This PepSite web server can be accessed programmatically via a simple RESTful interface.

The match method

Currently, the only implemented method is match, which queries an input peptide sequence against a protein surface identified via a PDB/chain code.

It is also possible to specify the receptor protein using a UniProt ID or accession instead of a PDB code and chain identifier. PepSite will attempt to use the corresponding PDB structure with the best resolution. For more detail, see the help page.

parameter	value	example	default
pdb	4-letter PDB code	1eak	none
chain	PDB chain id	W	none
uniprotid	UniProt ID	P53_HUMAN	none
uniprotid	UniProt accession	P04637	none
ligand	query peptide sequence	GPAGPPGA	none
format	output format (see below)	-	html

Supported output formats

format	decription
html	web page with matches, scores, and molecular visualization
txt	plain text output with list of matches and scores
pdb	match coordinates in pdb format
rpdb	coordinates of interacting receptor residues in pdb format
best_pval	p-value of the best match in plain text

Examples

The following examples should be run from a Unix terminal after setting a variable with the base URL:

url='http://pepsite2.russelllab.org/'

Predict binding of peptide sequence GPAGPPGA to protein from PDB 1eak, chain D, with returned as a plain text file:

curl -s "$url/match?pdb=1eak&chain=D&ligand=GPAGPPGA&format=txt" > result.txt

Retrieve only the p-value of the best match for the above example:

curl -s "$url/match?pdb=1eak&chain=D&ligand=GPAGPPGA&format=best_pval" > result.txt

Retrieve a PDB file with the top matches for the above example:

wget "$url/match?pdb=1eak&chain=D&ligand=GPAGPPGA&format=pdb" -O matches.pdb