GraphML specializations to codify ancestral recombinant graphs

摘要

Software which simulates, infers, or analyzes ancestral recombination graphs (ARGs) faces the problem of communicating them. Existing formats omit information either about the location of recombinations along the chromosome or the position of recombinations relative to the branching topology. We present a specialization of GraphML, an XML-based standard for mathematical graphs, for communication of ARGs. The GraphML <node> type is specialized to contain the node type, time, recombination location, and name. The GraphML <edge> type is specialized to contain the ancestral material passed along that edge. This approach, which we call ArgML, retains all information in the original ARG. Due to its use of established formats ArgML can be parsed, checked and displayed by existing software.