scvi-tools: Enabling Probabilistic Analysis for Single-Cell Genomics
Nir Yosef (University of California, Berkeley)
To maintain and further develop a community resource for probabilistic analysis of single-cell omics data, including an application interface for rapid development of new probabilistic models.
Probabilistic models have demonstrated state-of-the-art performance across a variety of single-cell omics analysis tasks; however, computational barriers limit their use. The goal of this proposal is to simplify both the use and further development of these statistically-sound models. scvi-tools is already among the most popular packages for single-cell data analysis in Python, with a growing user base of analysts and methods developers, culminating in thousands of downloads and hundreds of active users. This funding will support skilled machine learning engineers to focus on the continued deployment of emerging methods, supporting the end-user community, promoting, troubleshooting, and advancing model development in scvi-tools by other groups, and furthering integrations of emerging and powerful machine learning tools.