Skip to main navigationSkip to main content
The University of Southampton
Southampton Statistical Sciences Research Institute

Research project: Prediction of thermophysical properties of pharmeceutical compounds

Currently Active: 

Joint research with Chemistry is developing and applying methodology for continuous variable selection in data sets with a large number of potential explanatory variables.

The work is in collaboration with Pfizer UK through an EPSRC CASE award, and is motivated by problems in Crystallography. The scientific aim is to build predictive models for measured properties (which require experimentation) in terms of chemical properties and descriptors that can be calculated in-silico. The research has concentrated on the prediction of the thermophysical behaviour such as melting point of drug-like compounds. The main statistical issues concern the appropriate choice of variable selection techniques and statistical models, and the choice of compounds to synthesise in future experiments. The model building and variable selection problems are made more challenging by the sizes of available datasets and feasible experiments (small to medium throughput) which result in there often being more potential descriptors than data points.

Share this research project Share this on Facebook Share this on Twitter Share this on Weibo
Privacy Settings