TY - JOUR T1 - GeneVector: Identification of transcriptional programs using dense vector representations defined by mutual information JF - bioRxiv DO - 10.1101/2022.04.22.487554 SP - 2022.04.22.487554 AU - Nicholas Ceglia AU - Zachary Sethna AU - Florian Uhlitz AU - Viktoria Bojilova AU - Nicole Rusk AU - Bharat Burman AU - Andrew Chow AU - Sohrab Salehi AU - Farhia Kabeer AU - Samuel Aparicio AU - Benjamin Greenbaum AU - Sohrab P. Shah AU - Andrew McPherson Y1 - 2022/01/01 UR - http://biorxiv.org/content/early/2022/05/02/2022.04.22.487554.abstract N2 - Deciphering individual cell phenotypes from cell-specific transcriptional processes requires high dimensional single cell RNA sequencing. However, current dimensionality reduction methods aggregate sparse gene information across similar cells, without directly measuring the relationships that exist between genes. By performing dimensionality reduction with respect to gene co-expression, low-dimensional features can model these gene-specific relationships and leverage shared signal to overcome sparsity. We describe GeneVector, a scalable framework for dimensionality reduction implemented as a vector space model using mutual information between gene expression. Unlike other methods, including principal component analysis and variational autoencoders, GeneVector uses latent space arithmetic in a lower dimensional gene embedding to identify transcriptional programs and classify cell types. Using four single cell RNA-seq datasets, we show that GeneVector was able to capture phenotype-specific pathways, perform batch effect correction, interactively annotate cell types, and identify pathway variation with treatment over time. GeneVector is available as an open source python package at https://github.com/nceglia/genevector.Competing Interest StatementThe authors have declared no competing interest. ER -