Chemical Forums

Specialty Chemistry Forums => Biochemistry and Chemical Biology Forum => Topic started by: mir on August 29, 2007, 08:13:29 AM

Title: A data-analyzation tool for protein similarities
Post by: mir on August 29, 2007, 08:13:29 AM
My Bioinformatic lecturer showed a slide with this picture on. The analyzation tool behind the image, connects proteins (the dots) with each other with a line if they have anything to do with each other. What Im looking for is the software behind such charts. Do anybody know what its called and how to get it?
Title: Re: A data-analyzation tool for protein similarities
Post by: Yggdrasil on August 29, 2007, 01:55:36 PM
What connections are you looking at specifically (i.e. what do the lines represent)?  Do they represent homology between two protein sequences above a certain cutoff?  Do they represent functional linkage?  Do they represent direct interactions?  Also, is the proteome from one organism or a collection of organisms.  There are many different programs that will make mappings (although I don't know many of them by name, I've just seen them in various publications).

In general, maps like these are constructed using a spring-minimization algorithm.  Basically, the strength of the linkage between two proteins determines the spring constant of the line connecting the two.  The program then generates a random arrangement of the proteins in space then lets the springs relax into the lowest "energy" state.  An example, using functional linkages between proteins can be found in figure 4 of the following publication (yay for open access articles):

http://genomebiology.com/2004/5/5/R35

The software described by the article is found here:

http://mysql5.mbi.ucla.edu/cgi-bin/functionator/pronav
Title: Re: A data-analyzation tool for protein similarities
Post by: mir on August 30, 2007, 03:35:07 AM
I found a reference hidden on the same slide: Fraser, A. G. & Marcotte, E. M. Nat Genet, 2004, 36, 559-564 .

Im not an expert in biochemistry, but I guess its the interaction between proteins. Since my lecturer mentioned that the proteins that worked together where clustered together.

I was really only interested in the software, since I found another use of it. If it works I will "publish" it on my blog, than you see... ::)

Thanx!