Estimation of biological age of subjects through machine learning approaches: a new tool for population health screening?

 

Alessandro Gialluisi1, Augusto Di Castelnuovo1 and Licia Iacoviello1,2, on behalf of the Moli-sani Study Investigators

1 Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo - NEUROMED, Pozzilli, Italy

2 Department of Medicine and Surgery, University of Insubria, Varese, Italy

 

By 2050, over 21% of the global population will be over 60 years of age (1), leading to an increase in many age-related diseases and disabilities, with unsustainable costs for the public health systems. In this context, in the last years scientists have tried to develop estimators of Biological Age (BA), which can be defined as the hypothetical underlying age of an organism and can be computed based on a number of circulating and non-circulating biomarkers (2).

Therefore, identifying the biomarkers which have a prominent influence on BA and developing reliable models for its estimation is of fundamental importance for monitoring healthy ageing and could provide new tools to screen health status and the risk of clinical outcomes in the general population (3).

With this aim, we are testing both classical (regression-based) (4) and Machine Learning approaches (Deep Neural Networks, DNNs) (5,6) to estimate BA in the Moli-sani study, a prospective population-based cohort study of 24,325 subjects aged ≥35 years, resident in the Molise region. Within this cohort, we will apply DNN models to estimate BA based on circulating biomarkers, compare/integrate such estimate with classical linear regression-based estimation (4) and identify those biomarkers which will explain most of the variance in the model. Then we will compute the difference between BA and Chronological Age (CA), hereafter called DELTAage, to have a measure of higher or lower than expected ageing for each individual. Finally, we will test the influence of a number of environmental and genetic factors on such discrepancy, trying to identify those factors that improve (DELTAage < 0) and those that worsen ageing (DELTAage > 0). Moreover, since it has been proposed that biological ageing occurs at different rates in different tissues or cells within the same person – the so called “mosaic of ageing” (7) - we will apply the procedure of BA estimation to organ-specific sets of biomarkers, to compute organ-specific BA measures and compare them. This will allow us to identify potential factors exerting their influence more prominently on specific tissues, organs or systems (e.g., the central nervous system). Study design and preliminary results of the initiative will be discussed with participants and experts.

 

References

  1. Cohen JE. Human population: the next half century. Science 2003; 302:1172–5.
  2. Cole JH, Franke K. Predicting Age Using Neuroimaging : Innovative Brain Ageing Biomarkers. Trends in Neurosciences. 2017; 40(12): 681-690.
  3. Cole JH, et al. Brain age predicts mortality. Molecular psychiatry, doi:10.1038/mp.2017.62.
  4. Klemera P, Doubal S. A new approach to the concept and computation of biological age. Mech Ageing Dev. 2006; 127(3): 240-8.
  5. Mamoshina P, et al. Population specific biomarkers of human aging: a big data study using South Korean, Canadian and Eastern European patient populations. The journals of gerontology. Series A, Biological sciences and medical sciences, doi:10.1093/gerona/gly005.
  6. Putin E, et al. Deep biomarkers of human aging: Application of deep neural networks to biomarker development. Aging 2016; 8: 1021-1033, doi:10.18632/aging.100968 (2016).
  7. Cevenini E, Invidia L, Lescai F, Salvioli S, Tieri P, Castellani G et al. Human models of aging and longevity. Expert Opin Biol Ther 2008; 8: 1393–1405.