Hello everyone.
I have doubts regarding the interpretation of data in ANCOM-BC in the barplot, with 16S metabarcoding data.
If I wanted to evaluate the significant taxa by testing several variables and see which taxa are significant at different times, as in this case:
But I get the files to display, as if they were analysed 1 single metadata at a time. i need the aggregated result.
So I tried this with *:
qiime composition ancombc
--i-table table_16S.qza
--m-metadata-file Metadata.txt
--p-formula ‘A* B * TIME’
--p-reference-levels TIME::T0
--o-differentials data.qza
But it gives me the error --> An error was encountered while running ANCOM-BC in R (return code 1), please inspect stdout and stderr to learn more. Parsing the error, it tells me --> Error: No residual degrees of freedom! The model is over-parameterized
A more parsimonious model is needed.
I hope this was clear and helpful.
Thanks in advance.
Thank you for your kind reply.
I have 3 different types of food matrices, analysed together. in total I have 60 samples (3 matrices x 20 each). I did 2 ancom-bc tests, 1 on the total 60 samples and 1 test on the samples analysed separately (20 samples). I got the same error with both and had originally done this test to see if the problem was having them analysed separately.
The metadata are common to all the samples. Two treatments were carried out on them.
The metadata columns are as follows: SampleID, A (first treatment, categoriacal metadata), B (second treatment, categorical metadata), Time (times 0, 7, 18, 21, 27, numerica metadata).
