OMG! Orthologs for Multiple Genomes -Competing Formulations (Keynote Talk)

摘要

Multiple alignment of the gene orders in sequenced genomes is an important problem in comparative genomics [1]. A key aspect is the construction of disjoint orthology sets of genes, in which each element is orthologous to all other genes (on different genomes) in the same set. Approaches differ as to the nature and timing and relative importance of sequence alignment, synteny block construction, and paralogy resolution in constructing these sets. We argue that these considerations are best integrated in the construction of pairwise synteny blocks as a first step, followed by the conflation of the pairwise orthologies into larger sets. The two advantages of this are: first, the availability of finely tuned pairwise synteny block software (e.g., SynMap in the CoGe platform [2,3]) and second, the opportunity to dispense with parameters, thresholds or other arbitrary settings during the construction of the orthology sets themselves. The orthology sets problem becomes a pure graph algorithm problem.