What Is The Best Way To Calculate The Distance Between Two Proteins?

Question

[Sirus]

Hello, My question may be is simple but I am confused, I have a protein-protein interaction network and I want to express the distance between the proteins, you think which is the best way to express the distance or at least what are the usually used way in these circumstances. and does a package or a website that offer this functionality exist? Thank you in advance.

Answer 1

[Neilfws]

By distance, I assume that you mean "shortest path".

There are a lot of software solutions to this problem. For any kind of network analysis, I recommend the igraph package. It can be used from R, Python, Ruby or C. As an example, here is how you would read a network into R:

library(igraph)
# assume network is a file in GML format
g <- read.graph("myfile.gml", format = "gml")

Then consult the R documentation for functions related to "path". For example:

shortest.paths(g)

will return a matrix of N x N (where N is the number of vertices), containing the path lengths.

For a more visual approach, either Cytoscape or Gephi may be useful. In Gephi, for example, you can import a network file, select the "shortest path" tool from the menu and then click on 2 vertices to see the path length.

[Sirus]

Good answer, well I have seen that there is two kind of distances, topological and biological distance, for the topological one I think this one may suite, but is there any metric based on esome biological information that can be used, if there are some pre-existing tools, it would be fantastic :)

[Neilfws]

Well, "biological distance" is a rather nebulous term that could mean many things. I guess you are thinking about clustering the network into "functional modules". I'd suggest starting with a web or PubMed search for precisely those terms: "protein-protein interaction cluster functional modules".

[Sirus]

Yeah, in fact I am trying to implement a clustering algorithm for the protein-protein interaction network :). Thank you for your help, really appreciate it

Answer 2

[Rm]

• http://www.geneinfinity.org/sp/sp_proteininteraction.html
• http://arxiv.org/pdf/q-bio/0405028
• Deciphering Protein–Protein Interactions. Part II. Computational Methods to Predict Protein and Domain Interaction Partners
• Similarity of phylogenetic trees as indicator of protein–protein interaction
• Phylogenetic tree information aids supervised learning for predicting protein-protein interaction based on distance matrices.

[Sirus]

Thank you, those papers seems interesting I will take a look, the links are also good

Answer 3

[jackuser1979]

You can use cytoscape PPI network analysis

http://www.cytoscape.org/

Here is short tutorial on Protein interaction network analysis using Cytoscape

http://www.doc.ic.ac.uk/~natasha/course/docs/cytoscapeTutorial.pdf

For implementing cluster algorithm, you can install clusterMaker plugin in Cytoscape

http://www.cgl.ucsf.edu/cytoscape/cluster/clusterMaker.html