bokulich-lab/q2-protein-pca

QIIME2 plugin for PCA analysis of protein sequence phylogeny.

q2-protein-pca's Issues

add distance column to the table

To make it more obvious that the sorting is done based on euclidean distance, show it in an additional column.

add protein structure visualiser

It would be great to have a visualisation showing the conserved positions on a protein structure of choice. Maybe this could work.

add "most variable" colour

Similarly to the conservation threshold, one could add another slider to colour the most variable position within a sequence (should anyone be interested in finding which position are the ones that influence separation of sequences most)

add actual validation to the new types

consider other dimensionality reduction methods

q2 protein-pca plot-loadings

Following the tutorial, I am not able to execute the last command;
q2 protein-pca plot-loadings --i-pca-loadings thioredoxin-pca-loadings.qza --i-positions-mapping thioredoxin-mapped.qza --o-visualization thioredoxin-pca-loadings.qzv

is this command q2-protein-pca? What I am missing?

thanks

make the loadings visualisation zoomable

hide gap rows from the table

Similarly to the points in the plot, when the box Hide missing positions is selected the gaps should be hidden from the table.

add snakefile for an example workflow

Just so that once can easily run the entire analysis pipeline with one press of a button.

add git hook for flake8

Run flake8 on every commit.

add actual position to the point tooltip

The tooltip displayed when hovering over a point on the plot should (additionally) show the actual position of that point within the sequence selected via the drop-down.

review/add tests

Testing is very limited - review and add more unit tests.

JSON error when calling `plot-loadings`

For details of the bug see the original QIIME 2 forum post: https://forum.qiime2.org/t/error-while-qiime-protein-pca-settingwithcopywarning/23544.

It seems that when an alignment produced no gaps and the positions matrix has no NaNs, the type used by pandas to represent those values is not JSON-dumpable (it's pandas' Int64).

review the ranking code

This is super-slow now. There must be a way to make it run faster/parallelise.

update README

• add more context
• add examples
• make it look presentable