vardictcpp: a C++ reimplementation of VarDict variant caller

C++17 reimplementation of the VarDictJava variant caller, with identical variant calls

VarDict, originally written in Perl, and its reimplementation VarDictJava are widely used variant callers; the Perl version can take several hours, and VarDictJava, much faster, can consume large amounts of main memory, limiting the number of parallel jobs per machine. vardictcpp is a functionally identical reimplementation of VarDictJava in C++17. It is approximately four times faster in single-core execution and 9–11 times faster using eight cores, while requiring up to 17 times less main memory. Variant calls are identical to those of VarDictJava on the evaluated whole-exome samples, and the code is freely available on GitHub.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

No algorithmic novelty: the value is operational, for clinical sequencing laboratories with fixed compute resources already running VarDict, notably for somatic markers such as JAK2 V617F in myeloid neoplasms. Call identity is shown only on exomes whose number the abstract does not give; it should be checked on high-depth targeted panels with low allele fractions before any routine substitution.

Analysis by Dr Thibaut Benquey