Standard analyses of protein coinheritance depend upon building and clustering a coinheritance matrix, a 2D array in which the rows correspond to proteins, the columns correspond to sequenced genomes, and the entries correspond to the presence or absence of a protein in a sequenced genome (often denoted by a boolean, a BLASTP E-value, or a bitscore). However, now that we have in excess of 200 fully sequenced microbial genomes -- and hence more than 200 columns in our coinheritance matrix -- hierarchical clustering no longer allows ready human visualization. As more than 370 additional genomes are expected to be finished in the near future, the need for dimension reduction and visualization is clear. Phylogenomic mapping fills this need by reducing the coinheritance data set from N-dimensional space to 2 dimensions, enabling human visualization of the similarity relationships between coinherited proteins. As the N-dimensional space corresponds to the space of more than 200 sequenced genomes, this is a reduction in visual complexity of more than 2 orders of magnitude. A full description of the method along with extensive experimental confirmation is available in the Nature Biotechnology paper by Srinivasan et al.

Browsing the maps on your computer
To get up and running with the maps for a microbe of interest, a brief tutorial introduction is available here. The maps can be viewed with the aid of the VxInsight program. A restricted version suitable for viewing the maps associated with this paper is available here, along with the user manual.

Raw Data
The raw coinheritance matrix data used to generate the phylogenomic maps for each bacteria can be downloaded here, and the .faa files for the M. xanthus genome can be downloaded here.