Update ref based rna-seq

[lldelisle]

This PR in based on #3272 Additional changes in this PR:

• Use collection instead of individual datasets
• Use last up to date version of tools (except STAR which I could do).
• Add myself as contributor
• Update some phrasing: changing 'absolute fold-change > 2' to 'fold-change > 2 or < 0.5' etc.
• Change color scale of log2(counts)
• Change some images / text on strandness
• Add a step to check Pasilla gene status
• Propose an alternative with counts with STAR
• Use a choose your own adventure (CYOA) style thanks to @hexylena
• Use last version of heatmap2 and plot z-score without computing them. Moved the part on how to compute z-scores in a detailed section and do it on all genes.
• Use multiQC on cutadapt thanks to @bebatut Aside:
• I updated create_new file to set name at creation.
• @hexylena added a post on CYOA and updated 'Create new tutorial'.
• @bebatut updated copy datasets between histories

[lldelisle]

Questions I have in mind:

• Shouldn't the Z-score computation be part of heatmap2 tool.
• Why do we plot the under-represented KEGG pathway?

[lldelisle]

Ready for review @hrhotz @bebatut would you mind to check that the changes on strandness are ok for you (if you have time I am also interested in your review for the rest but this is the biggest change I made on what you wrote)?

[lldelisle]

Thanks @nsoranzo

[lldelisle]

I am still working on this PR (just the heatmap part because I just updated heatmap2 tool so it can compute the z-score).

[lldelisle]

Heatmap2 changes done. Waiting for last reviews before changing the workflow.

[bebatut]

Thanks a lot @lldelisle for this work. One last things, could you check if the tests for the workflows are still valid?

[lldelisle]

One last comment. Could you add the level to the metadata? Thanks a lot

Which one would you put? Introductory?

[gallardoalba]

One last comment. Could you add the level to the metadata? Thanks a lot

Which one would you put? Introductory?

I would say intermediate.

[bebatut]

I would have said Introductory for the level. We usually use it at the introduction to RNA-Seq analysis

[lldelisle]

I still need to update the workflow for DEG but I am a bit fighting...

[lldelisle]

Also what should I do with the tour? How can I test it?

[gallardoalba]

So I guess it can be merged.

[lldelisle]

Please wait that I update the workflow with tags.

[gallardoalba]

Please wait that I update the workflow with tags.

Ah, ok sorry!

[lldelisle]

What do we do with the tour? I update it?

[lldelisle]

@shiltemann @hexylena is it possible to hide the tour so I have time to update it after my course?

[bebatut]

I think no worries about the tour. I am quite sure it is a lot out of date. We should probably remove it

[lldelisle]

Then let's merge like this? Or I remove the tour?

[lldelisle]

Hi, My last changes are:

• Add a sentence about the discrepancies between the results of the first part and the count files on zenodo (this may disappear if we update zenodo).
• Give some hints on how to run the single-read datasets. In my opinion, the single reads datasets should be run with the option in STAR Length of the genomic sequence around annotated junctions to 74 as we explain above that it should be read length - 1. However, the counts in zenodo corresponds to STAR ran with 36 like in PE datasets. What do you think? Thanks

[lldelisle]

Hi there, I am back from holidays and our class is on Wednesday.

1. Should I remove the tour or I keep like this?
2. Should I remove the DEXseq part which is commented but out of date?

Thanks

[hexylena]

Let's just remove the tour. I think ok to leave commented out portions.

[lldelisle]

Thanks @hexylena Ready for review / merge

[bebatut]

The updated tutorial is online :tada: I just noticed sth. If nothing is selected for CYOA, then both content is displayed

[lldelisle]

@hexylena is there a way to select by default one of them?

[hexylena]

Yes, that is expected. I'm not sure there's a good solution for that, picking one by default seems sub-optimal, and having both displayed at least the user can see both simultaneously which feels useful? (Espeially when it's e.g. two separate hands-on boxes). Less so when it's really inline stuff.

[hexylena]

We could write some JS to select one (randomly or set a default.)

[lldelisle]

I think we can set a default (at least for this tutorial).

[hexylena]

What should the default be?

[lldelisle]

I prefer STAR but @pavanvidem said in the discussion that FeatureCount was more flexible for non standard gtf... So I would say FeatureCounts, also the end of the tutorial is based on these counts.

[hrhotz]

@bebatut @gallardoalba @hexylena @nsoranzo @jennaj Thank You very much for all your work and help getting @lldelisle PR into the GTN

[lldelisle]

Yes, thank you very much, I am really happy that we have an up to date tutorial.

[bebatut]

We "just" need now to record a new video for this tutorial :sweat_smile:

[hexylena]

two videos! :sob:

[lldelisle]

Let see if I have time to do this.