Irritable Bowel Syndrome (IBS) is a prevalent functional gastrointestinal disorder, affecting approximately 10-14% of the global population. Characterized by symptoms of abdominal pain and disordered bowel patterns, IBS is a chronic condition that significantly impairs the quality of life and work productivity for those affected. On average, individuals with IBS miss two days of work per month, and 23% of patients often stay home due to their symptoms. Despite its prevalence, IBS remains challenging to diagnose accurately and promptly due to its complex pathophysiology, which involves multiple risk factors, comorbidities, and a wide range of phenotypes.

Lack of a definitive diagnostic test for IBS means that clinicians often rely on a thorough medical history and a series of tests to rule out other conditions. This complexity leads to delays in diagnosis and treatment, further exacerbating the impact on patients’ lives. Chronic and relapsing nature of IBS necessitates long-term management, making it crucial to identify the most associated comorbidities and risk factors to improve diagnosis and treatment.

This research leverages multidimensional clinical data and unsupervised machine learning techniques to identify the endo-phenotypes of IBS. By analyzing natural data, the study uncovers the diseases, symptoms, and dietary practices that are strongly associated with persistent IBS symptoms. The goal is to provide clinicians with a more comprehensive understanding of IBS, enabling more timely and accurate diagnoses. Through this approach, the research aims to alleviate the burden on patients and improve the overall management of IBS by offering a data-driven solution to identify the most significant comorbidities and risk factors associated with this disorder.