TL;DR
VARAdb is a new, comprehensive database designed to annotate non-coding genetic variations in humans. It integrates a vast amount of genetic and epigenomic data, providing a scoring system to prioritize potentially functional variants and offering tools for downstream analysis, making it a valuable resource for understanding the role of non-coding variations in human diseases and biological processes.
What Problem Does This Paper Address?
The human genome is vast, and while much attention has been paid to coding regions, a significant portion of genetic variation occurs in non-coding regions. These non-coding variations are increasingly implicated in complex human diseases and phenotypes. However, identifying and interpreting their functional impact is challenging due to their location in regulatory elements like enhancers and promoters, and their complex interactions with genes. Existing databases often focus on coding variants or lack comprehensive annotation of regulatory information. This gap makes it difficult for researchers to effectively explore, prioritize, and understand the functional consequences of non-coding variations. VARAdb aims to bridge this gap by providing a unified, detailed, and user-friendly platform for annotating these crucial genetic elements.
Methods
VARAdb integrates data from numerous public sources and employs a multi-faceted approach to annotation. Key methods include:
- Data Integration: A massive collection of genetic and epigenomic data is curated and organized into five distinct annotation sections: 'Variation information,' 'Regulatory information,' 'Related genes,' 'Chromatin accessibility,' and 'Chromatin interaction.'
- Scoring Strategy: A novel scoring system is implemented, assigning a score from 0 to 9 to each variation based on its presence in nine annotation categories (risk SNP, eQTL, motif change, conservation, enhancer/super enhancer, promoter, TF binding, ATAC accessible region, and Hi-C). This allows users to prioritize variations with higher scores.
- Novel Variant Annotation: A dedicated tool is provided for annotating newly identified variants, enabling users to input their own data and receive comprehensive annotation across the five sections.
- Pathway Downstream Analysis: Users can input genes of interest and select from ten pathway databases to perform pathway enrichment analysis, identifying affected signaling pathways and regulatory networks.
- Genome Browser Integration: VARAdb links to the UCSC genome browser, allowing users to visualize variation data within the genomic context and explore regulatory information through custom tracks.
Datasets
VARAdb integrates data from a wide array of public resources, including:
- Genetic Variation Data: dbSNP (release 151), 1000 Genomes Project (phase 3), GWAS Catalog, GWASdb v2.0, GAD, Johnson and O'Donnell, and GRASP v2.0.
- eQTL Data: GTEx v7, PancanQTL, and HaploReg v4.1.
- Regulatory Element Data: ENCODE, Roadmap Epigenomics, UCSC, SEdb, HACER, FANTOM5, DiseaseEnhancer, EnDisease 2.0, VISTA Enhancer Browser, ENdb, TRInc, Remap, Cistrome, ChIP-Atlas, GTRD, and TF-Class.
- Chromatin Interaction Data: 4DGenome, ENCODE, and OncoBase.
- Other Data: TCGA, ICGC, PharmGKB, DisGeNET, and various ChIP-seq datasets.
VARAdb does not introduce new benchmark datasets but rather consolidates and annotates existing ones in a novel way.
Key Insights and Ideas
- Non-coding variants are functionally rich: The paper strongly emphasizes that non-coding variants are not "junk DNA" but play critical roles in gene regulation and are significantly enriched in regulatory elements.
- Integrated annotation is crucial: The sheer volume and diversity of genomic and epigenomic data necessitate a comprehensive, integrated approach to understand variant function. VARAdb's five-section structure highlights this.
- Prioritization is key: With millions of variants, a scoring system like VARAdb's is essential for researchers to efficiently identify the most promising candidates for further investigation.
- Connecting variations to pathways: The pathway downstream analysis tool offers a powerful way to understand the broader biological context and potential disease mechanisms associated with specific variants.
Key Results
- VARAdb catalogs 577,283,813 variations, including 10,386,595 coding region variations and 566,895,064 non-coding variations.
- It provides detailed annotation across five sections, covering a wide range of genetic and epigenomic features.
- The database offers five query methods for efficient searching and user-friendly browsing capabilities.
- Two online analysis tools are available: "Novel variant annotation" and "Pathway downstream analysis," demonstrating practical utility.
- Case studies illustrate how VARAdb can successfully predict the functional impact of known disease-associated variants, such as rs2279590 (Alzheimer's disease) and rs6983267 (colorectal cancer), by linking them to regulatory elements, transcription factors, and relevant genes.
Limitations and Future Work
- Data Accumulation: The authors acknowledge that annotation data sources are rapidly accumulating and plan to continue collecting and processing new data to enrich regulatory information.
- Scoring Strategy Refinement: Future work includes developing a new scoring strategy that more heavily considers TF binding sites, eQTLs, enhancers, super enhancers, ATAC accessible regions, and Hi-C data, potentially with tissue/cell-specific features.
- New Analysis Methods: The database aims to incorporate additional analysis methods, such as upstream analysis.
- Modular Updates: The database is designed with a modular structure to facilitate efficient updates and data management.
The paper highlights that VARAdb is a continuously evolving resource, with plans to expand its annotation capabilities and analytical tools to better serve the research community in understanding the complex roles of genetic variations.
Summary updated