Main

In this repo, we present code for producing results in the article: "Efficient algorithms for pangenome personalization".

C++ source code

Prerequisites

To compile the source code, the following tools should be installed:

1. KMC API headers, which can be downloaded with ./download_include.sh. Run the script from the root of the repo after cloning.
2. OpenMP library for C++ multithreading
3. OR-TOOLS library for min-cost flow.

OR-TOOLS can be downloaded from their offical website.

You can use cmake to compile code with rules from CMakeLists.txt. In that case you should add argument -DCMAKE_PREFIX_PATH="<path/to/or-tools/dir>" to cmake before running.

Programs

There are three executables, that can be compiled.

1. gfa_scorer is an executable that for a given graph in GFA, and $k$-mers database outputs scores of vertices in pgf format on standart output.
2. paths_finder is an executable that for a given graph in PGF runs 2-paths algorithm, and returns two paths in specificed format.
3. table_producer is an executable that for given set of GFAs and walks returns tables with stats that are further used in section "Recovery of true vertices".

Important

To run paths_finder properly, header of graph in PGF should be changed, which can be done by exps/scripts/change_header.py:

python change_header.py graph.gfa score_graph.pgf > scored_graph_with_header.pgf