Phylogenetic Analysis Of Gene Enrichment?

Question

[Biostar User]

Are there any tools to do phylogenetic analysis of gene enrichment? This is, I have a list of genes from an experiment performed in several species, with a z-score that can be described as "enrichment" associated to each gene in each species. It's not gene expression, it's enrichment of a certain marker in the gene. I would like to compare these lists, that are not fully overlapping, across the known phylogeny of the species (i.e., a known species tree with estimated branch lengths according to genetic distance), to see significant changes in the trend of enrichment at the gene level, or at the branch level. What tools could I use for this?

Species1 Z-score
Gene1    0.532
Gene2    0.531
...

Species2 Z-score
Gene1    0.51
Gene3    0.505
...

Answer 1

[Leonor Palmeira]

Assuming you have a "true" species tree, and assuming you have a table with gene in columns and species in lines, and that you want to determine:

• if and how the tree structure accounts for the variance per gene

then you have a simple phylogenetic principal component analysis problem, and you can use the adephylo and ade packages in R to help you solve this question. ppca() is basically the function you are going to work with.

You could also reverse the problem: have different gene trees for each gene, a table with species in columns and genes in lines, and determine how the gene tree structure accounts for the variance per species. This would probably answer the second version of your question: are there significant changes in the trend of enrichment at the gene level?

[Biostar User]

Yes, I would assume a known species tree. I then want to look at significant changes at the gene level or at the branch level.

Answer 2

[Vitis]

I would approach this as a classic phylogenetic character mapping problem. The gene function enrichment is a special kind of character. This has been conducted in very large scale as in this paper: http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1002411

[Biostar User]

This is an interesting suggestion although right now I can't wrap my head around the formatting of the input for this approach. In essence I know the species tree, and I want to use it to fit the changes I see in the list of genes and look at significant outliers.