Reproducibility of Variant Calls in Replicate Next Generation Sequencing Experiments

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Next-generation sequencing (NGS) is a powerful tool for detecting nucleotide alterations.
• However, sequencing errors can lead to false positive variant calls, impacting biological interpretation.
• Assessing the reproducibility of variant calls is crucial for reliable genomic analysis.

Purpose of the Study:

• To evaluate the reproducibility of nucleotide variant calls in replicate sequencing experiments.
• To determine the concordance rates of all variant calls and single nucleotide variants (SNVs).
• To identify factors influencing the reproducibility of SNV calls.

Main Methods:

• Targeted sequencing of the human kinome (~3.2 Mb) using the SOLiD v4 platform.
• Sequencing of 17 breast cancer samples in duplicate or triplicate.
• Analysis of concordance rates for all calls and SNVs, correlating with sequencing metrics.

Main Results:

• Overall concordance rates exceeded 99.99%, but SNV concordance ranged from 54.3% to 75.5%.
• Concordance increased with higher coverage, variant allele count (VAC), variant allele frequency (VAF), and quality metrics.
• Variant allele count (VAC) and variant allele frequency (VAF) were the most significant determinants of concordance.

Conclusions:

• Even with stringent quality control, SNV call reproducibility was around 80%, suggesting 20-40% error rates in single experiments.
• The study highlights substantial variability and potential for erroneous variant calling in NGS.
• Sequence data are available in the European Genome-phenome Archive (EGA) under accession EGAS00001000826.