Research Paper Advance Articles

AI-driven toolset for IPF and aging research associates lung fibrosis with accelerated aging

Fedor Galkin1, , Shan Chen2, , Alex Aliper1, , Alex Zhavoronkov1,2,3,4, , Feng Ren2, ,

  • 1 Insilico Medicine AI Ltd., Abu Dhabi, UAE
  • 2 Insilico Medicine Shanghai, Ltd., Shanghai, China
  • 3 Insilico Medicine US, Inc., New York, NY 10010, USA
  • 4 Insilico Medicine Hong Kong, Ltd., Hong Kong, China

Received: January 17, 2025       Accepted: July 15, 2025       Published: August 8, 2025      

https://doi.org/10.18632/aging.206295
How to Cite

Copyright: © 2025 Galkin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Abstract

Idiopathic pulmonary fibrosis (IPF) is a condition predominantly affecting the elderly and leading to a decline in lung function. Our study investigates the aging-related mechanisms in IPF using artificial intelligence (AI) approaches. We developed a pathway-aware proteomic aging clock using UK Biobank data and applied it alongside a specialized version of Precious3GPT (ipf-P3GPT) to demonstrate an AI-driven mode of IPF research. The aging clock shows great performance in cross-validation (R2=0.84) and its utility is validated in an independent dataset to show that severe cases of COVID-19 are associated with an increased aging rate. Computational analysis using ipf-P3GPT revealed distinct but overlapping molecular signatures between aging and IPF, suggesting that IPF represents a dysregulation rather than mere acceleration of normal aging processes. Our findings establish novel connections between aging biology and IPF pathogenesis while demonstrating the potential of AI-guided approaches in therapeutic development for age-related diseases.

Abbreviations

AI: Artificial intelligence; CV: Cross validation; IPF: Idiopathic pulmonary fibrosis; LLM: Large language model; MAE: Mean absolute error; OSF: Opens Science Framework; P3GPT: Precious3GPT; R2: Coefficient of determination; USPTO: United States Patent and Trademark Office.