Early drug discovery is becoming increasingly sophisticated, but one fundamental challenge remains: choosing the right biology to pursue from the outset.
Speaking as part of our Thought Leadership Series, Luke Alderwick, Associate Science Director within IQVIA Discovery Sciences, discusses how better target selection, more predictive models, AI and machine learning, and closer scientific collaboration could help reduce attrition and accelerate the development of new therapeutics.
Getting the biology right from the beginning
IQVIA Discovery Sciences supports programmes across the discovery continuum, from target identification and validation through screening, hit identification and lead optimisation, as well as predictive toxicology, efficacy models and downstream candidate selection.
For Alderwick, one of the biggest scientific challenges remains selecting the right target upfront, particularly when working in complex, multifactorial diseases.
While traditional reductionist approaches have helped identify targets that can be readily prosecuted in drug discovery programmes, developability does not necessarily mean biological relevance. Establishing a clearer connection between a target and the underlying causal mechanisms of disease is therefore becoming increasingly important.
This is also where AI and machine learning could have significant impact. When combined with the expertise of discovery scientists, computational approaches can help researchers interrogate complex biological datasets, identify meaningful relationships and prioritise targets with greater confidence.
Moving towards more predictive discovery
More predictive experimental models are also changing early-stage decision-making.
Rather than relying solely on biochemical or simplified cellular assays, drug discovery teams are increasingly looking towards patient-derived samples, organoids and other physiologically relevant models.
These approaches can provide a better indication of efficacy, potential toxicity and off-target effects much earlier in development. Combined with AI and ML, Alderwick believes this could enable teams to identify problems sooner, improve candidate developability and ultimately reduce pipeline attrition.