Deep Generative Models for Single-cell Transcriptomics
An artistic representation of single-cell RNA sequencing.
Single-cell Variational Inference (scVI) is a set of tools for probabilistic analysis of single-cell RNA-sequencing data. We propose to solve quantitative biological questions with tailored and scalable Bayesian machinery. Examples of biological questions include cell-type discovery, data integration, transfer of annotations and differential expression analysis. Examples of statistical tools include deep generative models, stochastic variational inference, Bayesian decision theory and domain adaptation.
All the models in our codebase may fit millions of transcriptomic measurements in a few hours. In addition to the original publication, we recommend the following ressources to get more familiar with scVI:
- our codebase and documentation which includes user-friendly notebooks
- the podcast recorded on the bioinformatics chat
- the YosefLab blog usually includes discussions about special topics
- the News & View in Nature Methods
- our post on the Berkeley AI Research blog
Romain Lopez
Postdoctoral Scholar
My research interests lie at the intersection of statistics, computation and modeling. A significant part of my research is driven by building more statistically accurate and faster machine learning software for analyzing biological data, with a focus on single-cell transcriptomics.
Posts
Bayesian Nets: from white-boarding to laser cutting
