In recent years, human microbiota, especially gut microbiota, have emerged as an important yet complex trait influencing human metabolism, immunology, and diseases. Many studies are investigating the forces underlying the observed variation, including the human genetic variants that shape human microbiota. Several preliminary genome-wide association studies (GWAS) have been completed, but more are necessary to achieve a fuller picture. Here, we announce the MiBioGen consortium initiative, which has assembled 18 population-level cohorts and some 19,000 participants. Its aim is to generate new knowledge for the rapidly developing field of microbiota research. Each cohort has surveyed the gut microbiome via 16S rRNA sequencing and genotyped their participants with full-genome SNP arrays. We have standardized the analytical pipelines for both the microbiota phenotypes and genotypes, and all the data have been processed using identical approaches. Our analysis of microbiome composition shows that we can reduce the potential artifacts introduced by technical differences in generating microbiota data. We are now in the process of benchmarking the association tests and performing meta-analyses of genome-wide associations. All pipeline and summary statistics results will be shared using public data repositories. We present the largest consortium to date devoted to microbiota-GWAS. We have adapted our analytical pipelines to suit multi-cohort analyses and expect to gain insight into host-microbiota cross-talk at the genome-wide level. And, as an open consortium, we invite more cohorts to join us (by contacting one of the corresponding authors) and to follow the analytical pipeline we have developed.

Typical data in a microbiome study consist of the operational taxonomic unit (OTU) counts that have the characteristic of excess zeros, which are often ignored by investigators. In this paper, we compare the performance of different competing methods to model data with zero inflated features through extensive simulations and application to a microbiome study. These methods include standard parametric and non-parametric models, hurdle models, and zero inflated models. We examine varying degrees of zero inflation, with or without dispersion in the count component, as well as different magnitude and direction of the covariate effect on structural zeros and the count components. We focus on the assessment of type I error, power to detect the overall covariate effect, measures of model fit, and bias and effectiveness of parameter estimations. We also evaluate the abilities of model selection strategies using Akaike information criterion (AIC) or Vuong test to identify the correct model. The simulation studies show that hurdle and zero inflated models have well controlled type I errors, higher power, better goodness of fit measures, and are more accurate and efficient in the parameter estimation. Besides that, the hurdle models have similar goodness of fit and parameter estimation for the count component as their corresponding zero inflated models. However, the estimation and interpretation of the parameters for the zero components differs, and hurdle models are more stable when structural zeros are absent. We then discuss the model selection strategy for zero inflated data and implement it in a gut microbiome study of > 400 independent subjects.