hictk: blazing fast toolkit to work with .hic and .cool files

We developed hictk, a toolkit that can transparently operate on .hic and .cool files with excellent performance. The toolkit is written in C++ and consists of a C++ library with Python bindings as well as CLI tools to perform common operations directly from the shell, including converting between .hic and .mcool formats. We benchmark the performance of hictk and compare it with other popular tools and libraries. We conclude that hictk significantly outperforms existing tools while providing the flexibility of natively working with both file formats without code duplication. We reanalyzed data from ENCSR410MDC (GSE63525) to generate Hi-C matrices in .hic v8, .hic v9 and .mcool format for benchmarking purposes. Matrices in .hic v8 and v9 format were generated using JuicerTools and HiCTools respectively while .mcool files were generated using cooler. In all cases, pair-wise interactions from ENCFF447ERX were used as input. BEDPE files with cis and trans queries were generated by randomly sampling TADs from ENCFF301CUL.