Utility of Genotyping in Pharmaceutical Target (gene) Discovery and Drug Response Anne Westcott EST-Informatics

Pharmaco-genetics/genomics A recent interest from pharmaceutical companies –5yrs Application across both Development and Discovery

Why are we interested now? Recent discovery of massive no.’s of genetic markers (SNPs, MS) Better, cheaper tools for genotyping Opportunity to collect and archive DNAs from populations –Research collections or clinical trial subjects

Application to Disc and Dev Research –Target identification via susceptibility gene discovery projects –Association studies Development –“personalised medicine” –Correlating drug response with genotype

Pharmacogenetics Adverse events – deaths US 1994, 2 million individuals with serious side-effects Variability in response 30% non-responders

What is a genotype? SNP Genotype

How to correlate There are 2 commonly used statistical approaches for correlating disease with genetic makeup. Linkage where 2 things are inherited together e.g. a SNP and a gene always seen together. Linkage Disequilibrium When things don’t go as the maths or model suggests Association By linkage, by haplotype, looking for transmission disequilibrium

Problems with these studies Poorly characterised phenotypes Small sample size (not enough DNAs) Subgroup analysis Random error Poor control group No replication Inability to replicate Multiple testing Over-interpretation Failure to detect linkage disequillibrium

From Informatics Perspective Sequence-based features –HGP, sequence, order and orientation of clones, physical distance Trait-based info. –Clinical trials data –Disease status –Bone density, drug response, liver enzyme levels, bronchial responsiveness –Ethnic diversity Need breadth and depth Need depth and consistency

What do we need? DNA collections Phenotypic data about those individuals Pedigree data (where approp) Controls Characterisation of disease