Specialization scoring: detecting specialization between protein paralogues.
     
For the impatient: click here to go straight to the results page.
     
Let's look at lactalbumin and locate the regions of specialization since it diverged from lysozyme C, some 300 to 400 million years ago. Download the example directory here (right click, save and unzip).
     
The directory contains two files with extension *.fasta, and one PDB (structural file). Fasta files contain a collection of eutherian orthologues for lysozyme-C and α-lactalbumin, obtained from OrthoDB database. 2fyd.pdb contains the structure of α-lactalbumin in complex with 1,4-galactosyltransferase, the enzyme it regulates.
     
You can upload both fasta files to Cube at once, by pressing ctrl while making the selection. If you upload the pdb file, the pymol session will be produced. You can use it to inspect the interface between the two proteins. Note that this interaction is specific for lactalbumin. Compare the levels of conservation across both lactalbumin and lysozyme families, with the specialization within lacatalbumin group of orthologues. It is a cautionary tale about using conservation as a guide when the collection of sequences is a mixture of different paralogues - proteins similar in sequence, but with diverged function.
     
Not happy with the alignment produced by the server? Read here about re-aligning sequences by yourself.