Statistical single cell multi-omics integration. Issue 7 (February 2018)

• Record Type: Journal Article
• Authors: Colomé-Tatché, M.; Theis, F.J.
• Is Part Of: Current opinion in systems biology. Issue 7(2018)
• Page Start: 54
• Page End: 59
• Publication Date: 2018-02
• Subjects: Systems biology -- Periodicals; 570
• DOI: 10.1016/j.coisb.2018.01.003
• Languages: English
• ISSNs: 2452-3100

Abstract

Abstract: Single cell high throughput genomic measurements are revolutionizing the fields of biology and medicine, providing a means to tackle biological problems that have thus far been inaccessible, such as the systematic discovery of new cell types, the identification of cellular heterogeneity in health and disease, or the cell-fate decisions taking place during differentiation and reprogramming. Recently implemented multi–omics measurements of genomes, transcriptomes, epigenomes, proteomes and chromatin organization are opening up new avenues to begin to disentangle the causal relationship between -omics layers and how these co-determine higher-order cellular phenotypes. This technological revolution is not restricted to basic science but promises major breakthroughs in medical diagnostics and treatments. In this paper we review existing computational methods for the analysis and integration of different -omics layers and discuss what new approaches are needed to leverage the full potential of single cell multi-omics data. Highlights: Single cell multi-omics measures are available to link genome, methylome, chromatin architecture, transcriptome and proteome. Current analysis methods compute correlations between -omics layers, or calculate and integrate independent single-cell maps. Other statistical approaches should include joint multi-omics kernels, view-specific classifiers and network analysis.