Sorry to follow up a question. I just want to make sure that I am on the right way. Thanks in advance.
I use the code as below,
qiime tools export
taxonomy.qza
–output-dir exported-taxonomy
qiime tools export
table.qza
–output-dir exported-feature-table
Here I got the taxonomy.tsv and feature-table.biom
Then I use the code, biom convert -i feature-table.biom -o feature_table.txt --to-tsv
Then I import the taxonomy.tsv and feature_table.txt into R, and merge them based on the featureID, which produces a final dataframe.
Then I can regard this table as the variants table(ie. the 100%OTU table) of my data? Also like the ZOTU in USEARCH/UNOISE3.
However, the later bar-plot(produced by code ‘qiime taxa barplot --i-table table.qza --i-taxonomy taxonomy.qza --m-metadata-file Mapping_3rd.txt --o-visualization taxa-bar-plots.qzv’) is just the summary file from different taxonomy perspective.
Hi @Lu_Yang,
I am not really sure what you are trying to do or what your question is. Merging feature tables can be performed using the commands that you already posted in your initial question in this thread. Exporting from QIIME and merging in R is absolutely unnecessary. The feature table in QIIME 2 already IS the "100% OTU table", i.e., the output of dada2 that has been denoised and dereplicated.
Yes, this is a barplot of taxa, not of individual features. It looks like you haven't followed any of the links I provided above regarding alternative visualizations. In particular, heatmap may provide the analysis that you are looking for.
Hi, @Nicholas_Bokulich,
Ahh, so sorry for not clearly stated my question.
My purpose is got the feature table(100% OTUs) and its corresponding taxonomy. I have tried the heatmap. I used the code ‘qiime feature-table heatmap --i-table table.qza --o-visualization heatmap’. But I got the error 'Plugin error from feature-table: Image size of 259406x4491 pixels is too large. It must be less than 2^16 in each direction. ’ So I thought it may be caused by the sample I provided is too much.
As I check the feature table(table.qza) and the taxonomy table(taxonomy.qza), they are the 100% OTUs and every feature ID’s corresponding taxonomy. So that is the reason I want to merge them in R to got one table with 100% OTUs and taxonomy.
Hope this time I stated clearly.
Thanks so much.
Hi @Lu_Yang,
Thank you for clarifying. I think I understand now – and no, QIIME 2 does not have a method for generating a QIIME1-style feature table (e.g., annotating a feature table with taxonomy IDs in addition to the feature ID). You are performing the correct steps to export and import to R if you are trying to, e.g., plot the abundance of each feature but label by taxonomic assignment. Is that what you are attempting to get?
Do you want to report this as a separate forum post? We can address this issue separately (and could be a support issue that we want to address for other users). It sounds like you probably have too many features to plot — 2^16 is a very large number of features (as dada2 output, it makes more sense for OTU picking without denoising), so you may want to check this number with feature-table summarize and review the distribution of features per sample. If this is expected (e.g., you are combining multiple HiSeq runs), then you may want to filter out low-abundance features with feature-table filter-features prior to generating a heatmap anyway.
Hi, @Nicholas_Bokulich,
Yes, that is what I want. And I will make another post about that heatmap error .
Thanks so much for your help along the way!
Best.
Hi @Lu_Yang,
No need to issue a new forum post — the previous thread has already been split to generate a separate thread relating to the heatmap size issue. Thank you!
Hello,
I received the same error:
qiime feature-table heatmap --i-table feature_allReads_rdp-table.qza --output-dir heatmap
Plugin error from feature-table:
Image size of 82935x1280 pixels is too large. It must be less than 2^16 in each direction.
Debug info has been saved to /home/cristina/tmp/qiime2-q2cli-err-k6loktr0.log
Hi @cesteban - this hasn’t been addressed yet, but it is worth noting, this error has to do with having a lot of features — personally I think the utility of this visualization decreases with too many features (or samples, for that matter). One solution to get you moving right now is to filter your feature table to a subset of features worth looking at that are relevant to your study. Another option is to collapse your table to some some taxonomic level. Keep us posted!
