Machine Learning to Unravel Anti-Ageing Compounds

The ageing process is one of the greatest scientific mysteries of our time. Ageing of the population is also a major biomedical, social and economic challenge. In the past two decades, a number of drugs have been identified that can modulate ageing and preserve health in model organisms, from invertebrates like worms to mammals such as mice. In spite of this recent progress, understanding the best strategies to develop human anti-ageing interventions and preserve health in old age is still very incomplete. Besides, although a few anti-ageing drugs are being clinically tested as treatments for age-related diseases, including drugs initially associated with ageing in worms, new methods are necessary to identify and prioritize drugs that can be suitable for human applications. Indeed, identifying new longevity drugs is of widespread interest.

In this project, we will develop new computational methods that take advantage of large amounts of multi-omics data available on the web, and our own compilations of the effects of hundreds of ageing-related drugs, to predict new drugs with anti-ageing properties. Specifically, methods will be developed that determine which attributes make some drugs extend animal lifespan. Furthermore, once we determine which attributes these drugs have in common, we can predict novel pro-longevity, health-promoting drugs. Worms will be employed for the experimental validation of the machine learning findings in this project because of their short lifespans and high fecundity.

All methods developed will be made available to the scientific community to help guide experiments, including in other organisms. Moreover, we will develop a webserver for predicting life-extending compounds that will be made freely available online. Overall, this project will provide a significant impetus to employing predictive biology in ageing studies.

Ageing can be manipulated in model systems by hundreds of genes and compounds. Nonetheless, ageing is still a poorly understood process, and identifying the most important modulators of ageing remains a challenge. Given the intrinsic costs of performing animal ageing studies, developing predictive computational tools is of utmost importance, and the accuracy and specificity of current predictive computational methods is still very limited.

In this project we will employ machine learning to provide insights into life-extending drugs and predict those with the greatest potential for human translation. Specifically, we aim to: 1) Apply machine learning methods to predict new life-extending effects of compounds in animal models; 2) Test if the main compounds with new life-extending effects predicted in Aim #1 extend worms' lifespan and healthspan; 3) Predict and prioritize compounds with potential for retarding mammalian ageing using machine learning.

The methods developed could be employed in other organisms and will be made available to the research community. Moreover, we will develop a webserver for predicting life-extending compounds that will be made freely available online. Therefore, this project will establish novel computational approaches suitable to guide hypothesis-driven studies on ageing that will be applicable to other systems.

Our proposed project will open up new directions of research, and potentially accelerate the development of medicines for ageing diseases and open opportunities for future translational research.