On 08/29/2017 09:16 PM,
t.cher...@gmail.com wrote:
> On Tuesday, August 29, 2017 at 12:25:45 AM UTC+4:30, Tayebe Cheraghi wrote:
>> hi dear
>> which databases are important for protein-protein interaction network in cytoscape?
>>
>> i want to import network from databases but there are different sources (databases) and i don't know which of them is useful for protein interaction network.
>>
>> when i use all active databases, a complex and indistinct network constructed.
>>
>> thanks
>
>
> thanks for your replay.
>
> I have a gene list.
>
> I tested STRING. in STRING they have no direct interaction but in cytoscape when I select "universal interaction databases client" and select all active databases, they have many interactions and vast network.
>
> which of them is correct?
Well, STRING has a confidence cutoff, so depending on your confidence
cutoff you could certainly get different results. Which is correct?
You would need to look at the evidence for the interaction that each one
is proposing. For the ones that have no interaction, is there any
evidence in the literature that they actually interact?
>
> If i restrict databases up to 1 or 2 it would become plain and simplex network.
> but which of these databses can i select (BioGrid, ChEMBL, IntAct, APID interactome, MINT, Reactome, InnateDB, EBI-GOA nonIntAct and etc...)??
You would need to look at the description of each of the databases.
BioGrid is a protein-protein interaction database, ChEMBL is a database
of small molecules, Reactome is a database of (mostly) human pathways,
etc., etc. There is no *right* database. Most of them have strengths
and weaknesses that really depend on the biological question you are asking.
-- scooter
>