Meta-analysis of human genome-microbiome association studies: the MiBioGen consortium initiative
It has now been well established that the microbial community living in our gut – the gut microbiome – impacts our lives in many ways, from affecting mood and quality of life, to predisposing us to particular diseases or influencing the efficacy of disease treatment.
Every one of us has a unique gut microbial ecosystem, and this uniqueness is determined by numerous factors that include our personal habits, food preferences and environment. Another element of this uniqueness is determined by our genetics, as demonstrated by several twin and family studies that showed some bacteria are more heritable than others, being more common in relatives and monozygotic twins compared to non-related individuals.
Several initial proof-of-concept studies examining the contribution of genetics to microbiome composition have now been carried out by four reasonably large-scale, population studies (>1000 individuals) performed on Canadian, German, Dutch and UK populations. Surprisingly, the overlap between the human genes these studies identified as affecting the microbiome was relatively small. The most consistent finding, now confirmed by several studies, is the association with genetic variants in the lactase gene, LCT, which determines our ability to digest the milk sugar lactose. Genetic variants in LCT were found to be associated with the abundance of Bifidobacteria, a commensal bacteria present in milk products.
There are several factors that can explain the very limited overlap seen in the results of the genetics-microbiome association studies published to date. First, the sample size of each cohort, between 1000 to 2000 individuals, is considered to be quite small for this type of analyses. Given that the potential effect of genetic variants on the level of different bacteria is relatively small, bigger cohorts are clearly needed to detect significant SNP-bacterial associations from millions of SNPs and hundreds of bacterial species. Second, the methods used to characterize the microbiome and data analysis used in these studies were very different, sometimes remarkably so, which makes impossible to directly compare the outcomes.
We moved to address both these problems by initiating the MiBioGen consortium, a collaborative initiative that aims to further explore the role of genetic background in determining gut microbiome composition. Currently, MiBioGen is made up of 20 research groups and includes data from more than 20,000 people. To address the difference in results due to differences in methods, we have developed uniform pipelines for the every step of the analysis, including both microbiome and genetic data processing, the exploration of associations, and the meta-analysis procedure.
We recently published a paper announcing the creation of the MiBioGen consortium in Microbiome. In this paper, we describe the analysis protocols, quality control and study design used in our uniform pipeline and provide information on the databases and software necessary to join the collaboration. We also describe our plans for further development of our collaboration and the potential of “microbiome genetics” for personalized treatment and prevention medicine. We welcome all groups with available genetic and gut microbiome data to join the MiBioGen consortium.
Our announcement paper in Microbiome can be found here:
Posted by Alexander Kurilshikov on behalf of MiBioGen consortium