vardictcpp: A C++ reimplementation of VarDict variant caller

Genomic variant calling is a critical computational procedure in disease research and diagnostics, with somatic mutations like JAK2 V617F acting as diagnostic markers in myeloid neoplasms. VarDict and its Java-based reimplementation, VarDictJava, are commonly used in the bioinformatics community, though the Perl-based VarDict struggles with runtime and memory requirements. VarDictJava significantly improves execution time but demands considerable memory, limiting parallel job execution on a single machine.

Introducing vardictcpp, a functionally identical C++17 reimplementation of VarDictJava developed by researchers at the Medical-Historical Institute of Hematology at the University of Cologne. vardictcpp demonstrates remarkable performance improvements, executing approximately four times faster in single-core execution and up to nine to eleven times more efficiently when utilizing eight cores. Importantly, vardictcpp's memory consumption is up to 17 times lower than that of VarDictJava.

The variant calling results generated by vardictcpp are identical to those produced by VarDictJava when tested on whole-exome samples. This compatibility ensures seamless integration into existing bioinformatics workflows. The significant reduction in memory requirements not only allows for a higher number of concurrent jobs on existing hardware but also enhances sample throughput, which is particularly advantageous for clinical sequencing laboratories operating within the constraints of fixed compute resources.

vardictcpp, written in C++17, is freely accessible on GitHub at https://github.com/MHH-Bioinformatics-Hematology/vardictcpp, offering researchers and practitioners a powerful and efficient tool for genomic variant calling.