Drug Design Using Molelcular Operating Environment.

The explosion of information has propelled the rapid development of bioinformatics. The future of postgenomic biology requires something that does not even exist today. There are numerous algorithms and new methodologies that need to be developed consistently. Bioinformatics is reckoned to revolutionize the disease and therapies research.

One has to consistently looking at accelerating the drug discovery process for genomics, proteomics and in silico biology. Simulating the in vivo to in silico is a challenging task today. But today it is possible to achieve with the advancement in High performance computing power and with the software developed my efficient algorithms. There are a few software available for in silico drug design. One such product is MOE, Molecular Operating Environment developed by chemical computing group Canada. Which contains all the programs from sequence analysis protein modeling and structure based drug design.

During this talk I would like to give a demo on how this software is helpful in the following areas.

MOLECULAR MODELING AND SIMULATIONS

Molecular Builders. Build small molecules, carbohydrates, proteins, and DNA and crystal structures in 3D with a collection of graphical interfaces.
Molecular Mechanics and Dynamics. Perform large-scale energy minimizations using AMBER '89/'94, MMFF94, PEFSAC95 or Engh-Huber force field parameters with an implicit solvent model.
Conformational analyses include both stochastic and systematic searches (including rings). Calculate dynamics trajectories in the NVE, NVT, NPT and NPH ensembles.
Implicit Solvent Electrostatics. Solve the non-linear Poisson-Boltzmann equation with a non-linear multi-grid method to produce the implicit solvent electrostatics field.
Molecule Alignment. Align a collection (flexibly or rigid-body) of small molecules that are presumed to have similar biological activity. Alignments are useful for the elucidation of pharmacophores, input to comparative field analysis and template forcing.

HIGH THROUGHPUT DRUG DISCOVERY

QSAR. Create large-scale linear or probabilistic predictive models of biological activity (or other properties). Use CCG's patent-pending Binary-QSAR method for HTS data analysis and ADME-based classification.
Combinatorial Library Design. Design combinatorial libraries using either maximum diversity or QSAR-biased methodologies.
Molecular Descriptors. Calculate over 480 molecular properties including topological indices, octanol/water logP, molar refractivity and CCG's VSA descriptors that have shown wide applicability in compound classification, QSAR and ADME property modeling.
Compound Clustering. A powerful non-parametric subdivision of probability space technique is used to rapidly cluster large databases. Compounds can be clustered by fingerprints (3D Pharmacophore, graph pharmacophore or MACCS keys) using the Jarvis-Patrick method.
Diverse Subset. Calculate maximal diverse subsets of compounds based on 3D conformations, molecular descriptors or molecular fingerprints

PROTEIN MODELING AND BIOINFORMATICS

Protein Structure Database. CCG has created a searchable database of over 15,000 protein structures from the Protein Data Bank. Each structure was imported using CCG's PDB reader, that corrects many of the errors commonly found in protein structure files.
Fold Identification. CCG has created a searchable database of structural families of proteins by exhaustively and iteratively clustering the Protein Data Bank. Incorporating rigorous sequence-to-family alignment, secondary structure prediction, hydrophobic fitness and Z-scoring, the search procedure
Multiple Alignments. CCG's unique technology for simultaneous multiple sequence and structure alignment superposition with no "master" sequence.
Consensus Features. Geometric criteria are used to rapidly determine the structurally conserved features in a family of proteins including conserved waters.
Structure Prediction. Homology modeling techniques are used to build complete high-quality 3D structures of proteins from a template.
Protein Mechanics and Dynamics. Proteins structures can be refined using either AMBER '89, '94, MMFF94 or Engh-Huber parameters augmented with an implicit solvent model. Molecular dynamics can be performed in either the NVE, NVT, NPT
or NPH ensembles.
Stereo chemical Quality. Statistical measures of bond lengths, angles, backbone dihedrals and non-bonded contacts are used to assess the overall stereo chemical quality of a protein structure.
Contact Analysis. Stabilizing contacts such as hydrogen bonds, salt bridges, hydrophobic contacts and disulfide bonds are often implicated in protein function.

STRUCTURE BASED DRUG DESIGN

Structure Preparation. Automatically connect and assign type information using element and coordinate information. Add hydrogens and refine/relax structures using AMBER '89, '94, MMFF94 or Engh-Huber parameters augmented with an implicit solvent model.
Visualization. Receptors and ligands can be displayed in a variety of styles including line, stick and CPK with full control over colors. Connolly, Accessible and van der Waals interaction surfaces can be colored in a number of ways including by Pocket and Hydrogen Bond. Both split-pair and full stereo modes are supported.
Active Site Detection. An alpha shape algorithm is used to determine potential active sites in 3D protein structures.
Ligand Docking. A simulated annealing search algorithm is used to flexibly dock a ligand into the active site of a receptor in an effort to predict the binding conformation. A grid-based energy evaluation is used to score the docked conformations.
MultiFragment Search helps to understand the interactions between chemical functional groups with the active site of a receptor. An ensemble of fragments are randomly placed into the active site and subjected to a special energy minimization protocol to determine the preferred locations of each functional group.

Text of lecture delivered at a workshop conducted by the FBAE.

Dr R Raghu, Tata Elxsi, Bangalore 560 025, rraghu@tataelxsi.co.in