CWI and VU develop algorithms for pharmaceutical research

In their search for new drugs, researchers increasingly use computer simulations to discover and test new molecules for pharmaceutical use. This does not only increase the chance of finding new or better drugs, but also prevents wasting resources on unsuccessful experiments.

Publication date: 26-10-2015

In their search for new drugs, researchers increasingly use computer simulations to discover and test new molecules for pharmaceutical use. This does not only increase the chance of finding new or better drugs, but also prevents wasting resources on unsuccessful experiments. PhD student Mohammed El-Kebir of Centrum Wiskunde & Informatica (CWI) and VU University Amsterdam has developed new algorithms for faster and more effective biological computer simulations. He will defend his thesis on Tuesday 27 October, at the VU University Amsterdam.

New techniques has made a wide range of biological data available, providing detailed information on the structure and interaction of genes, proteins and biomolecules. The main challenge for biologists is to make sense of this data and turn it into useful insights. Bioinformatician Mohammed El-Kebir presents in his thesis several algorithms that can improve pharmaceutical research for discovering new drugs.

One of these algorithms allows pharmaceutical researchers to discover new drugs more quickly by speeding up simulations predicting the effect of drugs on the human body. These simulations currently take a lot of time, as drug molecules often consist of a large number of atoms. To be able to perform the simulations efficiently, chemists divide every molecule in parts based on chemical properties. This is very time-intensive manual labour that takes several days per molecule. El-Kebir succeeded in capturing this chemical intuition in an algorithm that performs the same task in a split second. This large time gain tremendously improves the viability of computer simulations in drug research.

Another algorithm predicts the tranferability of experimental results between mice and humans. This increases the success rate of drug trials, and allows for fewer, more goal-directed experiments, thus saving lab animals. In a drug trial, pharmaceutical researchers investigate the effect of potential new drugs by administering these to model organisms, such as mice. If an effect is found, the drug is tested on a small set of humans. Only 25% of the experiments survive this second phase, mostly because the results found in the lab animals turn out to be non-transferable to humans. The algorithm developed by El-Kebir identifies patterns in large networks of genes to discover cellular mechanisms that are conserved throughout the course of evolution. If such a mechanism is conserved between mice and man, it will therefore probably be affected by drugs in the same way.

This research is funded through the Netherlands Organisation for Scientific Research (NWO).