Deep generative modeling for single-cell transcriptomics

• Published in: Nature methods Vol. 15; no. 12; pp. 1053 - 1058
• Main Authors: Lopez, Romain, Regier, Jeffrey, Cole, Michael B., Jordan, Michael I., Yosef, Nir
• Format: Journal Article
• Language: English
• Published: New York Nature Publishing Group US 01.12.2018; Nature Publishing Group
• Subjects: 631/114; 631/114/2397; Algorithms; Animals; Artificial neural networks; Biodiversity; Bioinformatics; Biological Microscopy; Biological Techniques; Biomedical and Life Sciences; Biomedical Engineering/Biotechnology; Brain - cytology; Brain - metabolism; Cluster Analysis; Clustering; Computational Biology - methods; Downstream effects; Gene expression; Genetic Variation; Hematopoietic Stem Cells - cytology; Hematopoietic Stem Cells - metabolism; High-Throughput Nucleotide Sequencing - methods; Humans; Leukocytes, Mononuclear - cytology; Leukocytes, Mononuclear - metabolism; Life Sciences; Mice; Models, Biological; Neural networks; Probabilistic inference; Proteomics; Sensitivity analysis; Sequence Analysis, RNA - methods; Single-Cell Analysis - methods; Transcriptome; Uncertainty analysis
• ISSN: 1548-7091; 1548-7105; 1548-7105
• DOI: 10.1038/s41592-018-0229-2

Single-cell transcriptome measurements can reveal unexplored biological diversity, but they suffer from technical noise and bias that must be modeled to account for the resulting uncertainty in downstream analyses. Here we introduce single-cell variational inference (scVI), a ready-to-use scalable framework for the probabilistic representation and analysis of gene expression in single cells ( https://github.com/YosefLab/scVI ). scVI uses stochastic optimization and deep neural networks to aggregate information across similar cells and genes and to approximate the distributions that underlie observed expression values, while accounting for batch effects and limited sensitivity. We used scVI for a range of fundamental analysis tasks including batch correction, visualization, clustering, and differential expression, and achieved high accuracy for each task. scVI is a ready-to-use generative deep learning tool for large-scale single-cell RNA-seq data that enables raw data processing and a wide range of rapid and accurate downstream analyses.