Computer-based methods can accurately predict drug side-effects
SCIENTISTS and pharma companies are on a constant lookout for new and better drugs. They work equally hard to ensure these new drugs remain free of side-effects. These unwanted secondary effects are a huge health concern and the main reason medicines are taken off market. Researchers from Kyushu University in Japan have now proposed a new in silico or computer-based method to predict side-effects of drugs.
Their study, published in the Journal of Chemical Information and Modeling on December 21, shows in silico methods can predict side-effects of a drug in the pre-clinical stages and can, therefore, play a critical role in drug discovery. These predictions need to be validated experimentally, but despite this, these early leads are crucial.
“Methods, such as this, give you a computerised safety panel, which will help to screen hundreds and thousands of compounds in drug development and subsequently help prioritise those that may be the safest,” says P K Diwan, president of the Indian Pharmacological Society, Hyderabad, a national body promoting need-based research in pharmacology. Traditionally drugs are tested for side-effects in vitro, using cell lines, lab animals and even human volunteers. Apart from being time consuming, these experimental methods also mean a huge loss of money, time and effort, if test results are negative. In silico methods work a step before the experimental work begins. They predict how a drug will behave when put inside the human body using the already available chemical data. The novelty of this method lies in integrating information about a candidate drug molecule’s chemical structure and the target protein, unlike the predictive methods currently in use.
“This type of data integration is essential for accurate prediction of side effects. For any given drug we do not know all of the targets. If we did we may be able to anticipate most side effects. Integrating data from chemical structure can improve prediction by sharing knowledge across drugs,” says Nicholas Tatonetti, assistant professor at the department of biomedical informatics at Columbia University in the US.
Lead researcher Yoshihiro Yamanishi and his colleagues tested their method on 658 drugs for simultaneous prediction of 969 side-effects and found it to be more accurate than other methods based on chemical or biological information. For further proof, they compared the information provided by the in silico method on side-effects of certain drugs, like betamethasone and lovonorgestrel, with independently existing literature on these drugs. They found the method predicted the side-effects in accordance with existing literature.
Apart from known drugs, a comprehensive side effect prediction for 730 uncharacterised drug molecules listed in drug banks was also done. Predicted profiles were found to match with predictions made by independent sources.
Some researchers have, however, questioned the efficacy of this validation procedure. “As for betamethasone a few important side-effects, like hyperglycemia and Cushing’s Syndrome, were not found in the list. Similarly, the method could predict some side effects for lovonorgestrel but some other important ones are missing,” says Ramasamy Raveendran of the department of pharmacology at Jawahar Lal Institute of Post Graduate Medical Education and Research in Puducherry.
We are a voice to you; you have been a support to us. Together we build journalism that is independent, credible and fearless. You can further help us by making a donation. This will mean a lot for our ability to bring you news, perspectives and analysis from the ground so that we can make change together.
.png)
Comments are moderated and will be published only after the site moderator’s approval. Please use a genuine email ID and provide your name. Selected comments may also be used in the ‘Letters’ section of the Down To Earth print edition.