AI helped explain how the NOD2 gene affects macrophages—protective cells in the intestine associated with inflammatory ailments, particularly Crohn’s disease. The neural network helped solve a 25-year-old medical mystery. There are two main categories of macrophages in the intestine. Combat macrophages fight microbes, and reparative ones repair tissues. In Crohn’s disease, there is an imbalance between these types, leading to an excess of inflammatory immune cells. The reason for this is unknown. In 2001, the NOD2 gene was discovered, and defects in it increase the likelihood of developing the ailment. How this gene is linked to inflammation remained a mystery, but scientists from the University of California managed to answer this question. Using machine learning tools, they studied macrophage cells in the colon of healthy mice and those with simulated Crohn’s disease. It turned out that the balance of inflammatory and reparative immune cells is maintained by the interaction of NOD2 with another protein—Girdin. The AI helped the report’s authors establish that when there are defects in the NOD2 gene, the contact with Girdin is disrupted, and the immune cells remain in an inflammatory state. Experiments on mice proved that rodents without Girdin quickly develop intestinal inflammation and are more likely to die from sepsis. This information not only explains how a hereditary defect leads to the disease but also opens up new avenues for treating inflammatory bowel diseases. According to preliminary estimates, about 10 million individuals worldwide suffer from them. They severely impair quality of life, and available treatments are mostly aimed at reducing symptoms rather than eliminating the root cause. This new information could change that. If researchers can restore the balance of macrophages disrupted by mutations in the NOD2 gene, it could form the basis for a completely new therapy for Crohn’s disease.