Calculate SMD and Hamming and Jaccard distances between each pair of samples in a set of variant files.

• libbio (provided as a Git submodule)

On Linux also the following libraries are required:

• libdispatch (provided as a Git submodule)
• libpthread_workqueue (provided as a Git submodule)
• libkqueue

• Reasonably new compilers for C and C++, e.g. GCC 7. C++17 support is required.
• GNU gengetopt (tested with version 2.22.6)
• Ragel State Machine Compiler (tested with version 6.7)
• CMake
• Boost

1. git clone https://github.com/tsnorri/vcfdistances.git
2. cd vcfdistances
3. git submodule update --init --recursive
4. Edit local.mk
5. make -j4

1. Clone the repository with git clone https://github.com/tsnorri/vcfdistances.git.
2. Change the working directory with cd vcfdistances.
3. Run git submodule update --init --recursive. This clones the missing submodules and updates their working tree.
4. Edit local.mk in the repository root to override build variables. Useful variables include CC, CXX, RAGEL and GENGETOPT for C and C++ compilers, gengetopt and Ragel respectively. BOOST_INCLUDE is used as preprocessor flags when Boost is required. BOOST_LIBS and LIBDISPATCH_LIBS are passed to the linker. See common.mk for additional variables.
5. Run make with a suitable numer of parallel jobs, e.g. make -j4

Useful make targets include:

all

Build everything

clean

Remove build products except for dependencies (in the lib folder).

clean-all

Remove all build products.

The tool takes one or more Variant Call Format files as its input. It reads the variants and makes pairwise comparisons between each sample. Remaining files are processed similarly. The requested distances are output as triangular matrices.

Please see src/vcfdistances --help for command line options.