Samples merged during cluster-features-de-novo

[Zoe_Buffington]

Hello,

I have run into issues in multiple areas that appear to be related, so I am trying to find (and hopefully fix) the source. Most noticeably, I have had trouble with analysis using phylogenetic alpha and beta diversities on one of my data sets. I get the error message below:

qiime diversity beta-phylogenetic
--i-table /homes/zbuffing/bats/communities/fastqout/id-filtered-table2.qza
--i-phylogeny /homes/zbuffing/bats/communities/fastqout/mafft-fasttree-output/rooted_tree.qza
--p-metric 'generalized_unifrac'
--o-distance-matrix /homes/zbuffing/bats/communities/fastqout/div/unifrac-dist-matrix-com.qza

Plugin error from diversity:
The table does not appear to be completely represented by the phylogeny.

I had no issues in previous data filtering and sequencing steps, but came across a potential problem in the clustered-table output from qiime vsearch cluster-features-de-novo. It appears that some of the samples have merged (there should be 21 unique samples, but there are only 19).

qiime vsearch cluster-features-de-novo
--i-table /homes/zbuffing/bats/communities/fastqout/table.qza
--i-sequences /homes/zbuffing/bats/communities/fastqout/rep-seqs.qza
--p-perc-identity 0.97
--o-clustered-table /homes/zbuffing/bats/communities/fastqout/table-dn-97.qza
--o-clustered-sequences /homes/zbuffing/bats/communities/fastqout/rep-seqs-dn-97.qza

https://view.qiime2.org/visualization/?src=38a085ac-da61-41d6-9284-b2df0ed4e97d&type=html

(From my metadata: samples 16 and 18, and samples 17 and 19 appear to have merged during de-novo)

The samples were all uniquely identified in the manifest when I imported them, but they contain the same information in my metadata. I have not provided qiime with the metadata as an argument in any previous steps, so I was wondering why it might be combining these, and how I can go about fixing it, as I believe it will solve some other issues I am having.

I am an undergraduate researcher who is very new to qiime and bioinformatics in general, so I apologize for anything that might have been lost in translation! If there is any more information I can provide, please let me know!

Thank you!

[jwdebelius (Justine Debelius)]

Hi @Zoe_Buffington,

It looks like you're having two errors,

@Zoe_Buffington

Plugin error from diversity:
The table does not appear to be completely represented by the phylogeny.

This notification tells you that there's a disconnect between your tree and your filtered table. Specifically, there are more features in your filtered table than in your rooted tree. So, you may want to figure out why/where you lost features and/or filter your table by excluded tips.

@Zoe_Buffington

I had no issues in previous data filtering and sequencing steps, but came across a potential problem in the clustered-table output from qiime vsearch cluster-features-de-novo. It appears that some of the samples have merged (there should be 21 unique samples, but there are only 19).

...

It looks like you're displaying your results by sample date, since you're showing two samples on those dates. If you change your metadata category on the interactive sample drop down, this will shift the count.
(I can't see the q2-view file you linked; it either needs to be hosted online or we need the original file)

Best,
Justine