Viewing the MaAsLin2 Results

By navigating to the "Comparative Analysis" menu, you'll be able to view your generated MaAsLin2 comparatives. The comparative analysis will contain data in 6 different tabs (as described below).

MaAsLin2 Results

Input Data

A pre-formatted .tsv table containing taxonomic or functional profiling results (values dependent on input parameters).

Input Metadata

A pre-formatted .tsv table containing all sample metadata.

Association Results

Perhaps the most important output from MaAsLin 2 is the list of significant associations displayed in the Association Results, which displays significant associations between metadata and microbial features along with effect size, p-values, and q-values.

The full list of associations that pass MaAsLin 2's significance threshold is ordered by increasing q-values. Columns are:

Metadata: the variable name being associated with a microbial feature.
Feature Name: the microbial feature (taxon, gene, pathway, etc.).
Value Name: for categorical features, the specific feature level for which the coefficient and significance of association is being reported.
Coefficient/EffectSize: the model coefficient value
- Coefficients for categorical variables indicate the contrast between the each metadata category versus the reference.
Standard Error: the standard error from the model.
N: the total number of samples used in the model for this association (since minimum abundance/prevalence can exclude samples from comparisons).
N not Zero: the total of number of these samples in which the feature is non-zero.
P-value: the nominal significance of this association.
Q-value: the corrected significance is computed with p.adjust with the correction method (BH, etc.).

Box and Scatter Plots

For each of the significant associations in the Association Results, individual visualizations are generated for each feature (boxplots for categorical variables, scatter plots for continuous). These can be viewed by exporting the data as a .ZIP file. These visualization can be viewed within the CosmosID-HUB in the near future.

Heatmap

Top 50 (unless otherwise specified) features with significant associations.

Volcano

Volcano plot depicting all significant variables for each fixed effect

Forest

Forest plot helps to visually summarize the significant features effect size from the model contrasts along with their confidence intervals.

X-axis represents the contrast effect size; Y-axis represents the significant features that were found in respective contrasts of interest. The dot and the lines from the dot represents the individual feature’s effect size and the confidence interval of the effect size respectively

Exporting the Data

All data can be exported as a .ZIP file by navigating to the comparative analysis and selecting the "Export" button in the top right corner. This .ZIP file will contain all input tables, results table, all figures (PDF), and a maaslin2.log for debugging purposes.