Well, I just gave my talk on phylogenomic and functional predictions and am going to try and catch up with blogging.
In my talk I discussed how an understanding of function and prediction of function requires an understanding of how functions have evolved.
I am trying to get my talk slides posted here but, alas, I need to deal with some Copyright issues first (OK – here is a little slideshow of my talk … no audio sorry)
Patricia Babbitt gave a talk after mine on another aspect of phylogenomics and functional predictions. She has done some really interesting studies (see her lab site here) of functional diversifications and the molecular level by integrating genomic, structural, biochemical and phylogenetic analyses. She showed some really nice tools for clustering and visualization protein families that, although not phylogenetic, seemed to be very useful for the onslaught of genome data. Unfortunately, most of her publications are not in OA journals so I cannot use any of the figures here and am not going to bother linking to the papers.
Kimmen Sjolander is talking now about her phylogenomic studies. She is discussion “Phylofacts” which are precomputer phylogenetic trees of gene families from across the tree of life. One great thing about her work – most of it is published in OA journals and most of her software is available for free download.
More detail on Phylofacts is available in her Genome Biology paper here. Kimmen has done some really great work on automating phylogenomic functional predictions and this is one example. Also see her Flower Power and Sci-Phy and Satchmo and Intrepid and Phylo builder other tools (downlaods and other information are available at her website here.)
Gretta Serres will be speaking next.
I will post on here talk soon … And now I am posting on it …
She is discussing “linking metabolic diversity to protein families” os something to that effect. She discussed something I never have thought of doing which is the following – take genomes, identify the size of protein families in each genome and then cluster genomes by their similarity of the protein family size. I assume some others must have done this but it seems like a good way to identify similar duplication pressures on distantly related organisms.