Gneiss Issues With Heatmap

gneiss
(Kristina) #1

My Gneiss analysis will not shoot out a full heat map and a readable dendrogram. Is there something I can do to adjust my data. I know there are compositional differences as the taxa-bar-plots figure is drastically different between the CON and TRT groups, I just can’t get the heat map to shoot anything out. Here is my code and I have no error messages.

I am using qiime2-2018.8

$ qiime gneiss correlation-clustering --i-table finaltable.qza --p-pseudocount 0.5 --o-clustering hierarchy.qza

$ qiime gneiss assign-ids --i-input-table finaltable.qza --i-input-tree rooted-tree.qza --o-output-table labeledtable.qza --o-output-tree labeledhierarchy.qza

$ qiime gneiss ilr-hierarchical --i-table finaltable.qza --i-tree labeledhierarchy.qza --o-balances balances.qza

$ qiime gneiss ols-regression --p-formula “Inoculum+Feed” --i-table balances.qza --i-tree labeledhierarchy.qza --m-metadata-file DVSMetadata.txt --o-visualization regression_summary.qzv

$ qiime gneiss dendrogram-heatmap --i-table finaltable.qza --i-tree labeledhierarchy.qza --m-metadata-file DVSMetadata.txt --m-metadata-column Feed --p-color-map seismic --o-visualization heatmapfeed.qzv

Saved Visualization to: heatmapfeed.qzv

(Nicholas Bokulich) #2

Please share the heatmap QZV here. Sounds like this is either a browser issue or a memory issue (the heatmap is too large to load in your browser). Maybe try a different browser?

(Kristina) #3

Good Day Nicholas!

I figured out that I should sort out the features with less than 50 per feature and then probably collapse the features first. I think everything was too much, so I sorted and collapsed to L5 and was able to finally get good heat maps.

So, if anyone has this issue, sort out the low frequency features and collapse the tables.

Thank you for your response, but I think I have it =)