P2P ConPred

Prediction of residue contacts and residue correlations in proteins using pair-to-pair substitution matrix

P2PConPred is a program to calculate correlation scores between columns in multiple sequence alignments. The calculation is based on a weighted sum of all pair to pair (P2P) substitution scores between pairs of sequences. The scores are taken from a large P2P matrix which includes a score for each of the (400*400) possible pair to pair substitutions. The scores reflect the degree of correlation between the two sites. High correlation usually arise from a spatial contact between the residues in these two sites.

The program also calculates and reports:
1. Conservation of each site
2. Entropy of each site
3. Standard deviation of the individual scores which compose the overall score between each two positions.

At this site you can download the source code for p2p for different Unix/Linux platforms and for windows. You can also use our web server for the calculation. The results are sent back by Email and also saved on the server. Visual analyses of the results are supported.

get p2p code

To use the web server follow the instructions below. You must fill one of the options:

Submit your own sequence (the server creates alignment using Blast hits and Clustalw)
Submit your own multiple sequence alignment (FASTA format)
Specify a PFAM id
Submit a PDB code. The server create multiple alignemnt based on the structure or uses PFAM entries which contains domains of the structure

The alignment should be in Fasta format (example). A first line starting with ">" is required for each sequence, although it might contain no additional information.

Paste a sequence (fasta format)

Paste a multiple alignment (fasta format) file

Enter a multiple alignment (fasta format) file

Enter PFAM id (e.g PF00018)

Enter PDB file

Enter a pdb id

Check the next box for calculation of solvent accessibility. Considering only buried residues improves significantly the contact prediction accuracy. The prediction of accessibility values is done using SABLE and might take several minutes, usually more than the calculation of the P2P scores

Perform SABLE prediction

Matrix for calculation

Output format

Email address

The P2P calculations take about a minute for a query sequence of L=200 residues, based on a MSA of N=100 sequences, and the computation time scales as O(N²L²). For long tasks, users have the option of submitting their e-mail address below, so as to receive an e-mail informing them about the completion of the job, and pointing them to the web address where the results are stored. Otherwise, you can wait and be directed to the same address at the completion of the job. In any case the email address is not saved on the server and does not serves for any other purpose.

Send questions and comments to Eran Eyal - e y a l @ c c b b . p i t t . e d u