I have a list of 190 human proteins. I need interaction among all of these proteins not other proteins. Is there any database where I can give all of these proteins as an input and get an interaction as an output? or is there any tools? Thank you
See http://string-db.org/ and select the tab named "multiple names".
There are plenty of tools - and sources of interaction data. Here's how I'd do it in R using data from the Human Protein Reference Database.
First, make a text file with your 190 proteins, 1 per line. I'll assume that they have HGNC symbols, e.g. GRB7 and the file is called myfile.txt, with the column header "hgnc".
Next, visit the Human Protein Reference Database download page, complete their tedious registration form (every time!) and grab the file called HPRD_Release9_041310.tar.gz (or whatever the most recent version is called). Extract the zipped tar file, which creates a directory in which you'll find the file BINARY_PROTEIN_PROTEIN_INTERACTIONS.txt.
Finally, read both files into R and select only those rows from the interaction data where both partners are in your list:
hprd <- read.table("BINARY_PROTEIN_PROTEIN_INTERACTIONS.txt",
sep = "\t", header = F, stringsAsFactors=F)
colnames(hprd) <- c("hgnc.1", "hprd.1", "refseq.1", "hgnc.2",
"hprd.2", "refseq.2", "expt", "pmid")
myprot <- read.table("myfile.txt", header = T, stringsAsFactors = F)
hprd.190 <- subset(hprd, hgnc.1 %in% myprot$hgnc & hgnc.2 %in% myprot$hgnc)
The MINT database provides a "connection tool" : http://mint.bio.uniroma2.it/mint/search
Just paste the Uniprot acs in the second form and make sure you checked the "Only consider proteins in this list" option. If you're working with human proteins, you will have the option to switch to HomoMINT, where additional informations are transferred from model organisms.
I use DAPPLE, a tool available from the Broad Institute. From the web page: DAPPLE stands for Disease Association Protein-Protein Link Evaluator. DAPPLE looks for significant physical connectivity among proteins encoded for by genes in loci associated to disease according to protein-protein interactions reported in the literature. The hypothesis behind DAPPLE is causal genetic variation affects a limited set of underlying mechanisms that are detectable by protein-protein interactions.[?]
DAPPLE will build direct and indirect interaction networks from proteins encoded for by seed genes. It will then assess the statistical significance of a number of network connectivity parameters as well as of the connectivity of individual proteins to other seed proteins using a within-degree node-label permutation method. The individual protein scores are then used to propose candidate genes in large loci, and the user can decide whether to iterate until it can no longer propose any more new genes. It will then email the user a number of output files including figures, data summaries and resultant significance scores.
I find the second paragraph of importance - the permutations and statistical evaluations of the built network.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Once you have uploaded your list go to the dropdown menu "Inter-relations". Select either predicted or experimental interactions. Once you see the list of interactions in your screen, click on the drop down menu "Export". You will get the data in the sif format that you can directly use with Cytoscape.
It is really useful database/tool. Thanks a lot. But how can I get the network?