Dataset Information

Deep learning models predict regulatory variants in pancreatic islets and refine type 2 diabetes association signals.

ABSTRACT: Genome-wide association analyses have uncovered multiple genomic regions associated with T2D, but identification of the causal variants at these remains a challenge. There is growing interest in the potential of deep learning models - which predict epigenome features from DNA sequence - to support inference concerning the regulatory effects of disease-associated variants. Here, we evaluate the advantages of training convolutional neural network (CNN) models on a broad set of epigenomic features collected in a single disease-relevant tissue - pancreatic islets in the case of type 2 diabetes (T2D) - as opposed to models trained on multiple human tissues. We report convergence of CNN-based metrics of regulatory function with conventional approaches to variant prioritization - genetic fine-mapping and regulatory annotation enrichment. We demonstrate that CNN-based analyses can refine association signals at T2D-associated loci and provide experimental validation for one such signal. We anticipate that these approaches will become routine in downstream analyses of GWAS.

SUBMITTER: Wesolowska-Andersen A

PROVIDER: S-EPMC7007221 | biostudies-literature | 2020 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning models predict regulatory variants in pancreatic islets and refine type 2 diabetes association signals.

Wesolowska-Andersen Agata A Zhuo Yu Grace G Nylander Vibe V Abaitua Fernando F Thurner Matthias M Torres Jason M JM Mahajan Anubha A Gloyn Anna L AL McCarthy Mark I MI

eLife 20200127

Genome-wide association analyses have uncovered multiple genomic regions associated with T2D, but identification of the causal variants at these remains a challenge. There is growing interest in the potential of deep learning models - which predict epigenome features from DNA sequence - to support inference concerning the regulatory effects of disease-associated variants. Here, we evaluate the advantages of training convolutional neural network (CNN) models on a broad set of epigenomic features ...[more]

PMID: 31985400

Similar Datasets

Project description:OBJECTIVE:Type 2 diabetes (T2D) is a complex disease characterized by pancreatic islet dysfunction, insulin resistance, and disruption of blood glucose levels. Genome-wide association studies (GWAS) have identified > 400 independent signals that encode genetic predisposition. More than 90% of associated single-nucleotide polymorphisms (SNPs) localize to non-coding regions and are enriched in chromatin-defined islet enhancer elements, indicating a strong transcriptional regulatory component to disease susceptibility. Pancreatic islets are a mixture of cell types that express distinct hormonal programs, so each cell type may contribute differentially to the underlying regulatory processes that modulate T2D-associated transcriptional circuits. Existing chromatin profiling methods such as ATAC-seq and DNase-seq, applied to islets in bulk, produce aggregate profiles that mask important cellular and regulatory heterogeneity. METHODS:We present genome-wide single-cell chromatin accessibility profiles in >1,600 cells derived from a human pancreatic islet sample using single-cell combinatorial indexing ATAC-seq (sci-ATAC-seq). We also developed a deep learning model based on U-Net architecture to accurately predict open chromatin peak calls in rare cell populations. RESULTS:We show that sci-ATAC-seq profiles allow us to deconvolve alpha, beta, and delta cell populations and identify cell-type-specific regulatory signatures underlying T2D. Particularly, T2D GWAS SNPs are significantly enriched in beta cell-specific and across cell-type shared islet open chromatin, but not in alpha or delta cell-specific open chromatin. We also demonstrate, using less abundant delta cells, that deep learning models can improve signal recovery and feature reconstruction of rarer cell populations. Finally, we use co-accessibility measures to nominate the cell-specific target genes at 104 non-coding T2D GWAS signals. CONCLUSIONS:Collectively, we identify the islet cell type of action across genetic signals of T2D predisposition and provide higher-resolution mechanistic insights into genetically encoded risk pathways.

Project description:The function and viability of cultured, transplanted, or encapsulated pancreatic islets is often limited by hypoxia because these islets have lost their vasculature during the isolation process and have to rely on gradient-driven passive diffusion, which cannot provide adequate oxygen transport. Pancreatic islets (islets of Langerhans) are particularly susceptible due to their relatively large size, large metabolic demand, and increased sensitivity to hypoxia. Here, finite element method (FEM) based multiphysics models are explored to describe oxygen transport and cell viability in avascular islets both in static and in moving culture media.Two- and three-dimensional models were built in COMSOL Multiphysics using the convection and diffusion as well as the incompressible Navier-Stokes fluid dynamics application modes. Oxygen consumption was assumed to follow Michaelis-Menten-type kinetics and to cease when local concentrations fell below a critical threshold; in a dynamic model, it was also allowed to increase with increasing glucose concentration.Partial differential equation (PDE) based exploratory cellular-level oxygen consumption and cell viability models incorporating physiologically realistic assumptions have been implemented for fully scaled cell culture geometries with 100, 150, and 200 microm diameter islets as representative. Calculated oxygen concentrations and intra-islet regions likely to suffer from hypoxia-related necrosis obtained for traditional flask-type cultures, oxygen-permeable silicone-rubber membrane bottom cultures, and perifusion chambers with flowing media and varying incoming glucose levels are presented in detail illustrated with corresponding colour-coded figures and animations.Results of the computational models are, as a first estimate, in good quantitative agreement with existing experimental evidence, and they confirm that during culture, hypoxia is often a problem for non-vascularised islet and can lead to considerable cell death (necrosis), especially in the core region of larger islets. Such models are of considerable interest to improve the function and viability of cultured, transplanted, or encapsulated islets. The present implementation allows convenient extension to true multiphysics applications that solve coupled physics phenomena such as diffusion and consumption with convection due to flowing or moving media.

Dataset Information

Deep learning models predict regulatory variants in pancreatic islets and refine type 2 diabetes association signals.

Publications

Deep learning models predict regulatory variants in pancreatic islets and refine type 2 diabetes association signals.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets