How to identify the feature ID in Gneiss

Hi, we use Gneiss to analyze the species which showed significant abundances between groups. However, most of them are uncultured bacteria. I was just wondering if we can get the feature IDs for them to figure out who they are.

Thank you so much for the help!

Junfei

Hi @junfei_guo, did you have any luck extracting IDs from balance-taxonomy? There is an example of how to run this command in the last step in the tutorial here.

Thanks for the reply!
Yes I have downloaded IDs from balance-taxonomy. But how can I figure out which one is the one showed in proportion plot since most them are D_6_uncultured bacterium?

Junfei

Did you have any luck downloading the csv files at the bottom? They should give the exact sequence ids for features that you are interested in.

Yes, I downloaded the csv files at the bottom. It contains thousands of feature IDs and it just showed 10 species in proportion plot, so we got confused how to specify the species which under the name of uncultured bacterium. I have uploaded my qzv file, could you please check it for me? Maybe I didn't explain my questions well.

Thank you so much!

Junfeiy0_RF_taxa_summary-BH-Con-l6.qzv (323.6 KB)

Hi @junfei_guo, that sounds about right – sometimes balances can contain hundreds of features. It is difficult to make sense of a proportion plot with hundreds of features - so we only present 10 features from the numerator, and 10 from the denominator. Looking at the files you posted, it looks like there are quite a few features under column 6 that are uncultured. There are 14 taxonomic fields which is a little weird, but likely not problematic, since the columns after column 7 appear to be redundant.

Thank you for your patience!
Does it possible to find the uncultured features if we can compare those feature IDs to mcrobiome database?Or that is the final information we can get from the balance taxonomy?

Junfei

Yup. You should be able to match those Feature IDs back to the output of qiime feature-table tabulate-seqs (see here). That should give you more information about the exact sequence and possible strains / genomes that this sequence is found in.

Like in the csv file I sent before, there are two D_5_prevotella 1 in numerator proportion plot figure. I still don’t know which one belongs to which feature ID.
Also, are the 10 features showed in proportion plot significantly difference in two treatment groups?

Thank you!

Junfei

Hi @junfei_guo, the proportion plots don’t currently have feature ids. It sounds like you may want to a little exploration on your data - I’d consider exporting the FeatureTable, and investigating the individual taxa (see the exporting data tutorial )

Concerning significance tests, gneiss will test for significant balances, which will boil down to a pair of groups. It won’t necessary give you information about individual taxa. It may be worthwhile to sanity check against ANCOM to see if those uncultured microbes are also significant there.

Oh I see. Will try ANCOM again.
Thank you so much for the help!

Junfei

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