Imported FeatureTable[Frequency] index error

[jcthomas (JT)]

Hello. I've imported a Biom file as a FeatureTable[Frequency], but when I try to do anything with it I get a Pandas index error. For example:

qiime feature-table summarize --i-table wanda.featuretable.qza --o-visualization wanda.featuretable.qzv --m-sample-metadata-file sample_names.tsv

Yields:
Plugin error from feature-table:

"None of [Index(['s2-04F', 's2-03C', ],\n dtype='object')] are in the [index]"

The metadata file is just sample names, I've made the files available here:

qiime

Not sure where to start in figuring this out

[Mehrbod_Estaki]

Just a quick guess here. The error seems to come from these samples

@jcthomas

‘s2-04F’, ‘s2-03C’

and in your metadata file these samples appear with an underscore instead of a dash i.e. s2_04F. I'm wondering if this may the issue. Try renaming those 2 samples and see if that helps while you wait for the experts to give a more knowledge-based suggestions

[thermokarst]

Hi @jcthomas! @Mehrbod_Estaki hit the nail on the head --- the sample IDs in your metadata file don't match the sample IDs saved in the feature table (although, it isn't just the two samples @Mehrbod_Estaki listed, it looks like all of the samples are mismatched). Some of them are underscore/dash swaps, but other samples are missing wholesale. One way you can interrogate your feature table is to rerun the feature-table summarize command above, omitting the metadata:

$ qiime feature-table summarize \
  --i-table wanda.featuretable.qza \
  --o-visualization wanda.featuretable.sansmetadata.qzv

You will see something like this:
check-out-those-ids.qzv (313.9 KB)

The easiest option for you is to update your metadata file to match the sample names in your feature table, otherwise, you would need to rename your samples in your feature table. Let us know if you get stuck!

[jcthomas (JT)]

Whoops, yeah that was it. Cheers!