super(1)H NMR based metabolic profiling in Crohn's disease by random forest methodology

• Published in: Magnetic resonance in chemistry 2014-07, Vol.52 (7), p.370-376
• Main Authors: Fathi, Fariba, Majari-Kasmaee, Laleh, Mani-Varnosfaderani, Ahmad, Kyani, Anahita, Rostami-Nejad, Mohammad, Sohrabzadeh, Kaveh, Naderi, Nosratollah, Zali, Mohammad Reza, Rezaei-Tavirani, Mostafa, Tafazzoli, Mohsen, Arefi-Oskouie, Afsaneh
• Format: Article
• Language: English
• Subjects: Classification; Correlation; Crohn's Disease; Forests; Mathematical models; Metabolites; Serums; Zinc
• ISSN: 0749-1581; 1097-458X
• DOI: 10.1002/mrc.4074

The present study was designed to search for metabolic biomarkers and their correlation with serum zinc in Crohn's disease patients. Crohn's disease (CD) is a form of inflammatory bowel disease that may affect any part of the gastrointestinal tract and can be difficult to diagnose using the clinical tests. Thus, introduction of a novel diagnostic method would be a major step towards CD treatment. Proton nuclear magnetic resonance spectroscopy ( super(1)H NMR) was employed for metabolic profiling to find out which metabolites in the serum have meaningful significance in the diagnosis of CD. CD and healthy subjects were correctly classified using random forest methodology. The classification model for the external test set showed a 94% correct classification of CD and healthy subjects. The present study suggests Valine and Isoleucine as differentiating metabolites for CD diagnosis. These metabolites can be used for screening of risky samples at the early stages of CD diagnoses. Moreover, a robust random forest regression model with good prediction outcomes was developed for correlating serum zinc level and metabolite concentrations. The regression model showed the correlation (R super(2)) and root mean square error values of 0.83 and 6.44, respectively. This model suggests valuable clues for understanding the mechanism of zinc deficiency in CD patients. Copyright copyright 2014 John Wiley & Sons, Ltd. The metabolites that caused changes in people with crohn's disease were identified using random forest and the obtained classification model showed 94% correct classification of crohn's disease and healthy subject for the external test set. As an innovative approach, a model was made for correlation between these metabolites and serum levels of zinc. The regression model of levels of zinc and metabolites resulted correlation (R super(2)) value of 0.83 for the external test set.