Solving Biological Complexity to Engineer Medicines

Drug discovery R&D is in a productivity crisis. It takes 10-15 years to bring a medicine to market, costing $2.6bn, yet just 5% make it to market. Less than 30% of drugs entering clinical trials make it to phase 3; over 40% of drugs fail phase 3, by then $1.4bn has already been invested. Reasons for late-stage failure are attributed to 50% of drugs lacking efficacy and 30% lacking safety. Attrition rates vary according to therapeutic area. Oncology medicines experience a 97% failure rate.

We don't fully understand why the majority of drugs fail or succeed. We need to understand how medicines are working. If we understand the therapeutic mechanisms more deeply and holistically, then we'll be in a better position to engineer medicines that treat illnesses effectively.

The crisis stems from the failure to understand mathematically, biological complexity. Biological theories haven't reached the level of generality and power currently evidenced in physics or chemistry. We need to develop mathematical theories of disease biology and therapeutics; however, there are significant technical challenges to developing such insights because modelling biological systems is exceptionally hard.

The behaviour of biological systems are characterised by nonlinear dynamical processes, functioning at multiple levels of organisation - cells, tissues, organs - in space and time; we don't yet have the mathematical or computational tools to describe this level of complexity into overarching models.

We're developing a technology to help us do just that.

In this project, we're developing a computational system augmented by AI to help us understand how medicines work at different levels and scales in the human body. Such a holistic understanding will give us a fuller picture of how to both understand disease biology and how best to treat disease.

Taking multiple myeloma as our case study, we aim to build quantitative systems models of the disease biology and how medicines impact the disease. Based on these mechanistic models, we use AI to make predictions about multiple myeloma therapies.

We aim to generalise the methods from this project and apply them to all medicines - to tackle diseases, which are having the greatest impact on human health.

The project seeks to disrupt how we model and understand biological complexity, by building a novel computational system augmented by AI, thereby enabling us to create a new paradigm in drug discovery and development - to engineer biology and program medicines that transform human health.