LATERAL GENE TRANSFER DETECTION USING MULTIPLE AGREEMENT FORESTS

Abstract

Phylogenetic trees are used to illustrate evolutionary relationships between and among species. However, lateral gene transfers (LGTs) can cause different evolutionary histories for genes compared to the species. One method to identify possible LGT scenarios uses mathematical models called maximum agreement forests (MAFs). Previous MAF-based models require vast sequence data and cannot identify specific transfers, such as antibiotic resistance origins. This study extends single MAF analysis to identify particular LGT using several MAFs, focusing on transfers found in all MAFs of two phylogenetic trees, called obligate transfers. We present a method for enumerating all MAFs using modified branching rules and cluster reduction along with a method for identifying obligate transfers without enumerating all MAFs. Our findings through experiments suggest listing all MAFs is feasible for identifying obligate transfers. Furthermore, we propose methods for tracing the LGT endpoints for non-binary reference trees to improve running time by performing non-binary LGT analysis.