search

dputhier / pygtftk

A python package and a set of shell commands to handle GTF files
GNU General Public License v3.0
45 stars 6 forks source link

Implement strategy for N-fold bed samples comparison #80

Closed guillaumecharbonnier closed 4 years ago

guillaumecharbonnier commented 5 years ago

If a comma-separated list of bed files is provided as --peak-file, ologram should run the analysis for each bed then provide a heatmap instead of barplot like I do in the "greatr" package. I would like to work on this once my PhD manuscript is in good shape if nobody is eager to develop that soon.

qferre commented 5 years ago

This would be pretty straightforward. We already provide the tsv file, so it's just a matter of writing the plotnine code based on the file.

Of course, you do not have to read the output file ! The pandas dataframe is accessible in Python code. We can talk about it in person if you are interested in implementing this.

dputhier commented 5 years ago

The comma separated is supposed to allow N-fold comparisons. As a workaround, I have implemened merge_ologram_stats that simply takes a set of OLOGRAM tsv output files to build a heatmap. Available in 6541a23cd72bc4f30c3b36ba96abd633788f6902

qferre commented 5 years ago

I think this is a good plugin, and I think it is also a good idea to separate those two usages, especially since I am developing a multiple-overlap solution that will also use this type of syntax. It could get confusing for the user if we merged the two.

I think this workaround (the merge_ologram_stats plugin) should be the permanent solution, with an explanation in the documentation.

qferre commented 5 years ago

I have first draft of a plugin to do an analysis on an entire set of bed files at once (multiple overlaps) and find interesting combinations. We discussed this at length in-person. More detailed explanations will be found in the documentation.

I will push this to the develop branch in the upcoming days. If you've got time to discuss it and/or try it, I would welcome the feedback :)

dputhier commented 4 years ago

This is implemented in 1.2.7