Dataset Information

Deep learning predicts DNA methylation regulatory variants in specific brain cell types and enhances fine mapping for brain disorders.

ABSTRACT: DNA methylation (DNAm) is essential for brain development and function and potentially mediates the effects of genetic risk variants underlying brain disorders. We present INTERACT, a transformer-based deep learning model to predict regulatory variants impacting DNAm levels in specific brain cell types, leveraging existing single-nucleus DNAm data from the human brain. We show that INTERACT accurately predicts cell type-specific DNAm profiles, achieving an average area under the Receiver Operating Characteristic curve of 0.98 across cell types. Furthermore, INTERACT predicts cell type-specific DNAm regulatory variants, which reflect cellular context and enrich the heritability of brain-related traits in relevant cell types. Importantly, we demonstrate that incorporating predicted variant effects and DNAm levels of CpG sites enhances the fine mapping for three brain disorders-schizophrenia, depression, and Alzheimer's disease-and facilitates mapping causal genes to particular cell types. Our study highlights the power of deep learning in identifying cell type-specific regulatory variants, which will enhance our understanding of the genetics of complex traits.

SUBMITTER: Zhou J

PROVIDER: S-EPMC10827166 | biostudies-literature | 2024 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning predicts DNA methylation regulatory variants in specific brain cell types and enhances fine mapping for brain disorders.

Zhou Jiyun J Weinberger Daniel R DR Han Shizhong S

bioRxiv : the preprint server for biology 20240121

DNA methylation (DNAm) is essential for brain development and function and potentially mediates the effects of genetic risk variants underlying brain disorders. We present INTERACT, a transformer-based deep learning model to predict regulatory variants impacting DNAm levels in specific brain cell types, leveraging existing single-nucleus DNAm data from the human brain. We show that INTERACT accurately predicts cell type-specific DNAm profiles, achieving an average area under the Receiver Operati ...[more]

PMID: 38293210

Similar Datasets

Project description:5-methylcytosine (5mC) is the most common chemical modification occurring on the CpG sites across the human genome. Bisulfite conversion combined with short-read whole genome sequencing can capture and quantify the modification at single nucleotide resolution. However, the PCR amplification process could lead to duplicative methylation patterns and introduce 5mC detection bias. Additionally, the limited read length also restricts co-methylation analysis between distant CpG sites. The bisulfite conversion process presents a significant challenge for detecting variant-specific methylation due to the destruction of allele information in the sequencing reads. To address these issues, we sought to characterize the human methylation profiling with the nanopore long-read sequencing, aiming to demonstrate its potential for long-range co-methylation analysis with native modification call and intact allele information retained. In this regard, we first analyzed the nanopore demo data in the adaptive sampling sequencing run targeting all human CpG islands. We applied the linkage disequilibrium (LD) R2 to calculate the co-methylation in nanopore data, and further identified 27,875, 50,481, 26,542 and 51,189 methylation haplotype blocks (MHB) in COLO829, COLO829BL, HCC1395 and HCC1395BL cell lines, respectively. Interestingly, while we found that majority of the co-methylation were in a short range (≤200bp), a small portion (1~3%) showed long distance (≥1,000bp), suggesting potential remote regulatory mechanisms across the genome. To further characterize the epigenetic changes related to transcription factor binding, we profiled the 5mC percentage changes surrounding various motif sites in JASPAR collection and found that CTCF and KLF5 binding sites showed reduced methylation, while FOXE1 and ZNF354A sites showed increased methylation. To further investigate the allele-specific 5mCG in the prostate genome, we designed a target region covering methylation quantitative trait loci (mQTL) and genome-wide association study (GWAS) risk germline variants and generated long reads with adaptive sampling run in the 22Rv1 cell line. To identify the allele-specific methylation in the 22Rv1 cell line, we performed long-read based phasing and compared the 5mCG signals between the two haplotypes. As a result, we identified 6,390 haplotype-specific methylated regions in the 22Rv1 cell line (p-MWU ≤ 1e-5 and delta ≥ 50%). By examining haplotype-specific methylated regions near the phasing variants, we identified examples of allele-specific methylated regions that showed allelespecific accessibility in the ATAC-seq data. By further integrating the ATAC-seq data of 22Rv1, we found that methylation levels were negatively correlated with chromatin accessibility at the genome-wide scale. Our study has revealed native methylome profiling while preserving haplotype information, offering a novel approach to uncovering the regulatory mechanisms of the human prostate genome.

Project description:Common genetic variants confer susceptibility to a large number of complex brain disorders. Given that such variants predominantly localize in non-coding regions of the human genome, there is a significant challenge to predict and characterize their functional consequences. More importantly, most available computational methods, generally defined as context-free methods, output prediction scores regarding the functionality of genetic variants irrespective of the context, i.e., the tissue or cell-type affected by a disease, limiting the ability to predict the functional consequences of common variants on brain disorders. In this study, we introduce a comparative multi-step pipeline to investigate the relative effectiveness of context-specific and context-free approaches to prioritize disease causal variants. As an experimental case, we focused on schizophrenia (SCZ), a debilitating neuropsychiatric disease for which a large number of susceptibility variants is identified from genome-wide association studies. We tested over two dozen available methods and examined potential associations between the cell/tissue-specific mapping scores and open chromatin accessibility, and provided a prioritized map of SCZ risk loci for in vitro or in-vivo functional analysis. We found extensive differences between context-free and tissue-specific approaches and showed how they may play complementary roles. As a proof of concept, we found a few sets of genes, through a consensus mapping of both categories, including FURIN to be among the top hits. We showed that the genetic variants in this gene and related genes collectively dysregulate gene expression patterns in stem cell-derived neurons and characterize SCZ phenotypic manifestations, while genes which were not shared among highly prioritized candidates in both approaches did not demonstrate such characteristics. In conclusion, by combining context-free and tissue-specific predictions, our pipeline enables prioritization of the most likely disease-causal common variants in complex brain disorders.

Dataset Information

Deep learning predicts DNA methylation regulatory variants in specific brain cell types and enhances fine mapping for brain disorders.

Publications

Deep learning predicts DNA methylation regulatory variants in specific brain cell types and enhances fine mapping for brain disorders.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets