A conversation with Professor Linda Partridge
Professor Dame Linda Partridge is one of the world’s most renowned scientists in the field of longevity and healthy aging. Professor Emerita at University College London (UCL), she was director of the Institute of Healthy Ageing at this university for many years. She was also the founder and director of the Max Planck Institute for Biology of Ageing in Cologne, Germany.
Her pioneering research has revealed that it is possible to extend healthy life through genetic and nutritional interventions, inspiring new pharmacological approaches to prevent diseases such as Alzheimer’s and Parkinson’s.
Throughout her professional career, she has received numerous awards and honorary degrees. A member of the Royal Society and the Academy of Medical Sciences, she received the title of Dame Commander of the Order of the British Empire (DBE) for her contributions to science.
Professor Linda is considered one of the most influential voices in the pursuit of longevity with health and purpose.
AI has been used to generate various types of ageing clocks. Could you explain us what are ageing clocks and how IA can be useful in the context of clinical medicine?
An aging clock can be constructed from any trait that changes during ageing. The original clocks were made using changes in a type of molecule that sticks to the genetic material, DNA. These are called methylation clocks. In the population as a whole, they tick at a steady rate during ageing – there are predictable changes in the methyl groups on the DNA. But some individuals experience these changes faster than average – their methylation clock is ticking faster and their methylation age is greater than their chronological age. Others have clocks that tick slower than average, and their methylation age is younger than their chronological age. Importantly, faster ticking of these clocks can predict future ill health and time to death, not perfectly, but there is a definite predictive value. A faster ticking clock says that the person has a higher biological age than their chronological age.
Now we also have clocks that use RNA, proteins, metabolites, and combinations thereof, and we have many ways of using the data to make the clocks, including ones based on AI. Recent work suggests that using AI can make the clocks more accurate in terms of predicting future health and lifespan. Models that can include non-linear changes with age are particularly useful.
Even more important, very recently it has been shown that clocks based on changes in proteins can detect the health of individual organs – your heart, lungs, kidney, brain. So the biological age of one organ can be greater than that of another within the same person.
Ageing clocks could help spot early signs of declining health, enabling preventative strategies and interventions before disease onset. This is particularly true of organ-specific clocks because they could be very useful for early detection of risk of specific diseases. They may also allow people to proactively track their health, make better lifestyle choices, and take steps to stay healthy for longer.
Some drugs are proving to be geroprotective – they combat mechanisms of ageing and prevent age-related diseases, as least in animals. Can AI help identify such drugs.
AI is proving very useful generally for identifying potential drugs targets and then for predicting molecules that will interact with those targets. This is particularly used in industry, but increasingly also in academia. AI can also predict whether existing drugs could be repurposed to delay or prevent age-related diseases, and to identify ones that would be better tolerated by users. AI systems gather and integrate diverse datasets, including existing drug information, genomic data, proteomics, clinical trial results, and aging-related molecular markers.
AI enables structure-based drug design, predicting protein structures. AI models can perform virtual screens of libraries of chemical compounds to identify potential lead candidates that have a higher probability of success in later development stages.
Generative AI models can even create novel molecular structures from scratch, designing compounds with specific characteristics to enhance drug efficacy and reduce toxicity. It can even design drugs for targets that were previously considered undruggable. AI can also improve other essential properties, including Machine learning (ML) algorithms predict the physicochemical properties of drug candidates, including solubility, bioavailability, and toxicity, helping to optimize the drug development process.
AI can also optimise clinical trials, by improving patient recruitment, designing optimal trial protocols, and even generating synthetic control arms using real-world data, reducing the time and cost of drug development and lowering the risk of failure in trials.
By analysing individual genetic profiles, lifestyle factors, and other data, AI can help design personalized treatments tailored to a patient’s specific needs, leading to better outcomes and fewer adverse effects.
es, AI can significantly help identify existing drugs suitable for geroprotection (promoting healthy aging) by analysing vast datasets of drug properties, biological pathways, and aging-related molecular signatures to predict potential candidates, thereby accelerating the drug repurposing process and reducing development costs and timelines.
Some individuals age exceptionally healthy and have long lives with minimal periods of ill health late in life. Can AI help identify the reasons for this healthy aging, and potentially use the information to promote healthy aging more widely?
AI can help to identify the factors contributing to healthy aging by analysing complex datasets including genetics, lifestyle, and environmental factors to uncover patterns and biomarkers associated with longevity and resilience to age-related diseases. This information be used to develop personalised medicine based on predicting risks and creating tailored interventions.
A starting point is analysis of the genome, to detect genetic variants associated with healthy longevity. AI can also process large datasets that contain multiple types of information, including genetic, proteomic, metabolomic, and lifestyle. Its ability to pick up patterns in the data allows it to identify complex interactions and correlations that are crucial for understanding healthy aging. AI is also identifying novel biomarkers of risk of disease, for instance tissue-specific proteomes, which give early warning that an organ is failing. Analysis of personalised risk can also allow proactive monitoring and health management.
AI can help identify the psychological and social factors contributing to positive aging, such as optimism, social engagement, and adaptability, which can then be integrated into broader public health initiatives. It can also analyse lifestyle patterns to provide highly personalized recommendations for diet, exercise, social engagement, and other factors critical for healthy aging.
What ethical issues are raised by the use of AI in longevity medicine?
The use of AI in longevity medicine raises significant ethical issues.
Data privacy and security. AI systems require access to large amounts of highly sensitive personal health data, including genetic and clinical information. There is a risk of data breaches and unauthorized access, and also misuse by the main custodian, leading to privacy violations. Wearable devices can lead to near-constant surveillance and collection of excessive data.
AI could also exacerbate health disparities, and it will be important to ensure equity and access. Algorithms trained on non-representative data can lead to biased outcomes and worsen existing healthcare disparities. Ensuring equitable access to AI-driven longevity technologies across different socio-economic and geographic groups is a significant ethical challenge.
Autonomy and Consent. A lack of clear methods to obtain informed consent for AI involvement in treatment decisions is a challenge, as patients may not understand or trust AI’s role. Older adults, in particular, like to keep control over their data and decision-making.
There could be a danger of dehumanization of care because of an absence of empathy and compassion, which could lead to loss of patient confidence in health care, especially in older adults.
The outcome of AI analysis can lack transparency because it is not clear how the black-box system arrived at its conclusions. This can make it difficult for the clinician to explain to the patient how decision-making is being done. There is also a challenge for healthcare professionals to ensure that AI designs and algorithms, are appropriate and ethical guidelines need to be in place for accountability when AI is used in clinical settings.
It is also important that there is clarity of aims when using AI. Most people say they would like to live longer only if those extra years are healthy and not if they are unhealthy or excessively medicalised. These are complex outcomes for an AI analysis
Large amounts of data are being generated from wearable devices. What is and will be their impact for early detection and preventative measures?
A huge advantage of a wearable device is that the person can be monitored as they go about their normal life rather than having to attend a clinic.
Wearable devices and in-home monitoring can track vital signs, activity levels, and sleep patterns and thereby detect health issues early and enable personal decision making, personalised interventions and preventive care. Wearable devices can detect early signs of and monitor diseases, such as heart conditions by monitoring heart activity or diabetes by tracking glycaemic levels. By enabling self-monitoring, these devices empower individuals to take a more active role in their own health management, for instance by walking more or faster, sleeping better, and maintaining a healthy weight. Data accuracy, privacy and affordability are issues, as is the use of AI to analyse data outputs (see below). Devices should also be linked to the local healthcare system, which is not generally the case for most countries.
