Graduate student Anna Bogdanova has constructed a machine learning model to predict whether a bacterial flagellin is likely to bind to and/or activate the TLR5 receptor in the human gut. She developed FlaPro, a computational pipeline for quantifying flagellins in microbiome sequencing data, and discovered that inflammatory bowel diseases (IBD) are associated with altered proportions of “silent” versus “stimulatory” flagellins’ expression. Many thanks to everyone who contributed to this work. Please read the preprint here.