I need to build up a virtual library of the ligands for my in silico drug design. So I need guidance on how to build up the ligand library, the software to use and how to make sure the database does not have redundant ligands.
Building of ligand library is very crucial in terms of drug designing. Now to start you will have to set a reference molecule which is already commercialized drug and till date it is the best, now the aim is to find a entity which is more effective than this one which is already commercialized. Based on the reference one, you select basic scaffold or the or the part which is responsible for bringing an biological effect , and search in the public repositories such as Pubchem or zinc database, etc to find the strcutural derivatives of the compound the selection criteria should be lipinski rule of 5, and so you can build up a compound library further you will have to filter out your library depending on molecular descriptors and many conditions to find one best effective compound.
So if I already have the reference, then I already download the related structure from PubChem, Zinc and so on, how we want to know the databases that we downloaded not redundant?
when you prepare your ligand database try to address the following issues:
- 3D conformations (if required)
- tautomer representation
- formal charges representation
- enantiomeric expansion
The more you can improve the representation of your molecules the more accurate results you will get.
Based on what you are planning to do (e.g. docking rather than pharmacophore or similarity etc) have a look if the software you are planning to use to analyse your database offers some tools to build the database itself. Most commercially available packages have specific modules to do so (e.g. LigPrep in Schrodinger as an example) that will also address the issues I mentioned at the beginning.
Hello there. There is actually a non-redundant database of ligands that is available from the NCI that is ready to use. It is a the NCI Diversity Set. You can download them at https://wiki.nci.nih.gov/display/NCIDTPdata/Compound+Sets. You might also want to refer to a previous and related post in ResearchGate for relevant information: https://www.researchgate.net/post/I_predicted_a_highly_druggable_pocket_in_a_protein_How_could_I_prepare_a_library_for_virtual_screening?_tpcectx=profile_answers
In addition to the previous answers, for the practical aspects of managing and manipulating large compound libraries in various formats I would highly recommend OpenBabel software which is freely available from openbabel.org
Thanks for the answer. After we build the library, do we need to validate the library? Another things, which one is better, either build library based on reference molecule before screen or just screen all the molecules in available library?
Library building is part of designing process. After you build library you will screen the molecules accordingly. I would suggest to build library on the basis of reference molecule with a basic scaffold. Hope this helps
My purpose to build up the library is to find the suitable drug that can inhibit the targets. After that, optimize the molecule before synthesize it. I thought the library need to validate whether it reliable or not. Is it?
ChemAxon Markush enumeration is an important source to generate scaffold based limited library. Beside this Chem T , autoclickchem and Similib like soft tools can be used to organised the library.