Exploring the causal role of the human gut microbiome on colorectal cancer: Application of Mendelian randomization


Session type:

Kaitlin Wade1,Long Nguyen2,Amit Joshi2,Robert Carreras-Torres3,Andrew Chan2,Marc Gunter4,Yi Lin5,Graham Casey6,Jane Figueiredo7,Stephen Gruber8,Jochen Hampe9,Heather Hampel10,Mark Jenkins11,Temitope Keku12,Ulrike Peters13,Catherine Tangen14,Anna Wu15,David Hughes1,Malte Rühlemann16,Jeroen Raes17,Nicholas Timpson1

1MRC Integrative Epidemiology Unit, Population Health Sciences, Bristol Medical School,2Massachusetts General Hospital,3IDIBELL Bellvitge Biomedical Research Institute,4International Agency for Research on Cancer,5Fred Hutchinson Cancer Research Center,6Center for Public Health Genomics,7Department of Medicine, Samuel Oschin Comprehensive Cancer Institute,8Department of Preventive Medicine, USC Norris Comprehensive Cancer Center,9Department of Medicine I, University Hospital Dresden,10Division of Human Genetics, Department of Internal Medicine, The Ohio State University,11Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health,12Center for Gastrointestinal Biology and Disease, University of North Carolina,13Public Health Sciences Division, Fred Hutchinson Cancer Research Center,14SWOG Statistical Center, Fred Hutchinson Cancer Research Center,15University of Southern California,16Institute of Clinical Molecular Biology, Christian Albrechts University of Kiel,17Center for Microbiology, VIB, Department of Microbiology and Immunology, Rega Instituut



Despite compelling results from in vivo and in vitro models providing evidence supporting associations between the human gut microbiome and colorectal cancer (CRC), few findings have been translated between model organisms and human studies have been unconvincing in their ability to offer causal evidence. Such discrepancies are mainly due to the challenges in multi-omic technologies (e.g., sequencing and metagenomics) and sensitive experimental models. Additionally, existing epidemiological studies are of cross-sectional or case-control design; therefore, confounding, reverse causation and bias are important limitations to consider. Such evidence that is lost in translation impedes any opportunity for harnessing the gut microbiome – an ideally malleable trait – for improving population health. Mendelian randomization (MR) is a method that uses human genetic variation as instruments for a clinically relevant trait to improve causal inference.


We used MR to interrogate the causal impact of the gut microbiome on CRC combining summary-level data from genome-wide association studies (GWASs) of host genotype and gut microbiome variation from the Flemish Gut Flora Project and two German cohorts (n=3890) with the Genetics and Epidemiology of Colorectal Cancer Consortium (n=120328). Analyses were conducted in MR-Base and pleiotropy was assessed using PhenoScanner and MR-TRYX.


Of the 157 microbial traits (MTs) assessed in the GWAS of the gut microbiome, there was evidence for host genetic contributions to 13 MTs. Of these, we found evidence that presence of a genus within a certain order of bacteria increased the risk of CRC by 8% (95% CI:2-15%;P=0.02), with no strong evidence that the SNP used as an instrument was associated with other traits, discounting the likelihood of pleiotropy.


Our study confirmed observational evidence suggesting that genera within this bacterial order are more present in CRC cases than controls and provides further evidence that this may be due to a causal impact of these bacteria on CRC.