Improved Human Age Prediction by Using Gene Expression Profiles From Multiple Tissues.

Area of Science:

• Genomics
• Aging Research
• Systems Biology

Background:

• Understanding aging requires analyzing transcriptome changes across tissues.
• Previous studies focused on single tissues or mouse models, leaving human multi-tissue aging unexamined.
• The Genotype-Tissue Expression (GTEx) project provides a rich dataset for this research.

Purpose of the Study:

• To develop a quantitative model for predicting human chronological age using multi-tissue gene expression data.
• To assess the predictive accuracy of single-tissue versus multi-tissue transcriptomes for age estimation.
• To investigate aging as a systemic process involving coordinated changes across the human body.

Main Methods:

• Utilized gene expression data from 46 human tissues from the GTEx project.
• Developed a model to predict biological age from single-tissue gene expression profiles.
• Systematically combined gene expression data from pairs of tissues to evaluate improvements in predictive accuracy.
• Measured performance using root-mean-square error (RMSE).

Main Results:

• The best single-tissue prediction (pituitary) achieved an RMSE of 3.92 years.
• Combining two tissues (pituitary and muscle) reduced the RMSE to 3.6 years.
• Demonstrated that different tissues possess varying potential for age prediction.
• Showcased improved prediction accuracy by integrating data from multiple tissues.

Conclusions:

• Aging is a complex, systemic process influenced by coordinated changes across multiple tissues.
• Multi-tissue transcriptome analysis offers a more accurate approach to estimating biological age.
• The developed quantitative model provides a novel tool for aging research.