Dataset Information

Precious1GPT: multimodal transformer-based transfer learning for aging clock development and feature importance analysis for aging and age-related disease target discovery.

ABSTRACT: Aging is a complex and multifactorial process that increases the risk of various age-related diseases and there are many aging clocks that can accurately predict chronological age, mortality, and health status. These clocks are disconnected and are rarely fit for therapeutic target discovery. In this study, we propose a novel approach to multimodal aging clock we call Precious1GPT utilizing methylation and transcriptomic data for interpretable age prediction and target discovery developed using a transformer-based model and transfer learning for case-control classification. While the accuracy of the multimodal transformer is lower within each individual data type compared to the state of art specialized aging clocks based on methylation or transcriptomic data separately it may have higher practical utility for target discovery. This method provides the ability to discover novel therapeutic targets that hypothetically may be able to reverse or accelerate biological age providing a pathway for therapeutic drug discovery and validation using the aging clock. In addition, we provide a list of promising targets annotated using the PandaOmics industrial target discovery platform.

SUBMITTER: Urban A

PROVIDER: S-EPMC10292881 | biostudies-literature | 2023 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Precious1GPT: multimodal transformer-based transfer learning for aging clock development and feature importance analysis for aging and age-related disease target discovery.

Urban Anatoly A Sidorenko Denis D Zagirova Diana D Kozlova Ekaterina E Kalashnikov Aleksandr A Pushkov Stefan S Naumov Vladimir V Sarkisova Viktoria V Leung Geoffrey Ho Duen GHD Leung Hoi Wing HW Pun Frank W FW Ozerov Ivan V IV Aliper Alex A Ren Feng F Zhavoronkov Alex A

Aging 20230613 11

Aging is a complex and multifactorial process that increases the risk of various age-related diseases and there are many aging clocks that can accurately predict chronological age, mortality, and health status. These clocks are disconnected and are rarely fit for therapeutic target discovery. In this study, we propose a novel approach to multimodal aging clock we call Precious1GPT utilizing methylation and transcriptomic data for interpretable age prediction and target discovery developed using ...[more]

PMID: 37315204

Similar Datasets

Project description:BackgroundBeyond mere prognostication, optimal biomarkers of aging provide insights into qualitative and quantitative features of biological aging and might therefore offer useful information for the testing and ultimately clinical use of gerotherapeutics. Using a large proteomic cohort in the UK Biobank (UKB), we aimed to develop a proteomic aging clock for all-cause mortality risk as a proxy of biological age (BA).MethodsParticipants in the UK Biobank Pharma Proteomics Project (UKB PPP) were included with ages between 39 and 70 years (n=53,021). Data were split into a training (70%, n=37,115) and a test set (30%, n=15,906), including 2,923 plasma proteins assessed using the Olink Explore 3072 assay®. We developed a proteomic aging clock (PAC) for all-cause mortality risk as a surrogate of BA using a combination of Least Absolute Shrinkage and Selection Operator (LASSO) penalized Cox regression and Gompertz proportional hazards models. The validation for PAC included assessing its age-adjusted associations with, and predictions for all-cause mortality and 18 incident diseases, and head-to-head comparisons with two biological age measures (PhenoAge and BioAge) and leukocyte telomere length (LTL). Additionally, a functional analysis was performed to identify gene sets and tissues enriched with genes associated with BA deviation, based on different BA measures.FindingsThe Spearman correlation between PAC proteomic age and chronological age was 0.76. 10.9% of the combined training and test samples died during a mean follow-up of 13.3 years (SD=2.2), with the mean age at death 70.1 years (SD=8.1). PAC proteomic age, after controlling for age and other covariates, showed stronger associations than PhenoAge, BioAge, and LTL, with mortality and multiple incident diseases in the test set sample and in disease-free participants, such as mortality, heart failure, pneumonia, delirium, Chronic Obstructive Pulmonary Disease (COPD), and dementia. Additionally, PAC proteomic age showed higher predictive power for the conditions above compared to chronological age, PhenoAge, and BioAge, based on Harrell's C-statistics. Proteins associated with PAC proteomic age deviation (from chronological age) are enriched in various hallmarks of biological aging, including immunoinflammatory responses, cellular senescence, extracellular matrix remodeling, cellular response to stressors, and vascular biology.InterpretationPAC showed robust age-adjusted associations and predictions for all-cause mortality and the onset of various diseases. The diverse hallmark gene sets linked with PAC proteomic age deviation highlight the potential efficacy of geroscience-guided interventions. Further validation is essential to ascertain the use of PAC across different settings.

Dataset Information

Precious1GPT: multimodal transformer-based transfer learning for aging clock development and feature importance analysis for aging and age-related disease target discovery.

Publications

Precious1GPT: multimodal transformer-based transfer learning for aging clock development and feature importance analysis for aging and age-related disease target discovery.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets