Indira Ghosh

 Title: Virtual High Throughput Screening: will it help for lead identification?

 Bioinformatics Center, University of Pune, Pune-41107 

Abstract : 

Improving the process in the early drug discovery phases remains one of the difficult challenges in the scientific community. In the past decade High Throughput Screening (HTS) technologies have leveraged robotics and miniaturized assays to automate the identification of proprietary lead molecules and subsequently validate their effects on drug targets. Although HTS is a widely applicable technology, the challenge is to further leverage the initial lead molecules by using them as a template molecule so that a number of proprietary molecules with similar activity but better drug profile can be developed. Only limitation is that this process is highly costly and chemical intensive. Also the attrition rates are very high. Hence an alternative in silico virtual High Throughput Screening(vHTS) method has become very popular to filter out unpromising set of compounds at the very beginning.

In a vHTS, each chemical compound available in the library is sequentially docked at every possible pose at the active site of the target molecule (often protein). Docked molecules with the lowest free energy of interactions are ranked against all docked molecules in the list and the best ranked could be considered as potential 'hits'. These sets of compounds can be subjected to experimental assays to determine if they are appropriate to select as lead for the optimization studies.

Virtual Screening programs consist of two essential parts: an algorithm that searches the conformational, rotational and translational space available to a candidate molecule within the binding site, and an objective function, representing a crude measure of binding affinity (called scoring) to be minimized during this process. In order to be successful as a virtual screening tool, the program must be able to find solutions (called poses) for active molecules in accordance with experiment and it should be able to separate active compounds from inactive ones. Several review articles have been published addressing the challenges associated with both the issues. Computational requirements for vHTS often are a limiting factor for productive screening campaigns. Advances in CPU speed, increasing parallelism, and distributed computing (Grid computing) promises to reduce the run time.

An attempt has been made to study the effect of the active site of receptor on the vHTS. In the present study, two-model receptors have been chosen: (i) Plasmepsin-II - large active site, presence of acidic residues and charged residues in the active site.  (ii) Farnesyl transferase (FTase)- large active site, presence of basic residue, charged residues and Zinc metal in the active site.

Plasmepsin-II vHTS has been performed using two systems: (i) Grid-computing enabled platform, using LigandFit docking module of Cerius2. (ii) Heterogeneous linux/irix cluster environment, using FRED docking tools. The dataset was 1.4 million small organic compounds. FTase vHTS has been performed using FRED  (www.eyesopen.com) in a heterogeneous linux/irix cluster environment. The dataset was 3.5 million small organic compounds, obtained after applying customized filter (to remove compounds that non-drug and non-lead like). The protonation states of the residues in the active site of proteins have been assigned. Tautomers and conformers generation, along with charge assignments and refinement (MMFF94 force field) of the compounds have been performed using OMEGA .   

The virtual screening of the both the targets with 4.6 million (3.2 for FTase and 1.4 Plasmepsin-II) compounds has been done within a week time, so as to promise an initial set of leads for experimentalist.

Analysis of the screening data to derive the probable hit list will be discussed in relation to the of the parameterization of scoring system based on  re-docking of the conformers of experimentally bound ligand and enrichment factor on the test set of 1000 compounds, comprised of known experimental inhibitors along with diverse random decoy chemicals. A novel method of ranking compounds based on soft fuzzy criteria developed in-house will also be discussed. 

'.$link.'';}?>