Towards improved reconstruction of ancestral gene order in angiosperm phylogeny

Whole genome doubling (WGD), a frequent occurrence during the evolution of the angiosperms, complicates ancestral gene order reconstruction due to the multiplicity of solutions to the genome halving process. Using the genome of a related species (the outgroup) to guide the halving of a WGD descendant attenuates this problem. We investigate a battery of techniques for further improvement, including an unbiased version of the guided genome halving algorithm, reference to two related genomes instead of only one to guide the reconstruction, use of draft genome sequences in contig form only, incorporation of incomplete sets of homology correspondences among the genomes, and addition of large numbers of "singleton" correspondences. We make use of genomic distance, breakpoint reuse rate, dispersion of sets of alternate solutions, and other means to evaluate these techniques, and employ the papaya (Carica papaya) and grapevine (Vitis vinifera) genomes to reconstruct the pre-WGD ancestor of poplar (Populus trichocarpa), as well as an early rosid ancestor. A significant result is that the papaya genome has rearranged at a greater rate from the rosid ancestor than phylogenetic relationships would predict.