Christopher Reynolds

E-mail	cr308@imperial.ac.uk
Telephone	+44 (0)20 7594 5737
Mailing Address ✉	Structural Bioinformatics Group Centre for Bioinformatics Division of Molecular Biosciences Faculty of Life Sciences Imperial College London SW7 2AZ UK
Supervisor	Professor Mike Sternberg

Inductive Logic-Based Drug Design

Since October 2008, I have been undertaking PhD research into the subject of Inductive Logic-Based Drug Design.

Introduction

New drug leads are always needed, but the problem of finding them is the vast size of chemical space to search through. The number of synthetically feasible, drug-like molecules is estimated to be in excess of 10⁶⁰ (Bohacek, et al., 1996). Instead of physically testing compounds using high-throughput screening, virtual screening is often used to simulate the bioactivity of compounds, in order to search as much of chemical space as possible in a reasonable amount of time. Virtual screening using machine learning has made many recent advances. Databases of drug-like molecules are available, and virtual screening methods are available to scan them for potential drug leads, but the chemical space of available molecules is just a small part of the chemical space of all synthetically accessible molecules.

Aim

A means of using machine learning to search for molecules that are easily synthetically accessible is needed. This project will take an existing logic based machine-learning approach to identifying bioactive compounds, then further validate this methodology, and incorporate rules of chemical synthesis to design new, original, and more effective pharmaceutical drugs.

Objectives

To achieve the aim, the overall objectives will be to:

Produce a tool that can contribute to future drug design.
Test the success and viability of this approach for automatically designing effective drugs by cross-validation, and by withholding a validation subset of the data from the training set and testing if this data can be retrieved from a screening of the entire mol ecule database with the validation subset concealed within it.
Identify at least one small molecule with improved activity over exist ing drugs with the same target.
Submit some of the molecules produced for pharmaceutical testing.
Disseminate the results of the project through conferences and scientific publications.

Presentation

On 7th November 2011, I gave a talk at the 7th German Conference of Chemoinformatics on my work. The talk can be downloaded here.

Bohacek, R.S., McMartin, C. and Guida, W.C. (1996) The art and practice of structure-based drug design: A molecular modeling perspective, Medicinal Research Reviews, 16, 3-50.