Hebrew University of Jerusale

Prof. Amiram Goldblum's

School of Pharmacy
Molecular Modeling and Drug Design Group



Projects in our lab are mainly of two types: methodology developments and applications. In most cases, methodology development is performed by PhD students and involves writing of computer code in one or a few of the prominent languages. MSc thesis may involve program/algorithm development or analysis as well as new algorithm development, depending on the background of students. Quite a few MSc students are developing ideas in drug design for kinases. Our laboratory currently has Two associates, 4 PhD students and 8 MSc students.

A new algorithm

Much of the current activity in my lab is to extend the scope and applications of a general optimization method that we started to develop a few years ago. It is called "Iterative Stochastic Elimination"(ISE) and allows us to find sets of best solutions to highly complex combinatorial problems. These sets (populations or ensembles) have been shown to include the global minimum for problems in which ISE solutions could be compared to exhaustive computations, with problem sizes up to 108-109. As we deal with problems of size 10200 and more, there is no proof that global minima are detected in such immense combinatorial problems, but there are a few indications that encourage us to use those best results. On several problems, we compared our results to those from other stochastic methods and found them to be better (closer to global minima) or more representative of the problem's space.

The requirements for applying ISE are:

  1. the problem may be presented as a set of variables that may assume, each, a set of values
  2. variables “interact” with each other (dependence)
  3. A reliable scoring function may be constructed to evaluate each (random) choice of variables’ values
ISE is a generic algorithm that may be applied to any complex optimization problem. In the fields of protein interactions, it is applied to construct ensemble of conformations of protein side chains and loops, of ligands, and for flexible ligand docking. For small molecules, we have used ISE for conformational analysis and for computing properties of molecular ensembles, for distinguishing between active and inactive molecules acting on specific targets (Molecular Bioactivity Index, MBI) and for distinguishing between drugs and non-drugs (Drug Like Index, DLI).

Recent and current work on ISE and its applications in my group:

  1. Flexible ligand docking (including side chain flexibility)
  2. Conformational ensembles and properties for molecules with many rotatable bonds
  3. Cheminformatics - distinguishing between active and inactive (or less active) molecules (on a specific target, on drug likeness, solubility, BBB passage, toxicity etc.)
  4.  Protein Design problems: optimizing protein-protein interactions,creating new functions for proteins
  5. Distinguishing between folding and non-folding sequences
  6. Protein – protein interactions
  7. New methods in QSAR: 3D pharmacophore optimization
  8. Metabolism by P450 isoenzymes - chenoinformatics for assigning substrates and inhibitorss
  9. Computational Biology issues

Other subjects that we consider for applications of ISE are:

  1. Multiple structure/sequence alignment methods
  2. Fold recognition

Dr. Anwar Rayan is our main developer and consultant for ISE applications.


Last Update : January 19, 2011 | Copyright ©, 2009, The Hebrew University of Jerusalem. All Rights Reserved.