New methods to measure residues coevolution in proteins

Getting bored at worked and have to stay a bit longer as my computer is creating a back up. I can not work either because for some reason BWA is not working on our cluster so what I am intending to write a blog about an article recently published in BMC Bioinformatics. This work is done in Shanghai University, mainly by Hongyun and Jun Wang.

Already getting a bit disappointed on the article as I was expecting it to detail more of sophisticated  algorithm but apparently this article is based on mutual information theory with modification on the MIB model. If you don't know then MIB stands for  Mutual information with background distribution.

In conclusion they write "our analysis also reveals that the covariation of physicochemical properties is a new aspect of coevolution information", really ???

When there is a mutation of functionally or structurally important residues, compensatory mutations may occur to preserve or restore the function or structure of the protein. - To the point.

There are two ways of methods for coevolution calculations:

1. Parametric method

• Maximum likelihood approximation
• Bayesian proababilities
• Phylogenetic approches

 2. Nonparametric model

• Mutual information
• ELSC

BLOSUM62 is used in the article as the amino acid background distribution.  Pfam was used for downloading MSAs with number of sequences for an alignment should be greater than 125. Too few or too many?? I don't know.

They have validated this model on a non-comprehensive way and protein pictures are not the best one either. Here is an example...

 




That is all, not interested to write more on this article , hopefully next time something interesting.