global test and pairwise comparisons giving pvals of all NA or all 1

[jpkarl]

Hello.
Thank you for creating ANCOM BC2. I am trying to use the method to analyze species-level read counts from shotgun sequencing data from a crossover study in which I have 3 treatment groups (repeated measure) and several other factors to include in the model. I am using the code below. The results of the primary analysis look fine. However, in the global test I get W = 0 and p-val = NA for all taxa. For pairwise comparisons all p-vals = 1 (lfc and SE seem to be calculated fine).

Any guidance on what might be going wrong is appreciated.
Thank you

out = ancombc2(data = phylo,
fix_formula = "Treatment + Sequence + Phase",
rand_formula = "(1|Subject)",
p_adj_method = "BH",
pseudo_sens = TRUE,
prv_cut = 1,
lib_cut = 0,
s0_perc = 0.05,
group = "Treatment",
struc_zero = TRUE,
neg_lb = FALSE,
alpha = 0.20,
n_cl = 1,
verbose = TRUE,
global = TRUE,
pairwise = TRUE,
dunnet = TRUE,
trend = FALSE,
iter_control = list(tol = 0.01, max_iter = 20, verbose = TRUE),
em_control = list(tol = 1e-05, max_iter = 100),
lme_control = lme4::lmerControl(),
mdfdr_control = list(fwer_ctrl_method = "holm", B = 100),
trend_control = list(contrast = NULL, node = NULL, solver = "ECOS", B = 100)

[Maggie8888]

Could you please provide more information regarding your data?