Our Sponsors



Download BioinformaticsOnline(BOL) Apps in your chrome browser.




Question: Question: Features to select non-redundant protein-ligand dataset

Muhammad Ali
3841 days ago

Question: Features to select non-redundant protein-ligand dataset

Hi,

I have a huge protein-ligand dataset which is also redundant, I want to remove that redundancy using CD-HIT program, I want to know different features on the base of which I can filter(narrow down) my data (for example just take those ligands which are drug like molecule, neglect all others) any suggestion in this context will be highly appreciative.

Thanks in advance.

 
Answers
3

Hi Muhammad,
You can try using lipinski rules also known as rule of five (http://preadmet.bmdrc.org/?option=com_content&view=article&id=70:lipinskis-rule&catid=36:druglikeness&Itemid=87)
These rule classify the ligands which have potential to become drugs.(tool: http://biochemlabsolutions.com/Molecule%20Docking/DRUG-LIKELINESS/Drug-like.html)  But remember exceptions are always there. There are many marketed drugs which are exception to rule of 5.
Depending upon your ultimate aim you have to decide this on your own.


Also for clustering the ligands on basis of similarity see this discussion (http://www.biostars.org/p/17766/)
one of the tool for the same i found was Ligand.info (http://ligand.info/)

PS: I have never used these tools so you have to see they are free or not. And u can find another tools also.

Let me know if u need any more info/help.

0

Hi Ankita,

Thank you very much for your valuable suggestions, the first tool you mentioned (filtering drug like molecules: http://biochemlabsolutions.com/Molecule%20Docking/DRUG-LIKELINESS/Drug-like.html) it's release is pending, I am exactly searching for a tool like that but I am unable to find it. I am still searching for it, if you can provide me some information regarding such a tool it will be a great help. Thanks again.