Broken silence: 22,841 predicted deleterious synonymous variants identified in the human exome through computational...

Area of Science:

• Genomics
• Bioinformatics
• Human Genetics

Background:

• Synonymous single nucleotide variants (sSNVs) were historically considered neutral as they do not alter protein sequences.
• Emerging evidence highlights the role of sSNVs in various human diseases.
• Current variant prioritization strategies often overlook the potential impact of sSNVs.

Purpose of the Study:

• To identify deleterious synonymous variants within human exomes.
• To develop a robust framework for evaluating the deleteriousness of sSNVs for improved variant prioritization in genetic studies.

Main Methods:

• Analyzed 125,748 human exomes to identify sSNVs using two in silico prediction tools (SilVA and CADD).
• Employed Heterogeneous Ensemble Feature Selection to investigate key features influencing deleterious predictions.
• Assessed the frequency and gene location of identified detrimental sSNVs and investigated population-specific selection patterns.

Main Results:

• Identified 22,841 deleterious sSNVs, representing 1.8% of all synonymous variants.
• Impact on amino acid sequence and evolutionary conservation were the most significant predictors of deleteriousness.
• Discovered 39 detrimental sSNVs in disease-associated genes and 10 sSNVs under positive selection in specific populations.

Conclusions:

• A small but significant proportion of sSNVs are predicted to be deleterious and contribute to disease.
• The proposed framework enhances the evaluation of sSNVs, moving beyond the 'silent' classification.
• Findings underscore the importance of considering sSNVs in genetic research and clinical variant interpretation.