Comparison of genetic variants in matched samples using thesaurus annotation

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate detection of DNA changes, such as somatic mutations in cancer, necessitates analyzing multiple matched samples.
• Low-mappability genomic regions present challenges for conventional variant calling methods, leading to under-reporting of critical events.
• Thesaurus annotation offers a novel approach to accurately describe variants by linking multiple genomic loci.

Purpose of the Study:

• To introduce and evaluate software for thesaurus annotation to detect DNA point changes in matched samples.
• To benchmark the performance of thesaurus annotation against conventional methods for mutation calling.
• To demonstrate the applicability of thesaurus annotation for analyzing de novo mutations in family studies.

Main Methods:

• Development of Java-based software (GeneticThesaurus) for thesaurus annotation.
• Implementation of mutation calling within an R package (RGeneticThesaurus).
• Benchmarking on matched normal/tumor samples to assess performance metrics like true event recovery and false discovery rate.

Main Results:

• Thesaurus annotation recovered 5–10% more true DNA events compared to conventional approaches in matched normal/tumor samples.
• The technique demonstrated a strict limitation on false discovery rates.
• The method proved fully consistent with established variant analysis workflows and applicable to de novo mutation analysis.

Conclusions:

• Thesaurus annotation is a robust technique for enhancing mutation calling accuracy, particularly in difficult genomic regions.
• The developed software provides a valuable tool for researchers studying genetic variations in cancer and hereditary diseases.
• This approach improves the comprehensive analysis of genomic variants from matched samples.