# From Best Matches to Gene Families: How to use paralogs in phylogenomics

**Ponente**: Peter F. Stadler

**Institución**: University of Leipzig

**Tipo de Evento**: Investigación, Divulgación

Cuándo |
06/08/2019 de 17:00 a 18:00 |
---|---|

Dónde | Sala A2 del Centro Académico Cultural (CAC) |

Agregar evento al calendario |
vCal iCal |

Best match graphs (BMGs) arise naturally as the first processing intermediate in

algorithms for orthology detection. Let T be a phylogenetic (gene) tree T and

sigma an assignment of leaves of T to species. The best match graph (G,sigma)

is a digraph that contains an arc from x to y if the genes x and y reside in

different species and y is one of possibly many (evolutionary) closest

relatives of x compared to all other genes contained in the species sigma(y). I

will give two alternative characterizations of BMGs and show that a minimally

resolved tree that explains a BMG can be reconstructed in cubic time. The

symmetric part of a BMGs represents the empirical estimate for the orthology

relation on the gene set as inferred from a reciprocal best match heuristic.

BMGs are therefore close relatives of co-graphs, which describe perfect

duplication/speciation scenarios. Whenever a BMG deviates from a cograph

structure, this implies that the reciprocal best match heuristic has produced

incorrect orthology assignments. A reasonable approach therefore it to correct

the data by editing the BMG into its nearest co-graph. Cographs, in turn, are

equivalent to event-labeled gene trees that identify duplication and speciation

events. These trees also impose constraints on the species tree and the

possible reconciliation maps. Taken together, therefore, it is possible to

start from reciprocal best matches of the proteoms of a set of species and

eventually arrive at the phylogenetic tree of these taxa without the use of a

conventional tree reconstruction method. In fact, an analysis of the workflow

show that it only makes use of gene duplication events, while sets of 1-1

orthologs do not contribute at all. In this sense the approach is

orthogonal to classical phylogenetic methods.