MaAsLin3, summary plots not generating but association plots are

Downstream analysis and statistics MaAsLin
1

Hi,

I’m using the most recent version of maaslin3 to compare two treatments. I am using ASVs for my features. When I run maaslin3, it runs properly, and outputs the linear and logistic association plots. However, it is not generating the summary plot whatsoever. This is the only thing I am finding in the log file about the summary plot.

INFO::Writing summary plot of significant results to file: (sampletype)_maaslin3/figures/summary_plot.pdf

2026-07-01 12:36:16.77 INFO::The following specified variables were not found in the associations: Treatment

2026-07-01 12:36:16.77 INFO::Available associations: Treatment (Treatment_type1)

The association plots indicate that maaslin3 is able to recognize both treatment types from my metadata treatment column, as they are both present in the plotting.

Here is the code I am using:

maaslin3_ulva ← maaslin3(input_data = feature_abundance_table,

                      input_metadata = feature_metadata_table, 

                      output = '(sampletype)\_maaslin3', 

                      formula = c("Treatment"),

                      normalization = 'TSS',

                      transform = 'LOG',

                      augment = TRUE, 

                      median_comparison_abundance = TRUE, 

                      median_comparison_prevalence = FALSE, 

                      heatmap_vars = c("Treatment"),

                      plot_summary_plot = TRUE,

                      max_pngs = 250)

Hopefully I’m just missing something. Thanks.

2

Hi,

I think the heatmap_vars variable needs to be set to "Treatment (Treatment_type1)" or some version of what that info line is telling you. Does that work?

Will

3

Hi,

Thanks for your response. I think I have more of a data problem than a maaslin problem at this point, as I am able to get it to successfully generate with one version of this data set, but not another. Back to troubleshooting.

4

Glad it worked for at least one dataset, let me know if you can’t get it working for the other.