rbind warning #321
Description
I have been running ANCOM-BC2 on a multiple phyloseq objects with the same metadata and options but consistently get an error for one of the phyloseq objects (I do not get errors for the others). The command I run is:
Ancomb_output = ancombc2(Phyloseq_object, fix_formula = "A + B * C", rand_formula = "(1|D)", p_adj_method = "BH", transform_abundances("log10"))
The output I get is:
Checking the input data type ...
The input data is of type: phyloseq
PASS
Checking the sample metadata ...
The specified variables in the formula: A,B,C
The available variables in the sample metadata: A,B,C,D
PASS
Checking other arguments ...
PASS
Obtaining initial estimates ...
Estimating sample-specific biases ...
ANCOM-BC2 primary results ...
Conducting sensitivity analysis for pseudo-count addition to 0s ...
For taxa that are significant but do not pass the sensitivity analysis,
they are marked in the 'passed_ss' column and will be treated as non-significant in the 'diff_robust' column.
For detailed instructions on performing sensitivity analysis, please refer to the package vignette.
Along with the following warning message:
Warning message:
In rbind(c(`(Intercept)` = 82.678914248258, Acategory1 = 17.8380370857971, :
number of columns of result is not a multiple of vector length (arg 47)
I am worried that this warning message is misaligning results but cannot figure out why it is occurring. I can share all the data I used for this if needed.