Deep learning models predict regulatory variants in pancreatic islets and refine type 2 diabetes association signals
- University of Oxford, United Kingdom
- Genentech, United States
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.
Data availability
The datasets analysed during the current study are available in the public repositories under accessions listed in STable 1.
-
Integration of human pancreatic islet genomic data refines regulatory mechanisms at Type 2 Diabetes susceptibility lociEuropean Genome-Phenome Archive, EGAS00001002592.
-
Integration of ATAC-seq and RNA-seq Identifies Human Alpha Cell and Beta Cell Signature GenesNCBI Gene Expression Omnibus, GSE76268.
-
Duke_DnaseSeq_PanIsletsNCBI Gene Expression Omnibus, GSM816660.
-
UNC_FaireSeq_PanIsletsNCBI Gene Expression Omnibus, GSM864346.
-
DNaseI/FAIRE/ChIP Synthesis from ENCODE/OpenChrom(Duke/UNC/UTA)NCBI Gene Expression Omnibus, GSE40833.
-
BI Human Reference Epigenome Mapping Project: ChIP-Seq in human subjectNCBI Gene Expression Omnibus, GSE19465.
-
UCSF-UBC Human Reference Epigenome Mapping ProjectNCBI Gene Expression Omnibus, GSE16368.
-
Epigenomic plasticity enables human pancreatic alpha to beta cell reprogrammingNCBI Gene Expression Omnibus, GSE50386.
Article and author information
Author details
Agata Wesolowska-Andersen
Funding
Wellcome (099673/Z/12/Z)
- Matthias Thurner
Horizon 2020 Framework Programme (T2D Systems)
- Anna L Gloyn
NIH Clinical Center (U01-DK105535)
- Anna L Gloyn
- Mark I McCarthy
NIH Clinical Center (U01-DK085545)
- Anna L Gloyn
National Institute for Health Research (NF-SI-0617-10090)
- Anna L Gloyn
Wellcome (090532)
- Mark I McCarthy
Wellcome (106130)
- Anna L Gloyn
- Mark I McCarthy
Wellcome (098381)
- Mark I McCarthy
Wellcome (203141)
- Anna L Gloyn
- Mark I McCarthy
Wellcome (212259)
- Mark I McCarthy
Wellcome (095101)
- Anna L Gloyn
Wellcome (200837)
- Anna L Gloyn
Medical Research Council (MR/L020149/1)
- Anna L Gloyn
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
Copyright
© 2020, Wesolowska-Andersen et al.
This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 4,143
- views
-
- 629
- downloads
-
- 27
- citations
Views, downloads and citations are aggregated across all versions of this paper published by eLife.
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Deep learning models predict regulatory variants in pancreatic islets and refine type 2 diabetes association signals
eLife 9:e51503.
https://doi.org/10.7554/eLife.51503
Further reading
-
- Computational and Systems Biology
- Evolutionary Biology
As pathogens spread in a population of hosts, immunity is built up, and the pool of susceptible individuals are depleted. This generates selective pressure, to which many human RNA viruses, such as influenza virus or SARS-CoV-2, respond with rapid antigenic evolution and frequent emergence of immune evasive variants. However, the host’s immune systems adapt, and older immune responses wane, such that escape variants only enjoy a growth advantage for a limited time. If variant growth dynamics and reshaping of host-immunity operate on comparable time scales, viral adaptation is determined by eco-evolutionary interactions that are not captured by models of rapid evolution in a fixed environment. Here, we use a Susceptible/Infected model to describe the interaction between an evolving viral population in a dynamic but immunologically diverse host population. We show that depending on strain cross-immunity, heterogeneity of the host population, and durability of immune responses, escape variants initially grow exponentially, but lose their growth advantage before reaching high frequencies. Their subsequent dynamics follows an anomalous random walk determined by future escape variants and results in variant trajectories that are unpredictable. This model can explain the apparent contradiction between the clearly adaptive nature of antigenic evolution and the quasi-neutral dynamics of high-frequency variants observed for influenza viruses.
-
- Computational and Systems Biology
- Medicine
Sudden death after myocardial infarction (MI) is associated with electrophysiological heterogeneities and ionic current remodelling. Low ejection fraction (EF) is used in risk stratification, but its mechanistic links with pro-arrhythmic heterogeneities are unknown. We aim to provide mechanistic explanations of clinical phenotypes in acute and chronic MI, from ionic current remodelling to ECG and EF, using human electromechanical modelling and simulation to augment experimental and clinical investigations. A human ventricular electromechanical modelling and simulation framework is constructed and validated with rich experimental and clinical datasets, incorporating varying degrees of ionic current remodelling as reported in literature. In acute MI, T-wave inversion and Brugada phenocopy were explained by conduction abnormality and local action potential prolongation in the border zone. In chronic MI, upright tall T-waves highlight large repolarisation dispersion between the border and remote zones, which promoted ectopic propagation at fast pacing. Post-MI EF at resting heart rate was not sensitive to the extent of repolarisation heterogeneity and the risk of repolarisation abnormalities at fast pacing. T-wave and QT abnormalities are better indicators of repolarisation heterogeneities than EF in post-MI.
