Scientists have combined Artificial Intelligence (AI), genetics and gut microbiome analysis to shed new light on one of the most serious complications of Crohn's disease: intestinal fibrosis, the irreversible scarring of the bowel that often leads to surgery. 

Publishing their findings in Frontiers in Artificial Intelligence, the research team led by experts from the University of Birmingham Dubai reveals that bowel fibrosis is not a separate disease stage. Instead, it appears to be a related inflammatory condition driven by ongoing immune activation, damage to the intestinal lining and changes in gut bacteria. 

Using machine learning, the researchers identified a shared set of 43 key genes linked to disease progression and grouped them into three major biological patterns: 

Co-author Dr Animesh Acharjee, from the University of Birmingham Dubai, said: "Crohn's disease affects more than four million people worldwide. One of the biggest challenges facing patients and clinicians is that fibrosis can develop gradually over time, leading to bowel narrowing, obstruction and repeated hospital treatment. 

"Our findings help to explain the biological transition from inflammation to irreversible bowel damage and identify potential biomarkers that could one day help doctors identify high-risk patients earlier." 

Researchers in Dubai analyzed data from 448 intestinal tissue transcriptomic samples and 80 microbiome samples, spanning healthy controls, Crohn's disease and fibrotic Crohn's disease. Participants were predominantly adults, with median ages ranging from the 30s to early 60s. 

Key microbiome findings revealed depletion of beneficial gut bacteria that produce short-chain fatty acids, including Faecalibacterium, Anaerostipes, Coprococcus and Ruminococcus, alongside increases in potentially harmful bacteria such as Bilophila and Bacteroides. These microbial shifts appear closely linked to the biological mechanisms driving disease progression. 

The researchers used generative AI to create realistic synthetic gene-expression data based on known biological rules. This helped them explore disease patterns despite limited numbers of fibrosis samples. Using AI-enhanced datasets improved the performance of machine-learning models and helped identify important fibrosis-related genes, including IL-23R, TNF-α, and TGF-β. These genes are already known to play major roles in inflammation and tissue scarring. 

Our research demonstrates a practical scientific application of generative AI to address a major challenge in biomedical research: limited patient datasets. We provide evidence that AI can help uncover disease mechanisms and improve biomarker discovery." 

By generating biologically realistic synthetic data, AI improved the ability of machine learning models to identify important fibrosis-related genes, including genes already known to play important roles in inflammatory and fibrotic disease processes." 

Dr. Animesh Acharjee, University of Birmingham Dubai

Crohn's disease affects people throughout adulthood and is commonly diagnosed in teenagers and young adults. Fibrosis is currently difficult to predict, monitor and treat, meaning many patients ultimately require surgery. Up to 70% of patients with transmural Crohn's disease develop bowel narrowing complications within 10 years of diagnosis. 

In the future, the molecular markers identified in this study could help doctors to detect patients at higher risk of developing fibrosis as well as distinguishing active inflammation from permanent scarring. The markers could also help scientists to develop more targeted treatments aimed at preventing bowel damage before surgery becomes necessary. 

Source:

Journal reference:

Philip, D., et al. (2026). Generative AI-augmented transcriptomic and microbiome analysis across inflammatory and fibrotic disease states in Crohn’s disease. Frontiers in Artificial Intelligence. DOI: 10.3389/frai.2026.1881820. https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2026.1881820/full