EVOLUTIONARY TREE RECONSTRUCTION AND TRAVELING SALESMAN PROBLEM: A POWERFUL ALGORITHM FOR SHAGGY TREES

Summary Motivation: The correct Evolutionary Tree Reconstruction (ETR) based on the current genetic data is considered as an NP-complete problem. The number of “possible” trees rapidly grows with the number of “leaves”. The full ETR model includes the number of ancestor vertices and the number of mutation events on the origin tree. Efficient algorithms are needed to meet the challenges of current molecular evolution studies in order to allow simultaneous treatment of hundreds and thousands of individual genotypes. Results: We present a new ETR approach based both on the reduction the ETR problem to Traveling Salesman Problem (TSP) and on the minimization of the ETR model using Guided Evolution Strategy algorithm. The robustness of the model is defined by simulation experiments. The duration time on an ordinary computer, Pentium-4, is a few seconds for several hundreds of leaves. Availability: http://study.haifa.ac.il/~aburd/genetic.html