Bayesian gene/species tree reconciliation and orthology analysis using MCMC

• DOI: 10.1093/bioinformatics/btg1000
• Publication Date: 2003-07-10T23:49:03Z
• Authors: Lars Arvestad; Ann-Charlotte Ber...; Jens Lagergren; Bengt Sennblad
• Keywords: Tree (set theory)

Abstract Motivation: Comparative genomics in general and orthology analysis particular are becoming increasingly important parts of gene function prediction. Previously, reconciliation has been performed only with respect to the parsimony model. This discards many plausible solutions sometimes precludes finding correct one. In other areas bioinformatics probabilistic models have proven be both more realistic powerful than models. For instance, they allow for assessing solution reliability consideration alternative a uniform way. There is also an added benefit making model assumptions explicit therefore comparisons possible. analysis, uncertainty recently addressed using parsimonious combined bootstrap techniques. However, until now no methods available. Results: We introduce evolution based on birth-death process which tree evolves ‘inside’ species tree. Based this model, we develop tool capacity perform practical Fitch’s original definition, generally reconciling pairs trees. Our biologically sound (Nei et al., 1997) intuitively attractive. Bayesian MCMC facilitates approximation posteriori distribution reconciliations. That is, can find most probable reconciliations estimate probability any reconciliation, given observed gives way that pair genes orthologs. The main algorithmic contribution presented here consists algorithm computing likelihood reconciliation. To best our knowledge, first successful introduction type methods, flourish phylogeny into analysis. implemented and, although not yet being its final form, tests show it performs very well synthetic as biological data. Using standard correspondences, results carry over allele trees biogeography. Contact: {lottab,jensl}@nada.kth.se, {bengt.sennblad,lars.arvestad}@sbc.su.se *To whom correspondence should addressed.