Hello,
I used q2-amplicon-2024.2 in my current project and observed that the sample-ids in feature tables do not match the sample-id values of the metadata file compared to previous q2 releases. As examples, I added her two feature tables created by a filter-features command.
Using q2-2022.11 release:
qiime feature-table filter-features --i-table table.qza --m-metadata-file representative_sequences.qza --o-filtered-table table-filtered-14.qza --p-min-frequency 14
Using q2-amplicon-2024.2 release:
qiime feature-table filter-features --i-table table.qza --m-metadata-file representative_sequences.qza --o-filtered-table table-filtered-14-new.qza --p-min-frequency 14
Then visualized by q2 feature-table summarize:
table-filtered-14.qzv (785.8 KB)
table-filtered-14-new.qzv (748.9 KB)
Although the default sorting in the Interactive Sample Detail tab is different, the absolute numbers of reads per sample seem to be identical (assuming that the increasing numbers correspond to the sample ids in the old version). Downstream steps such as taxa barplots do use sample IDs as defined in the metadata file.
As far as I could test so far, this behaviour shows up when using q2-amplicon-2024.2. Here, the table created by dada2 and then visualized by feature-table summarize showed this sample ID formats, not only a subsequent filtering step.
I have tested this behaviour using tables created by dada2 of previous q2 releases (the sample ID is used as defined in the metadata file) and then executing a filtering step with q2-amplicon-2024.2 (the sample ID is now an increasing number).
I would very much appreciate if somebody from the developers could have a look at this.
Best,
