Similarities and differences between variants called with human reference genome HG19 or HG38

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Reference genome selection is critical for accurate next-generation sequencing (NGS) data analysis.
• Current studies often utilize either the HG19 or HG38 human reference genome versions.
• The impact of reference genome choice on single nucleotide variant (SNV) identification remains under-assessed.

Purpose of the Study:

• To rigorously assess the impact of human reference genome versions (HG19 vs. HG38) on SNV identification.
• To compare SNV calling and conversion rates between HG19 and HG38 using whole genome sequencing data.
• To provide recommendations for optimal reference genome selection in NGS studies.

Main Methods:

• Utilized whole genome sequencing (WGS) data from the Genome-In-A-Bottle (GIAB) project.
• Performed SNV calling using 26 distinct bioinformatics pipelines on both HG19 and HG38.
• Employed conversion tools to assess SNV compatibility between HG19 and HG38, analyzing conversion and discordance rates.

Main Results:

• Conversion rates from HG19 to HG38 averaged 99%, while HG38 to HG19 averaged 95%.
• Approximately 1.5% of SNVs showed discordant conversions between the two reference genomes.
• SNVs failing HG38 to HG19 conversion were often low confidence, had low read depth, or were GC-rich alleles.

Conclusions:

• A substantial number of SNVs are not directly convertible between HG19 and HG38.
• The HG38 reference genome is recommended for NGS SNV analysis due to superior performance.
• Caution is advised when translating SNVs identified using different human reference genome versions.