Plot sequence logos

[krish8484]

Description

Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context. List any dependencies that are required for this change. This PR aims to plot the sequence logos using python's logomaker which may be used along with the heatmaps in the future. Fixes #23

Type of change

Please delete options that are not relevant.

• [x] New feature (non-breaking change which adds functionality)

Checklist:

• [x] My code follows the style guidelines of this project
• [x] I have performed a self-review of my own code
• [x] I have commented my code, particularly in hard-to-understand areas
• [x] My changes generate no new warnings
• [x] I have added tests that prove my fix is effective or that my feature works
• [x] New and existing unit tests pass locally with my changes
• [x] I have not reduced the existing code coverage

[krish8484]

Are the sequence logos as expected? They are in output/sequence_logos.

[AngryMaciek]

Are the sequence logos as expected? They are in output/sequence_logos.

Dear Krish,

There is no directory called output anywhere I see, especially not uner the root of the repository... However, I have found some sample plots for HNRNPF undet unit test directory for this script so I will refer to these. It's OK but I think of 3 changes:

• change T to U since it is a RNA sequence, not DNA. This is important.
• is the y-axis showing information content? I would exect that the library automatically plots information content but then the y-axis would range from 0 to 2. As it is not the case - could you please find out what is the scale and what does it mean? This is important.
• Could you remove the top and right-side lines from the "box"? Just for aesthetics, not that important.

[krish8484]

Hi Maciek, The directory output/sequence_logos will be in tests/integration after you run the snakefile, though the png files in unit tests depict a similar output.

• I am not sure which T are you referring in the first point. If it is the one in motif_HNRNPF_824.png, then that T comes from the input file. See motif_HNRNPF_824 in the same folder, the value of T for 1st position is 97.115.
• The y axis represents the probability specified in the probability matrix in the input files (in unit tests folder). Hence the range 0 to 1.
• I could not find a function in the documentation which would remove the top and the right lines for us.

[AngryMaciek]

• Yes, by convention we operate now on DNA sequence always, but for the plot please change the alphabet for U.
• Is there a flag in this library that would allow to plot information content instead of raw probabilities?
• pyplot has these options, just google: https://stackoverflow.com/questions/9750699/how-to-display-only-a-left-and-bottom-box-border-in-matplotlib I see that use are using plt to show so this solution should wotrk.

[AngryMaciek]

If logomaker does not support plotting information content out-of-the-box, could you please calculate the contents youself? According to these specifications: https://en.wikipedia.org/wiki/Position_weight_matrix#Information_content We do not need to deal with background probabilities, the base of log is 2.

[AngryMaciek]

As I was goung through the Files changed section here on GitHub and reviewing the code I noticed that under the dedicated directory for unit-testing this script you have also added the png sample output files to the repository (ex: tests/unit/plot_sequence_logos/motif_HNRNPF_820.png). Could you please remind me - do we need them in the repository? It just flew through my head that maybe it is not necessary to include them, since every unit test will generate them once again, right?

If that is correct - please remove the sample output files from unit testing of this script as well as previous unit tests: combine_results, Plot-heatmap-for-motifs,

[krish8484]

I have fixed the indentation in the Snakefile at the places you mentioned, removed the top and the right axes from the graph, changed the T to U in the input file in the unit tests, and changed the script to display information instead of probability. Deleted the superfluous output files from unit tests and updated the md5sums of the output files.

[AngryMaciek]

Before I merge it please make sure that you have complied with all the items from the Checklist: in the initial post of the PR.

[AngryMaciek]

@krish8484 I have cleaned the code a little, indents (🙃), removed unnecessary expanions and importantly added two more sections to all your rules (which I forgot about earlier):

• benchmark - it allows snakemake to report the execution time per rule into a textfile
• log - to capture stdout and stderr streams of each script's execution into another two text files; very useful for debugging.