AI Tool Detects Hidden Signs of Cellular Aging

Researchers at the Bellvitge Biomedical Research Institute (Idibell), the Barcelona Supercomputing Center (BSC-CNS) and the Barcelona Institute for Global Health (ISGlobal) have developed an artificial intelligence tool capable of identifying signs of aging in blood stem cells by analysing microscopic images of their nuclei. The system, named ChromAgeNet, uses a convolutional neural network to detect subtle, age-associated patterns in the three-dimensional organisation of chromatin, the complex of DNA and proteins inside the cell nucleus. These structural changes are not visible to the human eye but influence gene activity and cell function. In a study published in the journal Aging Cell, the team trained ChromAgeNet using high-resolution 3D images of mouse hematopoietic stem cell nuclei stained with DAPI, a standard technique for visualising DNA. The model achieved a 77% probability of correctly classifying cells as young or aged based on nuclear appearance alone.

Key Features Linked to Aging Identified

ChromAgeNet outperformed earlier machine learning models based on manually defined chromatin features. By analysing which image characteristics contributed most to its predictions, researchers identified specific markers of aging: increased chromatin entropy, accumulation of heterochromatin at the nuclear periphery, and altered chromatin condensates. These findings suggest that the spatial arrangement of genetic material deteriorates with age in a measurable way, potentially contributing to declining blood and immune system function.

The research team treated aged cells with compounds known to modify chromatin packaging. Some treatments shifted the chromatin structure toward a more youthful configuration, changes that ChromAgeNet was able to detect. While it remains unverified whether this structural rejuvenation restores full cellular function, the results open avenues for exploring therapies aimed at preserving stem cell performance in older adults. Maria Carolina Florian, Icrea researcher at Idibell’s regenerative medicine programme, stated: “The tool we have developed is important because it helps us understand the mechanisms that can be a therapeutic target.”

“The tool we have developed is important because it helps us understand the mechanisms that can be a therapeutic target,” points out Maria Carolina Florian, Icrea researcher at the Idibell regenerative medicine program.

The next phase will involve testing whether ChromAgeNet can identify similar aging signatures in human blood stem cells and assessing its potential to screen for interventions that slow or reverse these changes. The project forms part of doctoral research by Pablo Iañez at ISGlobal and received funding from the European Research Council (ERC Consolidator Grant), the Spanish Ministry of Science, and the La Caixa Foundation.

Reported by en.ara.cat, medicalxpress.com, brightsurf.com, azolifesciences.com, idibell.cat, drugtodayonline.com, abc.com.py, p-cmrc.cat, gepris.dfg.de, Diari ARA.