The advent of single-cell RNA sequencing (scRNA-seq) technologies has revolutionized transcriptomic studies. However, large-scale integrative analysis of scRNA-seq data remains a challenge largely due to unwanted batch effects and the limited transferabilty, interpretability, and scalability of the existing computational methods. We present single-cell Embedded Topic Model (scETM). Our key contribution is the utilization of a transferable neural-network-based encoder while having an interpretable linear decoder via a matrix tri-factorization. In particular, scETM simultaneously learns an encoder network to infer cell type mixture and a set of highly interpretable gene embeddings, topic embeddings, and batch-effect linear intercepts from multiple scRNA-seq datasets. scETM is scalable to over 106 cells and confers remarkable cross-tissue and cross-species zero-shot transfer-learning performance. Using gene set enrichment analysis, we find that scETM-learned topics are enriched in biologically meaningful and disease-related pathways. Lastly, scETM enables the incorporation of known gene sets into the gene embeddings, thereby directly learning the associations between pathways and topics via the topic embeddings.

The advent of single-cell RNA sequencing (scRNA-seq) technologies has revolutionized transcriptomic studies. However, integrative analysis of scRNA-seq data remains a challenge largely due to batch effects. We present single-cell Embedded Topic Model (scETM), an unsupervised deep generative model that recapitulates known cell types by inferring the latent cell topic mixtures via a variational autoencoder. scETM is scalable to over 106 cells and enables effective knowledge transfer across datasets. scETM also offers high inter-pretability and allows the incorporation of prior pathway knowledge into the gene embeddings. The scETM-inferred topics show enrichment in cell-type-specific and disease-related pathways.

Motivation Single-cell RNA-sequencing (scRNA-seq) offers the opportunity to dissect heterogeneous cellular compositions and interrogate the cell-type-specific gene expression patterns across diverse conditions. However, batch effects such as laboratory conditions and individual-variability hinder their usage in cross-condition designs. Results Here, we present a single-cell Generative Adversarial Network (scGAN) to simultaneously acquire patterns from raw data while minimizing the confounding effect driven by technical artifacts or other factors inherent to the data. Specifically, scGAN models the data likelihood of the raw scRNA-seq counts by projecting each cell onto a latent embedding. Meanwhile, scGAN attempts to minimize the correlation between the latent embeddings and the batch labels across all cells. We demonstrate scGAN on three public scRNA-seq datasets and show that our method confers superior performance over the state-of-the-art methods in forming clusters of known cell types and identifying known psychiatric genes that are associated with major depressive disorder. Availability The scGAN code and the information for the public scRNA-seq datasets are available at https://github.com/li-lab-mcgill/singlecell-deepfeature Supplementary information Supplementary data are available at Bioinformatics online.

The recent proliferation of large scale genome-wide association studies (GWASs) has motivated the development of statistical methods for phenotype prediction using single nucleotide polymorphism (SNP) array data. These polygenic risk score (PRS) methods formulate the task of polygenic prediction in terms of a multiple linear regression framework, where the goal is to infer the joint effect sizes of all genetic variants on the trait. Among the subset of PRS methods that operate on GWAS summary statistics, sparse Bayesian methods have shown competitive predictive ability. However, existing Bayesian approaches employ Markov Chain Monte Carlo (MCMC) algorithms for posterior inference, which are computationally inefficient and do not scale favorably with the number of SNPs included in the analysis. Here, we introduce Variational Inference of Polygenic Risk Scores (VIPRS), a Bayesian summary statistics-based PRS method that utilizes Variational Inference (VI) techniques to efficiently approximate the posterior distribution for the effect sizes. Our experiments with genome-wide simulations and real phenotypes from the UK Biobank (UKB) dataset demonstrated that variational approximations to the posterior are competitively accurate and highly efficient. When compared to state-of-the-art PRS methods, VIPRS consistently achieves the best or second best predictive accuracy in our analyses of 18 simulation configurations as well as 12 real phenotypes measured among the UKB participants of ``White British'' background. This performance advantage was higher among individuals from other ethnic groups, with an increase in R-squared of up to 1.7-fold among participants of Nigerian ancestry for Low-Density Lipoprotein (LDL) cholesterol. Furthermore, given its computational efficiency, we applied VIPRS to a dataset of up to 10 million genetic markers, an order of magnitude greater than the standard HapMap3 subset used to train existing PRS methods. Modeling this expanded set of variants conferred modest improvements in prediction accuracy for a number of highly polygenic traits, such as standing height.

No abstract