Mechanistic modeling with a variational autoencoder for multimodal single-cell RNA sequencing data

Creators: Carilli, Maria; Gorin, Gennady; Choi, Yongin; Chari, Tara; Pachter, Lior

Style

An error occurred while generating the citation.

Abstract

We motivate and present biVI, which combines the variational autoencoder framework of scVI with biophysically motivated, bivariate models for nascent and mature RNA distributions. In simulated benchmarking, biVI accurately recapitulates key properties of interest, including cell type structure, parameter values, and copy number distributions. In biological datasets, biVI provides a route for the identification of the biophysical mechanisms underlying differential expression. The analytical approach outlines a generalizable strategy for representing multimodal datasets generated by single-cell RNA sequencing.

Additional Information

The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. M.C., G.G., T.C., and L.P. were partially funded by IGVF-1-UCI.IGVF and NIH U19MH114830. Y.C. was partially funded by T32 GM007377. G.G. thanks Drs. Ido Golding and Heng Xu for the inspiration leading to the explanatory model for the zero-inflated negative binomial distribution in Section S1.4. The RNA illustrations used in Figures 1, 2, S1, and S2 were derived from the DNA Twemoji by Twitter, Inc., used under the CC-BY 4.0 license. We thank the Caltech Bioinformatics Resource Center for GPU resources that helped in performing the analyses. Data availability. Simulated datasets, simulated parameters used to generate them, and Allen dataset B08 and its associated metadata are available in the Zenodo package 7497222. All analysis scripts and notebooks are available at https://github.com/pachterlab/CGCCP_2023. The repository also contains a Google Colaboratory demonstration notebook applying the methods to a small human blood cell dataset. The authors have declared no competing interest.

Attached Files

Submitted - 2023.01.13.523995v1.full.pdf

Supplemental Material - media-1.pdf

Files

2023.01.13.523995v1.full.pdf

Files (12.3 MB)

Name	Size	Download all
2023.01.13.523995v1.full.pdf md5:29f55873752dcb65b857a22e4b5c8c9a	5.0 MB	Preview Download
media-1.pdf md5:45b7525ab6332258610c9d1a8397d051	7.3 MB	Preview Download

Additional details

	All versions	This version
Views	27	27
Downloads	33	33
Data volume	229.6 MB	229.6 MB