Maaslin3 summary plot only showing 1 variable heatmap

Downstream analysis and statistics MaAsLin
1

Hello,

I ran a longitudinal analysis on a subset of my data because I was testing a new metadata format (each sample had 3 timepoints (T1-3) and a control (T0)) and the summary plot only showed a heatmap for T2 and then only showed the box and whisker plots for T1 and T3. I’ll attach it here. It did suggest using small_random_effects = True because of how small the sample size was, and I did re-run it, but nothing came back as significant. I wanted to ask if the summary plot problem was a result of the small sample size or the command/analysis being wrong before reformatting the rest of my metadata. Thanks!
This was the command I used:
fit_time ← maaslin3(

input_data = df_input_data,

input_metadata = df_t3_subset_meta_v1,

output = “t3_subset_only_longitudinal_smallrandeffects”,

fixed_effects = c(“Time_point”, “Group”),

random_effects = c(“sample_num”),

small_random_effects = TRUE

normalization = “TSS”,

transform = “LOG”,

reference = c(“Time_point,T0”, “Group,Healthy”))

summary_plot
summary_plot6250×3240 461 KB

2

Hi, when you refer to the “summary plot problem,” do you mean the fact that 2 variables are coming out as coefficient plots and the other one as a heatmap? If so, you can set which variables show up where using the arguments heatmap_vars and coef_plot_vars (or see this section of the wiki). If you meant something else was wrong, could you clarify what you mean?

Will