Which Genetics Variants in DNase-Seq Footprints Are More Likely to Alter Binding?

Area of Science:

• Genomics
• Regulatory Genomics
• Bioinformatics

Background:

• Characterizing the human regulatory genome is ongoing.
• A systematic definition of functional vs. silent non-coding genetic variants is lacking.
• Understanding variant impact is crucial for interpreting genetic association studies.

Purpose of the Study:

• To systematically predict the impact of genetic variants on transcription factor (TF) binding.
• To differentiate functional from silent genetic variants in non-coding regulatory regions.
• To improve the fine-mapping of causal variants in genome-wide association studies (GWAS).

Main Methods:

• Integrated DNaseI footprinting data with sequence-based TF motif models.
• Predicted variant effects on TF binding across 153 tissues and 1,372 TF motifs.
• Utilized allele-specific hypersensitivity (ASH) and GWAS data for validation.

Main Results:

• Identified 5.8 million genetic variants within DNaseI footprints.
• Predicted 66% of these variants affect TF binding.
• Found only 3,217 single nucleotide polymorphisms (SNPs) show evidence for functional impact (ASH), suggesting 97% are silent.
• Annotation improved GWAS fine-mapping by at least 2-fold for 86 SNPs.
• Tissue-specificity and affected TF binding site identity aided mechanistic interpretation.

Conclusions:

• The majority of genetic variants in regulatory regions are functionally silent.
• The developed computational approach accurately predicts variant impact on TF binding.
• This annotation enhances the interpretation of GWAS results by identifying potential causal variants and their mechanisms.