Overview
Our laboratory developed a unique method for finding sets of adequate solutions to extremely complex combinatorial problems, called the "Iterative Stochastic Elimination" (ISE) algorithm, which was applied to issues of size 10100 and more. It requires an input of variables, variable values, and a scoring method to evaluate any combination of variables and values. It had been applied to many problems associated with drug discovery and is involved purely for drug discovery in recent years in a plethora of projects. In addition to ISE, a "ligand-based" approach used to discover new and effective drugs, we also use "structure-based "methods such as docking and pharmacophore. In drug discovery, we focus on multitargeting single molecules and use databases of chemicals to screen by both "ligand-based" and "structure-based" studies.
The Iterative Stochastic Elimination algorithm
ISE
ISE is a generic algorithm capable of solving extremely complex combinatorial problems, such as finding the best solutions to a problem that has an enormous number of possible solutions that are not amenable to full examination by any means. The algorithm examines many possibilities and rejects possibilities in several "rounds" until the number of combinations is small enough to be fully computed. The significant advantage of this computational tool is its ability to suggest in silico (computerized) good solutions in a short time, which would be impossible to perform in the lab. It has already shortened the time for discovering new candidate drugs from years to months and even weeks for drug research.
ISE algorithm won the first prize in the Emerging Technologies Symposium of the "Computers in Chemistry Division, The American Chemical Society", Washington D.C. in August 2000, and the Kaye Innovation Award.
The range of problems to which ISE may be applied is enormous. It has been used to construct two start-up companies, one, Sensotrade, dealing with investments in the stock market. The other, Pepticom, predicts novel peptides based on known protein structures.
Read More ...
Platforms/software
We also use:
