ANCOM-BC analyis: comparing 3+ groups

I recently finished analyzing ANCOM-BC in qiime2 and was able to handle the basic analysis, but now I need to compare more than three groups, and I'm not sure which method is most appropriate.

Could anyone kindly guide me on how to do this or what can change in scripts I'd really appreciate your help or any resources you can share.

If see attachment of metadata.tsv, I want compare between gut_Treat X gut_ctr; also fecal_treat X Fecal_ctrl in the column “Treat_exp“.

qiime feature-table filter-samples --i-table table.qza --m-metadata-file ../sample-metadataa.tsv --p-where '[Treat_exp]="Fecal_ctr"' --o-filtered-table gut-table.qza

qiime composition ancombc --i-table ../dadaOutput_70_60/table.qza --m-metadata-file ../sample-metadataa.tsv --p-formula Treat_exp --o-differentials l6-ancombc-subject.qza

sample-metadataa.tsv (1.2 KB)

Thank you in advance!

Hello!

It appears that you need to split (qiime feature-table split) your feature table based on the sample type column to obtain “gut” (only gut samples) and “fecal” (only fecal) tables. Then, run each table separately and provide a reference (control) group for each table.

Best,

Thank you so much for quick reply.

I run this codes and it worked but I don’t know how do it for rep-seqs.qza?

qiime feature-table filter-samples --i-table ../dadaOutput/table.qza --m-metadata-file ../sample-metadataa.tsv --p-where "[Tissue.type]='Gut'" --o-filtered-table table_Gut.qza

Hello again.

  • You can use the same rep-seqs.qza file for both tables since it contains sequences from both
  • If needed, you can use this command (change file names) to filter it:

qiime feature-table filter-seqs
--i-data all_rep-seqs.qza
--i-table gut_table.qza
--o-filtered-data gut_rep-seqs.qza

Thanks . Kindly can check to me dada2 QC output, it looks something wrong .

Hi,

Please, always create a new topic if your questions deviates too much from the original one (from ancombc to dada2!).

I can’t see any issues with that based on the screenshot you provided. If you still have questions:

  • Create a new topic if it is not related to the original post
  • Provide more details regarding the issue you are facing

Best,

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.