search

girke-lab / ChemmineR-git-svn-bridge

This is defunct, now replaced with https://github.com/girke-lab/ChemmineR
2 stars 6 forks source link

check db size used for computing the E-value in the fpSim function #33

Open khoran opened 9 years ago

beyondpie commented 9 years ago

Hi, khoran. I just find this function, and it seems to be an interesting one to connect the p-value with chemical similarity, which is even good for us to predict the protein targets (similar with the SEA published in Nature 2009).
So I wonder is this bug important if I use it on around ten thousand compounds directly ? The compounds are not stored in a SQL-way. Thanks. Songpeng

khoran commented 9 years ago

Songpeng, I have an easy fix for this bug, I'm just waiting for the current release cycle of bioconductor to finish, which should be on April 17. Then I will push the fix to both the release version and the development version. The fpSim function does not require anything to be in an SQL database and should handle 10,000 compounds fine. The similarity function it uses is coded in C so its fairly fast.

Kevin

On 04/14/2015 02:38 AM, Songpeng Zu wrote:

Hi, khoran. I just find this function, and it seems to be an interesting one to connect the p-value with chemical similarity, which is even good for us to predict the protein targets (similar with the SEA published in Nature 2009).

So I wonder is this bug important if I use it on around ten thousand compounds directly ? The compounds are not stored in a SQL-way. Thanks. Songpeng

— Reply to this email directly or view it on GitHub https://github.com/girke-lab/ChemmineR/issues/33#issuecomment-92720347.

beyondpie commented 9 years ago

Hi kevin, Thanks a lot ! BTW, the ChemmineR in GitHub is the latest, right ? I mean, you've fixed the bug in the github version. Sorry for that if I miss your point. Songpeng

khoran commented 9 years ago

Songpeng, No, I've not fixed it in github because github syncs automatically with the bioconductor build system and they have stopped doing builds by now in preparation for the release. The bug only affects the scale of the resulting evalue.

Kevin On 04/14/2015 07:19 AM, Songpeng Zu wrote:

Hi kevin, Thanks a lot ! BTW, the ChemmineR in GitHub is the latest, right ? I mean, you've fixed the bug in the github version. Sorry for that if I miss your point. Songpeng

— Reply to this email directly or view it on GitHub https://github.com/girke-lab/ChemmineR/issues/33#issuecomment-92868849.

beyondpie commented 9 years ago

Kevin, I got it~ Thanks ! Songpeng