UltraProcessed Food Proxy Gene Signature Associated with Inflammation in Patients with Crohn's Disease
Video
Team Information
Team Members
Kathryn Whyte, Postdoctoral Research Fellow, Department of Medicine, Columbia University Irving Medical Center
Faculty Advisor: Dympna Gallagher, Professor of Nutritional Medicine (in Medicine and in the Institute of Human Nutrition) at the Columbia University Medical Center
Abstract
Ultra-Processed Food Proxy Gene Signature Associated with Inflammation in Patients with Crohn’s Disease
Kathryn Whyte1; Jacob Nye1; Jia Kang2; Elham Azizi1,2; Dympna Gallagher1
1Columbia University Irving Medical Center, New York, NY 2Fu Foundation School of Engineering and Applied Sciences, Columbia University, New York, NY
Ultra-processed food (UPF) intake is associated with increased risk of obesity and other suboptimal outcomes across the lifespan. In the absence of cross-discipline collaboration, the tools to reveal underlying mechanisms remain unknown. Fortunately, the required information to create these tools exist across several databases. The purpose of the study was to identify and annotate proxy gene markers related to UPF from these databases and investigate expression in patients with Crohn’s disease.
Genes with increased expression in humans associated with UPF exposure were identified from the WHO/FAO Joint Expert Committee on Food Additives and published literature. DAVID pathway enrichment analysis on UPF gene signature list identified additional markers. The final gene signature list (n=128) was used to annotate a dataset of n=22 ileal biopsies. Specifically, single cell RNAseq data was normalized, and filtered and cells were classified as inflamed or non-inflamed. Clustering, differential gene expression analyses and Mann-Whitney U tests were then used to determine observable significance of UPF markers in dataset.
Leiden clustering with kNN=15 revealed co-expression of UPF and established inflammatory markers. Differential gene analyses showed significantly greater cell scores of UPF gene signature in inflamed tissue (p<0.001). Greater overlap of UPF genes was observed in inflammatory tissue.
Consensus from experts in dietary assessment of UPF and bioinformatics would be required to establish validity and reliability of methods presented here. Automation towards identifying this information would allow for translational clinical intervention studies to elucidate mechanisms to build prediction algorithms for personalized recommendations related to individual UPF response.
Contact this Team
Contact: Kathryn Whyte (use form to send email)
