$W_{d}^{*}$-test: robust distance-based multivariate analysis of variance

Abstract

Background: Community-wide analyses provide an essential means for evaluation of the effect of interventions or design variables on the composition of the microbiome. Applications of these analyses are omnipresent in microbiome literature, yet some of their statistical properties have not been tested for robustness towards common features of microbiome data. Recently, it has been reported that PERMANOVA can yield wrong results in the presence of heteroscedasticity and unbalanced sample sizes. Findings: We develop a method for multivariate analysis of variance, W * d , based on Welch MANOVA that is robust to heteroscedasticity in the data. We do so by extending a previously reported method that does the same for two-level independent factor variables. Our approach can accommodate multi-level factors, stratification, and multiple post hoc testing scenarios. An R language implementation of the method is available at https://github.com/alekseyenko/ WdStar. Conclusion: Our method resolves potential for confounding of location and dispersion effects in multivariate analyses by explicitly accounting for the differences in multivariate dispersion in the data tested. The methods based on W * d have general applicability in microbiome and other 'omics data analyses.

A single strain of Bacteroides fragilis protects gut integrity and reduces GVHD

“…If statistical significance was not indicated, this means the relevant groups were compared but did not reach statistical significance ( P > 0.05). Community-level differences in microbiome were analyzed with Bray-Curtis dissimilarities using robust multivariate Welch MANOVA test Wd* from Hamidi et al 2019 ( 68 ). We applied univariate analyses to taxa that were present at more than 1% relative abundance in at least 1 sample.…”

Ecological Specialization Within a Carnivorous Fish Family Is Supported by a Herbivorous Microbiome Shaped by a Combination of Gut Traits and Specific Diet

“…Differences between species and diet types in terms of gut microbiome structure were tested using multivariate Welch MANOVA (W d test, Hamidi et al, 2019) implemented in the R package MicEco (Russel, 2020). This approach has the advantage, over classically used PERMANOVA, of being robust to heteroscedasticity in the data, i.e., differences in multivariate groups dispersion.…”

Unraveling the Microbial Mechanisms Underlying the Psychobiotic Potential of a Bifidobacterium breve Strain

“…Beta diversity was evaluated using PCA of the CLR‐transformed read counts dataset of the genus level and a multivariate Welch ANOVA test. [ 26,40 ] Every two groups showed a significant compositional difference (Figure 4E). Differentially abundant gut microbes between two groups were identified using the LEfSe.…”