Inferring duplication episodes from unrooted gene trees

Paweł Górecki,Jarosław Paszek

doi:10.1186/s12864-018-4623-z

Abstract

BackgroundOne of evolutionary molecular biology fundamental issues is to discover genomic duplication events and their correspondence to the species tree. Such events can be reconstructed by clustering single gene duplications inferred by reconciling a set of gene trees with a species tree.ResultsHere we propose the first solutions to the genomic duplication problem in which every reconciliation with the minimal number of single gene duplications is allowed and the method of clustering called minimum episodes under the assumption that input gene trees are unrooted.ConclusionsWe showed new theoretical properties of unrooted reconciliation for the duplication cost and apply them to design several exact and heuristic algorithms for solving the problem. Our evaluation study on empirical dataset confirmed several genomic duplication events from the literature and demonstrate that algorithms can be successfully applied.

Highlights

One of evolutionary molecular biology fundamental issues is to discover genomic duplication events and their correspondence to the species tree
We propose the solution to the unrooted minimum episodes problem, UME, in which allowed scenarios have the minimal number of gene duplications [36]
We show that every instance of UME can be transformed into at most 5k “simpler” instances that can be solved in linear time, where k is bounded above by special cases of S2 stars [47] in input trees

Summary

Introduction

One of evolutionary molecular biology fundamental issues is to discover genomic duplication events and their correspondence to the species tree. The whole-genome duplication occurred for numerous species and had a crucial impact on the evolution of crops [6,7,8,9]. The studies of this phenomenon focus on detecting its occurrences as well as its influence on introgressing novel metabolic traits [10] or its association with periods of increased environmental stress [11]. The methods of detecting whole-genome duplications can be divided into three categories based on synteny and colinearity comparison of genomes [1, 12, 13], the estimation of

Methods

Results

Discussion

Conclusion