Targeted Genomic Region Masking Supports Accurate Variant Calling While Suppressing Low-Complexity Sequencing...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• False-positive variant calls in low-complexity regions (LCRs) hinder clinical genomics.
• Short-read sequencing data often contains deterministic artifacts complicating variant analysis.

Purpose of the Study:

• Evaluate a targeted spatial masking strategy to suppress artifacts in LCRs.
• Preserve clinically actionable variants located outside LCRs in short-read sequencing data.

Main Methods:

• Implemented a selective masking protocol before variant calling.
• Applied the protocol to reference standards (EQA, NA12878) and breast cancer cohorts (n=25).
• Evaluated callsets for sensitivity, precision, mutational signatures, VAF, noise, and database annotations.

Main Results:

• Thousands of sequencing and alignment artifacts were removed, preserving biological variants.
• LCR masking maintained physiological mutation profiles and resolved pseudo-multiallelic noise.
• Excluded artifact calls showed distorted mutational and variant allele frequency (VAF) signatures, unlike retained calls.

Conclusions:

• Targeted spatial filtration is a critical early pipeline intervention for high-fidelity clinical genomics.
• Database annotation alone may be insufficient for variant authentication, showing cohort-dependent behavior.
• This strategy improves the analysis of non-LCR-associated germline variants using short reads.