MultiVI: deep generative model for the integration of multi-modal data

Abstract

The ability to jointly profile the transcriptional and chromatin land-scape of single-cells has emerged as a powerful technique to identify cellular populations and shed light on their regulation of gene expression. Current computational methods analyze jointly profiled (paired) or individual data modalities (unpaired), but do not offer a principled method to analyze both paired and unpaired samples jointly. Here we present MultiVI, a probabilistic framework that leverages deep neural networks to jointly analyze scRNA, scATAC and multiomic (scRNA + scATAC) data. MultiVI creates an informative low-dimensional latent space that accurately reflects both chromatin and transcriptional properties of the cells even when one of the modalities is missing. MultiVI accounts for technical effects in both scRNA and scATAC-seq while correcting for batch effects in both data modalities. We use public datasets to demonstrate that MultiVI is stable, easy to use, and outperforms current approaches for the joint analysis of paired and unpaired data. MultiVI is available as an open source package, implemented in the scvi-tools frame-work: https://docs.scvi-tools.org/.

Dictionary learning for integrative, multimodal, and scalable single-cell analysis

“…We next compared the performance of bridge integration against two recently proposed methods for integrated analysis of multimodal and single-modality datasets. Both multiVI 48 and Cobolt 49 utilize variational autoencoders for integration, and while they do not explicitly treat multi-omic datasets as a bridge, they aim to integrate datasets across technologies and modalities into a shared space. When applied to the previously described datasets, both methods were broadly successful in integrating scRNA-seq and scATAC-seq data, but did not identify matches at the same level of resolution (for example, neither method successfully matched ASDC in scATAC-seq data to the ASDC in the Azimuth reference) ( Fig.…”

Identifying commonalities between cell lines and tumors at the single cell level using Sobolev Alignment of deep generative models

“…Although we restricted our analysis to single cell gene expression data, our computational approach is versatile, and can easily be adapted for other molecular features, such as chromatin accessibility 56 , protein levels 57 or ribosomal profiling 58 . Recently developed deep probabilistic models tailored for such data 59,60 could be for instance employed to adapt our framework. We also solely compare cell line models to human tumors, but more complex models could be studied equally well, such as organoids or patient derived xenografts.…”

“…Recently developed deep probabilistic models tailored for such data 59,60 could be for instance employed to adapt our framework. We also solely compare cell line models to human tumors, but more complex models could be studied equally well, such as organoids or patient derived xenografts.…”

Integrated single-cell (phospho-)protein and RNA detection uncovers phenotypic characteristics of human antibody secreting cells

“…Computational strategies encounter several considerations as how to define anchors, scalability and handling missing data ( 43 ). Several of these challenges are being addressed by recently developed tools including MOFA+ ( 24 ), multiVI ( 44 ), COBOLT ( 45 ), StabMap ( 46 ) scMVP ( 47 ), and Bridge Integration ( 48 ). So-far there was no illustration of integrating mRNA datasets with a comprehensive intracellular phospho-protein and transcription factor dataset using a common set of surface proteins.…”