MultiVI: deep generative model for the integration of multimodal data

Ashuach, Tal; Gabitto, Mariano I.; Koodli, Rohan V.; Saldi, Giuseppe-Antonio; Jordan, Michael I.; Yosef, Nir

doi:10.1038/s41592-023-01909-9

Cited by 82 publications

(59 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1A). We then integrated the spot-by-cell type abundance data (39,563 spots) from 22 Visium sections (Fig. 1E).…”

Section: Spatial Transcriptomics Reveals Multicellular Nichesmentioning

confidence: 99%

See 1 more Smart Citation

Multiomic analysis reveals developmental dynamics of the human heart in health and disease

Cranley,

Kanemaru,

Bayraktar

et al. 2024

Preprint

View full text Add to dashboard Cite

Developmental dynamics encompass both the specification of cell types and their spatial organisation into multicellular niches. Here we harness the power of single-cell and spatial multiomics to unravel embryonic and foetal cardiac tissue niches, which lead to the development of a new tool, TissueTypist. We infer developmental cell trajectories, including evidence for lineage relationships based on shared somatic mutations, within first- and second-trimester human hearts. We reveal that cardiac-resident macrophages likely originate from the yolk sac, forming heterogeneous subsets. CX3CR1+ macrophages with a microglia-like profile localise in the sinoatrial node, which may contribute to axon guidance for the innervating autonomic neurons. Foetal pacemaker cells exhibit distinct characteristics compared to their adult counterparts, including the expression of genes that are known to promote parasympathetic innervation. By comparing somatic mutation profiles of cardiomyocytes, we identify an early branching point where pacemaker cells diverge from working cardiomyocytes. We highlight the enhancer-mediated gene regulatory networks governing atrial and ventricular cardiomyocyte specification. The maturation of atrial cardiomyocytes into distinct left and right phenotypes, driven by transcription factors linked to atrial septal defect genes, underscores the significance of this process for healthy heart development. In the ventricle, cellular and transcriptional gradients along both pseudotime and the transmural axis provide a new molecular understanding of myocardial compaction. Finally, generating data from Trisomy 21 hearts and comparing this with the euploid atlas, we reveal a reduced abundance of specific cell types including compact cardiomyocytes. Overall, this extensive dataset and our precomputed models will form a valuable resource for the field.

show abstract

“…1A). We then integrated the spot-by-cell type abundance data (39,563 spots) from 22 Visium sections (Fig. 1E).…”

Section: Spatial Transcriptomics Reveals Multicellular Nichesmentioning

confidence: 99%

“…However, in humans, tissue localisations and potential functions of cardiac macrophages in development are not fully clarified. To uncover this, we utilised gene expression and chromatin accessibility data of macrophages and monocytes and created an integrated embedding using multiVI 22 (Fig. 2A).…”

Section: Cardiac Macrophage Developmentmentioning

confidence: 99%

Multiomic analysis reveals developmental dynamics of the human heart in health and disease

Cranley,

Kanemaru,

Bayraktar

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Diagonal integration typically aims to construct a low‐dimensional latent space that captures the correlation between the data modalities, however even if gene expression and chromatin accessibility are correlated there is no guarantee that the latent representation can capture this information, making it a difficult endeavor. Commonly used diagonal integration methods include GLUE, LIGER, Cobolt, MultiVI, and Seurat V5, to name a few 112–116 . Second, current computational analysis methods are limited in their ability to learn the intricate relationship and cross talk between different data modalities.…”

Section: Single‐nucleus Transcriptomics Elucidates Cell Type–specific...mentioning

confidence: 99%

Gene networks and systems biology in Alzheimer's disease: Insights from multi‐omics approaches

Rahimzadeh,

Srinivasan,

Zhang

et al. 2024

Alzheimer's & Dementia

View full text Add to dashboard Cite

Despite numerous studies in the field of dementia and Alzheimer's disease (AD), a comprehensive understanding of this devastating disease remains elusive. Bulk transcriptomics have provided insights into the underlying genetic factors at a high level. Subsequent technological advancements have focused on single‐cell omics, encompassing techniques such as single‐cell RNA sequencing and epigenomics, enabling the capture of RNA transcripts and chromatin states at a single cell or nucleus resolution. Furthermore, the emergence of spatial omics has allowed the study of gene responses in the vicinity of amyloid beta plaques or across various brain regions. With the vast amount of data generated, utilizing gene regulatory networks to comprehensively study this disease has become essential. This review delves into some techniques employed in the field of AD, explores the discoveries made using these techniques, and provides insights into the future of the field.

show abstract

“…A number of computational approaches have been designed to perform joint dimension reduction on single-cell multimodal data types such as CITE-seq and 10X multiome. For example, the weighted nearest neighbor algorithm, the multi-omic factor analysis (), and the perform linear or non-linear dimension reduction on single-cell multimodal datasets that can be represented as genes and peaks [3–5]. Approaches based on variational autoencoders (VAEs) are especially powerful for learning joint representations from single-cell multimodal data.…”

Section: Introductionmentioning

confidence: 99%

Integrating single-cell multimodal epigenomic data using 1D-convolutional neural networks

Gao,

Welch

2024

Preprint

View full text Add to dashboard Cite

Recent experimental developments enable single-cell multimodal epigenomic profiling, which measures multiple histone modifications and chromatin accessibility within the same cell. Such parallel measurements provide exciting new opportunities to investigate how epigenomic modalities vary together across cell types and states. A pivotal step in using this type of data is integrating the epigenomic modalities to learn a unified representation of each cell, but existing approaches are not designed to model the unique nature of this data type. Our key insight is to model single-cell multimodal epigenome data as a multi-channel sequential signal. Based on this insight, we developed ConvNet-VAEs, a novel framework that uses 1D-convolutional variational autoencoders (VAEs) for single-cell multimodal epigenomic data integration. We evaluated ConvNet-VAEs on nano-CT and scNTT-seq data generated from juvenile mouse brain and human bone marrow. We found that ConvNet-VAEs can perform dimension reduction and batch correction better than previous architectures while using significantly fewer parameters. Furthermore, the performance gap between convolutional and fully-connected architectures increases with the number of modalities, and deeper convolutional architectures can increase performance while performance degrades for deeper fully-connected architectures. Our results indicate that convolutional autoencoders are a promising method for integrating current and future single-cell multimodal epigenomic datasets.

show abstract

MultiVI: deep generative model for the integration of multimodal data

Cited by 82 publications

References 34 publications

Multiomic analysis reveals developmental dynamics of the human heart in health and disease

Multiomic analysis reveals developmental dynamics of the human heart in health and disease

Gene networks and systems biology in Alzheimer's disease: Insights from multi‐omics approaches

Integrating single-cell multimodal epigenomic data using 1D-convolutional neural networks

Contact Info

Product

Resources

About