Many (if not most) clinical trials end up failing, but given the bias towards positive results sometimes the reasons for their not working remain unclear. Luckily, a team at Open Targets and EMBL’s European Bioinformatics Institute (EMBL-EBI) have analysed the reasons for aborted clinical trials mentioned in Clinicaltrials.gov, the open-source register for clinical trials worldwide, using machine-learning.

In their recently published paper, they analysed the free-text reasons explaining why 28,842 clinical trials –about half of them in cancer– stopped. Then, they evaluated target-disease relationships from a genetics point of view. Their analysis revealed that clinical trials terminated for reasons such as safety or efficacy concerns had less supporting genetic evidence for the link between drug candidate and the disease.

If you want to know more, go check the full article at MappingIgnorance.org