mi-Mic: better analysis, better results

In the world of big data, it is impossible to interpret results without employing statistical tests. In order to achieve reliable conclusions, the appropriate test should be applied on each type of dataset. Most statistical tests demand normal distribution of the results, as well as independence between the studied parameters, since any dependency may skew the results. Microbiome data, the identification of bacterial species in a sample, contains both difficulties. Bacterial samples tend to contain big portions of specific species, while most bacteria would not be represented at all. Moreover, the similarity between different strains, that are usually related to each other, adds a dependency factor that is not acceptable in most conventional statistical tests. In an article published in Genome Biology, Prof. Yoram Louzoun and his team present a new statistical tool called mi-Mic – Mann Whitney iMage Microbiome, a novel framework to apply differential abundance analysis to the non-normally distributed microbial data.