A Comparison of Variant Calling Pipelines Using Genome in a Bottle as a Reference

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-throughput sequencing, particularly exome sequencing, is a widely adopted diagnostic method.
• Assessing the performance of various bioinformatics tools for exome data analysis is essential for accurate variant detection.

Purpose of the Study:

• To validate an exome analysis pipeline using NIST Genome in a Bottle (GIAB) data.
• To identify the optimal combination of aligners and variant callers for human exome data analysis.

Main Methods:

• Utilized NIST GIAB results for validation of exome analysis pipelines.
• Evaluated 30 distinct pipelines, comprising six aligners and five variant callers.
• Assessed performance based on sensitivity and Positive Predictive Value (PPV) for single nucleotide variants (SNVs) and insertions/deletions (indels).

Main Results:

• The pipeline combining Novoalign (aligner) and GATK UnifiedGenotyper (variant caller) demonstrated the highest sensitivity for SNVs with a low false positive rate.
• Indel detection proved challenging across all tested pipelines, with average sensitivities below 33% and PPVs below 53%.
• The choice of aligner significantly impacts variant detection accuracy, comparable to the role of variant callers.

Conclusions:

• Novoalign and GATK UnifiedGenotyper offer a robust solution for SNV detection in exome sequencing data.
• Further development is needed to improve the accuracy of indel variant calling in exome analysis.
• Both aligner and variant caller selection are critical components for optimizing exome data analysis pipelines.