Predicting Brain Age With Machine Learning and Transcriptome Profiling

In this study, researchers investigated age-associated gene expression changes in the prefrontal cortex of male and female brains and used machine learning to develop age prediction models.

The human brain is a complex organ, and its aging process is influenced by a plethora of factors, both genetic and environmental. Aging-related changes in the brain can lead to cognitive decline and susceptibility to neurodegenerative diseases. Therefore, understanding the molecular mechanisms underlying these changes is crucial for developing therapeutic strategies to delay or prevent age-related cognitive decline.

Over the past few years, a myriad of scientific studies have been conducted to understand the intricate relationship between our genes and the aging process. In a new study, researchers Joseph A. Zarrella and Amy Tsurumi from Harvard T.H. Chan School of Public Health, Massachusetts General Hospital, Harvard Medical School, and Shriner’s Hospitals for Children-Boston explored the concept of genome brain age prediction, a groundbreaking area of study that employs advanced bioinformatics tools to analyze changes in gene expression associated with aging. On February 28, 2024, their research paper was published and chosen as the cover paper for Aging’s Volume 16, Issue 5, entitled, “Genome-wide transcriptome profiling and development of age prediction models in the human brain.”

“[…] we aimed to profile transcriptome changes in the aging PFC [prefrontal cortex] overall and compare females and males, and develop prediction models for age.”

Transcriptome Profiling in the Prefrontal Cortex

The prefrontal cortex (PFC) plays a significant role in the aging process. It is responsible for a host of cognitive functions, including decision-making and planning. Throughout the aging process, significant transcriptome alterations occur in the PFC compared to other regions of the brain. These alterations can influence cognitive decline and susceptibility to neurodegenerative diseases.

Delving deeper into the complexities of aging, researchers have turned to transcriptome profiling as a powerful tool to uncover the molecular changes occurring within the prefrontal cortex. Transcriptome profiling allows scientists to measure the expression levels of all genes in a cell or tissue. By analyzing the transcriptome of the PFC, researchers can identify genes that are differentially expressed during the aging process. These genes can serve as potential biomarkers for age prediction.

The Study

In their groundbreaking research, Zarrella and Tsurumi aimed to develop prediction models for age based on the expression levels of specific panels of transcripts in the PFC. They leveraged advanced machine learning algorithms, including the least absolute shrinkage and selection operator (Lasso), Elastic Net (EN), eXtreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM), to develop accurate prediction models for chronological age.

The researchers used postmortem PFC transcriptome datasets obtained from the Gene Expression Omnibus (GEO) repository, ranging in age from 21 to 105 years. They identified differentially regulated transcripts in old and elderly samples compared to young samples and assessed the genes associated with age using ontology, pathway, and network analyses.

Machine learning algorithms were used to develop accurate prediction models for chronological age based on the expression levels of specific transcripts. The study found that specific gene expression changes in the PFC are highly correlated with age. Some transcripts showed female and male-specific differences, indicating that sex may play a role in the aging process at the molecular level.

Key Findings & Implications

The study identified several key genes whose expression levels change significantly with age. These genes include Carbonic Anhydrase 4 (CA4), Calbindin 1 (CALB1), Neuropilin and Tolloid Like 2 (NETO2), and Olfactomedin1 (OLFM1), among others. Many of these genes have been previously implicated in aging or aging-related diseases, validating the results of this study.

The researchers also developed four highly accurate age prediction models using different machine learning algorithms. These models were validated in a test set and an external validation set, demonstrating their potential application in predicting chronological age based on gene expression levels.

“Our results support the notions that specific gene expression changes in the PFC are highly correlated with age, that some transcripts show female and male-specific differences, and that machine learning algorithms are useful tools for developing prediction models for age based on transcriptome information.”

Conclusions & Future Directions

This study sheds light on the complex relationship between gene expression changes and the aging process in the human brain. The findings underscore the potential of using transcriptome profiling and machine learning algorithms for age prediction. The identified genes could serve as potential biomarkers for age prediction and may offer new insights into the molecular mechanisms underlying the aging process.

However, further validation of these models in larger populations and molecular studies to elucidate the potential mechanisms by which the identified transcripts may be related to aging phenotypes would be beneficial. Additionally, more inclusive studies investigating the interplay between genetic markers and factors such as sex, lifestyle, and environmental exposures are warranted.

In conclusion, this study provides a promising foundation for future research on genome brain age prediction. It also underscores the potential of transcriptome profiling and machine learning for exploring the complex interplay between our genes and the aging process. This approach could pave the way for personalized medicine strategies aimed at preventing or delaying age-related cognitive decline and neurodegenerative diseases.

Click here to read the full research paper published in Aging.

Aging is an open-access, traditional, peer-reviewed journal that publishes high-impact papers in all fields of aging research. All papers are available to readers (at no cost and free of subscription barriers) in bi-monthly issues at

Click here to subscribe to Aging publication updates.

For media inquiries, please contact [email protected].

  • Follow Us