search

adam-sam-brown / ksRepo

ksRepo - A Generalized Tool for Computational Drug Repositioning
BSD 3-Clause "New" or "Revised" License
3 stars 3 forks source link

Gene-chemical interaction #2

Open vasilislenis opened 6 years ago

vasilislenis commented 6 years ago

Hi Adam,

My apologies for bothering you again but I have another question about ksRepo. I was wondering if it is possible apart from the association of the drug component with the gene(s) of interest to have also the information of the gene-chemical interaction.

For example, let's say that I found that the chemical component "CAPSAICIN" is related with the BCL2L1, BIRC5, and CASP2 genes. Is it possible to have as an output (extra column, maybe) the interaction that this chemical compound has with these genes (e.g. decreases/increases their expression)?

Thank you very much in advance, Vasilis.

adam-sam-brown commented 6 years ago

Hi Vasilis,

Short answer - not really.

Longer answer - for the repo, I've pulled the Comparative Toxicogenomics Database (ctdbase.org) as the source of drug-gene interactions. Given that it's a manually curated database, there's a lot of conflicting information - e.g. genes are noted as being both up- and down-regulated by a given compound. To give ksRepo flexibility and side-step that issue, I ignored the reported direction of effect.

That said, the gene interactions database from the CTD (http://ctdbase.org/reports/CTD_chem_gene_ixns.csv.gz) does have these interactions, so you could easily map them back. You'd then need to come up with a scheme for if there are conflicts.

Any other questions, feel free to reach out, Adam

vasilislenis commented 6 years ago

Thank you very much Adam!

What can I say, I really love the manually curated databases! :)

Ok, I believe that if I’ll map the results to the db and keep the cases that I don’t have a conflict I’ll be alright.

Thank you very much, Adam, for your support!

Kind regards, Vasilis.

On 8 Oct 2018, at 16:45, Adam Brown notifications@github.com wrote:

Hi Vasilis,

Short answer - not really.

Longer answer - for the repo, I've pulled the Comparative Toxicogenomics Database (ctdbase.org) as the source of drug-gene interactions. Given that it's a manually curated database, there's a lot of conflicting information - e.g. genes are noted as being both up- and down-regulated by a given compound. To give ksRepo flexibility and side-step that issue, I ignored the reported direction of effect.

That said, the gene interactions database from the CTD (http://ctdbase.org/reports/CTD_chem_gene_ixns.csv.gz http://ctdbase.org/reports/CTD_chem_gene_ixns.csv.gz) does have these interactions, so you could easily map them back. You'd then need to come up with a scheme for if there are conflicts.

Any other questions, feel free to reach out, Adam

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/adam-sam-brown/ksRepo/issues/2#issuecomment-427884026, or mute the thread https://github.com/notifications/unsubscribe-auth/AGhlfznxAnWWfzVAgdoo5i5TIQO2Z7nBks5ui3MDgaJpZM4XNK7U.