Retrieving interactions given compound name

[lorenzo.f...@gmail.com]

I'm trying to retrieve as much information as possible about the targets (genes, pathways, ...) of some small molecules and drugs. Is it possible to get this kind of information with pypath given the compounds' names?

[turei...@gmail.com]

Hi Lorenzo,

At the moment pypath is not able to provide drug-target interactions. The old ChEMBL support was based on MySQL and we removed this because noone wants to install a 30G database on their computers. We are working on a client for the web service and maybe connect also other databases like DGIdb or PubChem, etc. But these are not priority at the moment so will be ready in weeks or months later from now.

I would suggest to get your target protein list by another application or database and query pypath/OmniPath by this list of proteins.

Best wishes,

Denes

[lorenzo.f...@gmail.com]

Thank you for replying.

Suppose I have for each compound a list of targets, like UniProt IDs, would it be possible now to query bypath with such list?

[turei...@gmail.com]

Hi Lorenzo,

One way is to use the web service:

Let's say your proteins are P00533,O75385,P42345

Then you can get interactions by

http://omnipathdb.org/interactions?partners=P00533,O75385,P42345&genesymbols=1&fields=sources,references,type,tfregulons_level&datasets=omnipath,kinaseextra,ligrecextra,pathwayextra&types=PPI,TF&source_target=OR

Of course you can remove datasets and options according to your needs.

You can also query annotations, e.g. pathway memberships:

http://omnipathdb.org/annotations?proteins=EGFR,ULK1,MTOR&databases=SignaLink3,Signor,NetPath,KEGG

Other way is to use pypath:

import itertools

from pypath import main, data_formats

pa = main.PyPath()

pa.init_network(data_formats.pathway_all)

proteins = ['P00533', 'O75385', 'P42345']

neighbors = list(itertools.chain(*(pa.neighbors(uniprot).ids() for uniprot in proteins)))

neighborhood_graph = pa.graph.induced_subgraph(neighbors)

# annotations:

from pypath import annot

annot.init_db()

annot.db.search(protein = 'P00533')

I hope this helps.

Best wishes,

Denes