Direction of effect size

Downstream analysis and statistics MaAsLin
1

I have a question regarding the direction of effect sizes in the comparison. My data is zero-inflated so I have problems seeing which group has highest abundance of the genus. I have groups A and B and get a positive coef (2.61) for Neisseria. Does this mean that Neisseria is higher in group B compared to group A (Since A would be the reference, alphabetically)? Please note that the group-sizes are very different, which confuses the vizualisation.


> table(pdata$Group)

   A    B 
4876   72


fit_data = Maaslin2(
  input_data = counts.t, 
  input_metadata = pdata,
  analysis_method = "LM",
  transform = "NONE",
  min_prevalence = 0,
  min_abundance = 0.01,
  normalization = "NONE",
  output = "demo_output", 
  plot_scatter=T,
  plot_heatmap=T,
  fixed_effects = c("Group"))

head(fit_data$results)
feature metadata value coef stderr pval name qval N N.not.zero
1 Capnocytophaga Group B 0.8479733 0.06873646 1.840022e-34 GroupB 2.189627e-32 4948 55
2 Neisseria Group B 2.6103594 0.26466737 9.770062e-23 GroupB 5.813187e-21 4948 149

Group_2
Group_2960×960 55.2 KB

2

Hi @Robin_Mjelle,

Yes, you are interpreting this correctly. MaAsLin is seeing Neisseria has higher in group B. I agree also that due to the differences in group size the stats here are hard to see visually.

I hope this helps but let us know if you have any further questions.

Best,
Kelsey