Efficient sampling of transpositions and inverted transpositions for bayesian MCMC

  • Authors:

  • István Miklós;Timothy Brooks Paige;Péter Ligeti

  • Affiliations:

  • eScience Regional Knowledge Center, Eötvös Loránd University, Budapest, Hungary;Amherst College, Amherst, MA;Bioinformatics group, Alfréd Rényi Institute of Mathematics, Hungarian Academy of Sciences, Budapest, Hungary

  • Venue:
  • WABI'06 Proceedings of the 6th international conference on Algorithms in Bioinformatics
  • Year:
  • 2006

Quantified Score

Hi-index 0.02

Visualization

Abstract

The evolutionary distance between two organisms can be determined by comparing the order of appearance of orthologous genes in their genomes. Above the numerous parsimony approaches that try to obtain the shortest sequence of rearrangement operations sorting one genome into the other, Bayesian Markov chain Monte Carlo methods have been introduced a few years ago. The computational time for convergence in the Markov chain is the product of the number of needed steps in the Markov chain and the computational time needed to perform one MCMC step. Therefore faster methods for making one MCMC step can reduce the mixing time of an MCMC in terms of computer running time. We introduce two efficient algorithms for characterizing and sampling transpositions and inverted transpositions for Bayesian MCMC. The first algorithm characterizes the transpositions and inverted transpositions by the number of breakpoints the mutations change in the breakpoint graph, the second algorithm characterizes the mutations by the change in the number of cycles. Both algorithms run in O(n) time, where n is the size of the genome. This is a significant improvement compared with the so far available brute force method with O(n3) running time and memory usage.