Please login to the form below

Not currently logged in
Email:
Password:

Computer model successfully predicts drug toxicities

Could be used to filter out harmful drugs before preclinical trials

A new set of computer models developed at the University of California San Francisco (UCSF) has successfully predicted the side effect profiles of more than 600 current drug molecules in around 50 per cent of cases.

With additional refinement, the computational strategy could be used as a screen to filter out potentially toxic molecules in drug discovery programmes and focus efforts on compounds with a greater chance of success, according to the researchers.

The approach could also help drugmakers save the billions of dollars each year that are wasted on developing drugs that ultimately fail because adverse reactions are detected in clinical trials. Only one in 5,000 drug candidates that enter preclinical testing ever reaches the market.

While similar predictive models have been developed in the past, the new system is a step forward because it can handle hundreds of compounds at once according to the team at the UCSF, the Novartis Institutes for BioMedical Research (NIBR) and UCSF spin-out SeaChange Pharmaceuticals.

The system works by examining the structural similarity between test compounds and those known to be associated with a battery of 73 unintended side-effect target proteins that appear on Novartis' safety panel for testing drugs for side effects such as heart attacks.

"Approximately half of the predictions were confirmed, either from proprietary databases unknown to the method or by new experimental assays," according to the authors of the study, which is published online in Nature (June 10).

"This basically gives you a computerised safety panel, so someday, when you're deciding among hundreds of thousands of compounds to pursue, you could run a computer programme to prioritise for those that may be safest," said Michael Keiser, co-first author of the paper and a co-founder of SeaChange Pharma along with UCSF's Brian Shoichet and John Irwin.

The computer model identified 1,241 possible side-effect targets for the 656 drugs, of which 348 were confirmed by Novartis' proprietary database of drug interactions.

Another 151 hits revealed potential side effects that had never been identified for these drugs, which Novartis confirmed through lab testing.

13th June 2012

Share

Featured jobs

Subscribe to our email news alerts

PMHub

Add my company
EY Life Sciences

Our global reach Against a backdrop of regulatory, digital, financial and global transformation, Life Sciences companies are finding integration and...

Latest intelligence

ema1
The European Medicines Agency: PRIME’d for access?
Leela Barham examines the impact of the EMA's PRIME fast track system after two years...
How can pharma engage with key stakeholders on NHS service transformation?
Steve How, Paul Midgley and Oli Hudson, of the Wilmington Healthcare consulting team, explain how pharma should make its case for change...
michael elliot
The race for an HIV ‘cure’
Supercharging therapies as pharma and patients work together...

Infographics