Predicting physiological aging rates from a range of quantitative traits using machine learning

Sun, Eric; Qian, Yong; Oppong, Richard; Butler, Thomas J.; Zhao, Jing; Chen, Brian H.; Tanaka, Toshiko; Kang, Jian; Sidore, Carlo; Cucca, Francesco; Bandinelli, Stefania; Abecasis, Gonçalo R.; Gorospe, Myriam; Ferrucci, Luigi; Schlessinger, David; Goldberg, Ilya G.; Ding, Jun

doi:10.18632/aging.203660

Cited by 8 publications

(8 citation statements)

References 90 publications

(100 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, regression-based aging clocks trained on DNA methylation profiles from multiple tissues ('epigenetic aging clocks') [9][10][11][12][13] or blood plasma protein profiles [14][15][16][17] have striking performance to predict chronological age in humans. Aging clocks directly optimized to predict biological age have also been developed on functional phenotypes 12,13,18 or time remaining until death 19,20 . Interestingly, beneficial health interventions such as diet and exercise [21][22][23] and genetic manipulations [24][25][26] result in younger predictions from epigenetic aging clocks trained on chronological age.…”

mentioning

confidence: 99%

Cell-type-specific aging clocks to quantify aging and rejuvenation in neurogenic regions of the brain

et al. 2022

Self Cite

View full text Add to dashboard Cite

The diversity of cell types is a challenge for quantifying aging and its reversal. Here we develop ‘aging clocks’ based on single-cell transcriptomics to characterize cell-type-specific aging and rejuvenation. We generated single-cell transcriptomes from the subventricular zone neurogenic region of 28 mice, tiling ages from young to old. We trained single-cell-based regression models to predict chronological age and biological age (neural stem cell proliferation capacity). These aging clocks are generalizable to independent cohorts of mice, other regions of the brains, and other species. To determine if these aging clocks could quantify transcriptomic rejuvenation, we generated single-cell transcriptomic datasets of neurogenic regions for two interventions—heterochronic parabiosis and exercise. Aging clocks revealed that heterochronic parabiosis and exercise reverse transcriptomic aging in neurogenic regions, but in different ways. This study represents the first development of high-resolution aging clocks from single-cell transcriptomic data and demonstrates their application to quantify transcriptomic rejuvenation.

show abstract

mentioning

confidence: 99%

Cell-type-specific aging clocks to quantify aging and rejuvenation in neurogenic regions of the brain

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…The strongest markers of mortality and hospitalisation risk were Cystatin-C, N-terminal-pro hormone B-type natriuretic peptide (NT-proBNP), and gender. The Physiological Ageing score (PA) [(Sun et al 2021 ), Table 2 ] was derived from two independent cohorts of individuals in long-lived communities (SardiNIA and InCHIANTI). The ratio of PA to chronological age (PAR) was found to be a significant predictor of survival as well as a proxy for whole-body ageing.…”

Section: Ageing Scoresmentioning

confidence: 99%

“…Major contributors to BA were cystatin-C, NT-proBNP and gender. A decelerated BA was associated with higher physical and mental well-being, healthy lifestyle and higher socioeconomic status, while accelerated ageing was associated with smoking and obesity Physiological ageing rate-PAR (Sun et al 2021 ) Predict physiological ageing rate from quantitative traits. Identify genetic loci by GWAS.…”

Section: Ageing Scoresmentioning

confidence: 99%

Measuring healthy ageing: current and future tools

et al. 2023

View full text Add to dashboard Cite

Human ageing is a complex, multifactorial process characterised by physiological damage, increased risk of age-related diseases and inevitable functional deterioration. As the population of the world grows older, placing significant strain on social and healthcare resources, there is a growing need to identify reliable and easy-to-employ markers of healthy ageing for early detection of ageing trajectories and disease risk. Such markers would allow for the targeted implementation of strategies or treatments that can lessen suffering, disability, and dependence in old age. In this review, we summarise the healthy ageing scores reported in the literature, with a focus on the past 5 years, and compare and contrast the variables employed. The use of approaches to determine biological age, molecular biomarkers, ageing trajectories, and multi-omics ageing scores are reviewed. We conclude that the ideal healthy ageing score is multisystemic and able to encompass all of the potential alterations associated with ageing. It should also be longitudinal and able to accurately predict ageing complications at an early stage in order to maximize the chances of successful early intervention.

show abstract

“…In recent years, the number of publications describing machine learning models for the estimation of chronological and biological age have risen dramatically. The most well-known example is that of the “epigenetic clock,” although models have also been developed using transcriptomics, miRNA, proteomics, and clinical phenotypes ( Peters et al, 2015 ; Horvath and Raj 2018 ; Huan et al, 2018 ; Tanaka et al, 2018 ; Sun et al, 2021 ). Here, we define a “model” as a specific algorithm that uses a specific set of input variables ( e.g.…”

Section: Introductionmentioning

confidence: 99%

“…However, one unintended consequence of advances in generating large amounts of data and increased computing efficiency has been the relative ease and appeal of developing machine learning models using any available data. In some cases, available data included measures that were unique to the original dataset, thereby making independent validation of the model inherently challenging ( Sun et al, 2021 ).…”

Section: Introductionmentioning

confidence: 99%

Minimum standards for evaluating machine-learned models of high-dimensional data

Chen

2022

Front. Aging

Self Cite

View full text Add to dashboard Cite

The maturation of machine learning and technologies that generate high dimensional data have led to the growth in the number of predictive models, such as the “epigenetic clock”. While powerful, machine learning algorithms run a high risk of overfitting, particularly when training data is limited, as is often the case with high-dimensional data (“large p, small n”). Making independent validation a requirement of “algorithmic biomarker” development would bring greater clarity to the field by more efficiently identifying prediction or classification models to prioritize for further validation and characterization. Reproducibility has been a mainstay in science, but only recently received attention in defining its various aspects and how to apply these principles to machine learning models. The goal of this paper is merely to serve as a call-to-arms for greater rigor and attention paid to newly developed models for prediction or classification.

show abstract

Predicting physiological aging rates from a range of quantitative traits using machine learning

Cited by 8 publications

References 90 publications

Cell-type-specific aging clocks to quantify aging and rejuvenation in neurogenic regions of the brain

Cell-type-specific aging clocks to quantify aging and rejuvenation in neurogenic regions of the brain

Measuring healthy ageing: current and future tools

Minimum standards for evaluating machine-learned models of high-dimensional data

Contact Info

Product

Resources

About