNA plays a pivotal role in the cell, connecting genomes and epigenomes to proteomes and metabolomes. Multiomic researchers can use RNA sequencing to link these complex dynamics at the single cell level.
Advances in single cell technologies allow us to capture molecular snapshots of cell activity in more and more detail. 10x Genomics offers solutions to assay single cells in multiple dimensions, including gene expression, cell surface proteins, copy number variation, and chromatin accessibility. Multiomics combines genomic, epigenomic, transcriptomic, and/or proteomic metrics into a more complete and vivid picture of complex biology at fundamental resolution and massive scale.
Dr. Rahul Satija’s lab at New York Genome Center and NYU is developing statistical analysis tools to manage, combine, and decipher massive single cell datasets from different “-omics” technologies. In their recent paper, they explain why single cell RNA-sequencing (scRNA-seq) is key for multiomic data integration:
“Each technology has unique strengths and weaknesses and measures only particular aspects of cellular identity, motivating the need to leverage information in one dataset to improve the interpretation of another ... Our results suggest that scRNA-seq can serve as a general mediator for single-cell data integration. Not only is its application commercialized and routinely available, but also, transcriptome-wide gene expression data encodes multiple aspects of cellular identity and ‘metadata’ ... Its intermediate position in the central dogma allows for proximity to multiple molecular processes, including transcriptional, posttranscriptional, and translational regulation. We therefore suggest that scRNA-seq can serve as a ‘universal adapter plug’ for single-cell analysis, facilitating integration across multiple technologies and modalities, and enable a deeper understanding of cellular state, interactions, and behavior” (1, emphasis added).
scRNA-seq is an ideal foundation for multiomic analysis. Gene expression can serve as a proxy for protein expression, offering parallel information about cell types and states, without the need for antibodies to probe specific markers. In this way, transcriptomics provides an unbiased readout of cellular phenotype. Additionally, transcript levels often reflect copy number variation in the genome, or correlate with open chromatin regions (2–4). Beyond counting transcripts for gene expression profiling, certain scRNA-seq techniques can provide full-length gene sequence data for immune repertoire profiling or other applications.
Computational approaches can help extract the “metadata” from scRNA-seq information. Dr. Satija’s lab demonstrated how they transferred gene expression profiles onto ATAC-seq chromatin accessibility datasets to reveal regulatory programs and finer distinctions among cell types (1). InferCNV, an algorithm developed at the Broad Institute of MIT and Harvard, can use single cell gene expression data to infer large-scale chromosomal copy number variations in tumor cells.
Not all multiomics approaches are dependent on data integration after separate experiments. Newer single cell sequencing techniques build multiomics into the assay. Our Feature Barcode technology uses oligo-barcoded antibodies for highly multiplexed cell surface protein analysis, in parallel with gene expression profiling in the same cells. Feature Barcode technology can also use MHC-multimers to measure antigen specificity, or tag specific CRISPR perturbations for functional genomics screens. For example, researchers can link gene expression, cell surface protein, immune cell clonotype, and antigen specificity for single cells in the same experiment—or pick and choose which features to measure.
Ultimately, researchers hope to resolve complex biology by assembling a detailed, multidimensional view of each cell, in context with millions of other cells. Technologies to capture genomic, epigenomic, transcriptomic, and proteomic information at fundamental resolution, along with computational tools to integrate single cell data, will help make that possible. RNA can serve as the hub that connects genomes to phenotypes, and pulls multiomic data together into one full and vivid picture.
T Stuart et al. Comprehensive Integration of Single-Cell Data. Cell. 177, 1888-1902 (2019).
A Schlattl et al. Relating CNVs to transcriptome data at fine resolution: Assessment of the effect of variant size, type, and overlap with functional regions. Genome Res. 21, 2004–2013 (2011).
DA Cusanovich et al. A Single-Cell Atlas of In Vivo Mammalian Chromatin Accessibility. Cell. 174, 1309-1324 (2018).
BB Lake et al. Integrative single-cell analysis of transcriptional and epigenetic states in the human adult brain. Nat Biotechnol. 36, 70-80 (2018).