Indeed, I made sure to retain a maximum of samples, I chose a sampling depth of 24985 and Retained 124,925 (72.37%) features in 5 (100.00%) samples at the specifed sampling depth. So I think I don’t loose too many information, right ?
This seems perfectly reasonable and is more than likely not the cause of the issue!
There isn't another file to check, however, I am pretty sure that the issue lies in your feature IDs and not your sample IDs. So you might have checked the wrong column. I recommend following Matt's steps here again but checking the feature IDs.
If that does not work could you post your code and results(for Matt's steps) so that I have a better idea of what is happening. Also if you want to post your rooted tree and table that you are using I could also try to debug more effectively.
Thank you for your answer, I think I figured out what’s going on. I ran my qiime2 pipeline on a several .fna files. The one I looked was actually composed of 5 samples (ATB_1, _2, _3, _4 and _5). The problem is that for each sample, there exist multiple reads with an identifier as ATB_1_1502, ATB_1_1503, etc. In my table.qza, the samples only go from ATB_1 to _5 and in the rooted tree for each sample there exist multiple features, like one ATB_1_1502 and another one for ATB_1_1503. So I think, the rooted tree took all identifiers as unique and didn’t merge them as in the table.qza. I don’t know if it’s clear ?
Do you know a trick to force the rooted tree to merge all ATB_1 with each other and so on ?
Awesome! I am glad we were able to find the source of the problem
The next step is making sure that the rep-seqs.qza (what you use to make the rooted tree) have the same feature IDs. I believe that you will need to filter the repseqs.qza file so that it matches your current table using this command: