An Empirical Bayes Method for Differential Expression Analysis of Single Cells with Deep Generative Models

Boyeau, Pierre; Regier, Jeffrey; Gayoso, Adam; Jordan, Michael I.; Lopez, Romain; Yosef, Nir

doi:10.1101/2022.05.27.493625

Cited by 17 publications

(25 citation statements)

References 88 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…S10 ). Furthermore, this procedure can be naturally generalized for differential expression testing between case and control groups by applying the Bayesian inference framework ( 23 , 24 ).…”

Section: Resultsmentioning

confidence: 99%

Robust probabilistic modeling for single-cell multimodal mosaic integration and imputation via scVAEIT

Cai

Roeder

2022

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

Recent advances in single-cell technologies enable joint profiling of multiple omics. These profiles can reveal the complex interplay of different regulatory layers in single cells; still, new challenges arise when integrating datasets with some features shared across experiments and others exclusive to a single source; combining information across these sources is called mosaic integration. The difficulties lie in imputing missing molecular layers to build a self-consistent atlas, finding a common latent space, and transferring learning to new data sources robustly. Existing mosaic integration approaches based on matrix factorization cannot efficiently adapt to nonlinear embeddings for the latent cell space and are not designed for accurate imputation of missing molecular layers. By contrast, we propose a probabilistic variational autoencoder model, scVAEIT, to integrate and impute multimodal datasets with mosaic measurements. A key advance is the use of a missing mask for learning the conditional distribution of unobserved modalities and features, which makes scVAEIT flexible to combine different panels of measurements from multimodal datasets accurately and in an end-to-end manner. Imputing the masked features serves as a supervised learning procedure while preventing overfitting by regularization. Focusing on gene expression, protein abundance, and chromatin accessibility, we validate that scVAEIT robustly imputes the missing modalities and features of cells biologically different from the training data. scVAEIT also adjusts for batch effects while maintaining the biological variation, which provides better latent representations for the integrated datasets. We demonstrate that scVAEIT significantly improves integration and imputation across unseen cell types, different technologies, and different tissues.

show abstract

Section: Resultsmentioning

confidence: 99%

Robust probabilistic modeling for single-cell multimodal mosaic integration and imputation via scVAEIT

Cai

Roeder

2022

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

show abstract

“…S10). Furthermore, this procedure can be naturally generalized for differential expression testing between case and control groups by applying the Bayesian inference framework [18, 5].…”

Section: Resultsmentioning

confidence: 99%

Robust probabilistic modeling for single-cell multimodal mosaic integration and imputation via scVAEIT

Cai

Roeder

2022

Preprint

View full text Add to dashboard Cite

While emerging multi-omics single-cell datasets provides new insights into complex biological processes, it remains challenging to perform mosaic integration, which aims at combining datasets with different panels of measurements. Utilizing unshared features is valuable for intermediate integration when building a multimodal atlas, and also useful for late integration when mapping reference datasets to the atlas. However, existing mosaic integration approaches based on matrix factorization cannot efficiently adapt to non-linear embedding in intermediate integration and are hard to extend to late integration. In this paper, we propose a probabilistic variational autoencoder model to integrate and impute multimodal datasets with mosaic measurements. Our model incorporates a missing mask for learning the distribution of unobserved modalities and features conditioned on an arbitrary subset of unobserved features. We show that the masking procedure allows for generalization across unseen cell types, different technologies, and different tissues. It also significantly improves the integration of DOGMA-seq, CITE-seq, and ASAP-seq datasets.

show abstract

“…Here we focused on estimating s het , but our empirical Bayes framework, GeneBayes, can be used in any setting where one has a model that ties a gene-level parameter to gene-level observable data (Supplementary Note D). For example, GeneBayes can be used to find trait-associated genes using variants from case/control studies [71,72], or to improve power to find differentially expressed genes in RNA-seq experiments [73]. We provide a graphical overview of how GeneBayes can be applied more generally in Figure 6.…”

Section: Discussionmentioning

confidence: 99%

Bayesian estimation of gene constraint from an evolutionary model with gene features

Zeng

Spence

Mostafavi

et al. 2023

Preprint

View full text Add to dashboard Cite

Measures of selective constraint on genes have been used for many applications including clinical interpretation of rare coding variants, disease gene discovery, and studies of genome evolution. However, widely-used metrics are severely underpowered at detecting constraint for the shortest ∼25% of genes, potentially causing important pathogenic mutations to be overlooked. We developed a framework combining a population genetics model with machine learning on gene features to enable accurate inference of an interpretable constraint metric, shet. Our estimates outperform existing metrics for prioritizing genes important for cell essentiality, human disease, and other phenotypes, especially for short genes. Our new estimates of selective constraint should have wide utility for characterizing genes relevant to human disease. Finally, our inference framework, GeneBayes, provides a flexible platform that can improve estimation of many gene-level properties, such as rare variant burden or gene expression differences.

show abstract

An Empirical Bayes Method for Differential Expression Analysis of Single Cells with Deep Generative Models

Cited by 17 publications

References 88 publications

Robust probabilistic modeling for single-cell multimodal mosaic integration and imputation via scVAEIT

Robust probabilistic modeling for single-cell multimodal mosaic integration and imputation via scVAEIT

Robust probabilistic modeling for single-cell multimodal mosaic integration and imputation via scVAEIT

Bayesian estimation of gene constraint from an evolutionary model with gene features

Contact Info

Product

Resources

About