AI Models Demonstrate Links Between IPF, Aging Processes

A pair of new artificial intelligence (AI)-driven models are helping to clarify the relationship between aging and idiopathic pulmonary fibrosis (IPF). The new research suggests that IPF is not merely the result of the acceleration of normal aging but rather represents dysregulation of the aging process. The study was published in the journal Aging.¹

IPF is marked by declining lung function and eventual respiratory failure. It primarily affects patients over the age of 60. The latter fact has prompted a considerable amount of research into how aging-related mechanisms might affect the pathogenesis of IPF, noted corresponding author Alex Zhavoronkov, PhD, and colleagues. Zhavoronkov is the founder and chief executive of Insillico Medicine, a biotechnology company that has focused its work on using AI to accelerate drug discovery and life sciences research.

IPF is an important research area, the authors noted, because existing treatments cannot address the underlying causes of the disease. Instead, most therapies can only slow down the disease’s progression. Another treatment option is lung transplantation, which is the only known way to improve a patient’s prognosis, they said.

The IPF-Precious3GPT analysis found that aging lungs and fibrotic disease had shared gene expression patterns, but also unique patterns. | Image credit: Dr_Microbe - stock.adobe.com

In hopes of improving the prognosis for patients with IPF, the investigators developed a pair of AI-based models. First, they used proteomics data from the UK Biobank to train a fibrosis-aware aging clock that can assess a particular patient’s biological age. Next, they created IPF-Precious3GPT, an omics transformer that can generate differential gene expression profiles. The model allowed the investigators to create 2 distinct gene expression profiles. The first represented the classical IPF transcriptomic response, and the second represented the aging process in lung tissue over the time span from a patient’s 30th year through their 70th year. The model was able to discern the importance of particular genes to each respective biological process.

In cross-validation, the aging clock was able to predict a patient’s biological age with a high accuracy of R² = 0.84 and a mean absolute error of 2.68 years. After validating the model, the investigators applied it to a dataset of patients who were healthy or had moderate or severe COVID-19. The analysis found that patients with severe infections—those who were also more likely to develop lung fibrosis—also had a higher biological age (+2.77 years; P = .026).

The IPF-Precious3GPT analysis found that aging lungs and fibrotic disease had shared gene expression patterns, but also unique patterns. This suggests that IPF is more than just accelerated aging, but rather a unique pathological process. The authors identified 4 key pathways—TGF-β signaling, oxidative stress, inflammation, and extracellular matrix remodeling—that are key factors in IPF and aging, albeit with genetic-level differences.

“While numerous additional pathways play roles in IPF, these 4 were prioritized based on their documented centrality to disease progression and their known modulation during aging,” the authors said.²

Next steps will include validating their models in dedicated IPF cohorts and extending the approach to other fibrotic and age-related diseases.¹

“Beyond IPF, our approach holds potential for investigating other fibrotic conditions, such as liver cirrhosis, NAFLD (non-alcoholic fatty liver disease), kidney fibrosis, and systemic sclerosis, where aging-related mechanisms may similarly contribute to pathogenesis,” the authors wrote.

They also hope to use the tools for drug discovery, biomarker identification, and personalized medicine strategies, the company said.

The investigators also cautioned, though, that their models will benefit from additional validation. They added that their research primarily relied on transcriptomic and proteomic data, which may not adequately account for epigenetic factors that contribute to the relationship between IPF and aging.

References

1. Galkin F, Chen S, Aliper A, Zhavoronkov A, Ren F. AI-driven toolset for IPF and aging research associates lung fibrosis with accelerated aging. Aging (Albany NY). Published online August 8, 2025. doi:10.18632/aging.206295

2. Ren LL, Miao H, Wang YN, Liu F, Li P, Zhao YY. TGF-β as a master regulator of aging-associated tissue fibrosis. Aging Dis. 2023;14(5):1633-1650. doi:10.14336/AD.2023.0222