This PR aims to fix some issues I encountered when using the pipeline:
The logic in rules/common.smk::is_single_end was somewhat backwards
fgsea:
fgsea has an option to collapse enriched pathways, such that they become independent; this now also gets plotted
fgsea now only plots (at most) 1000 significant (w.r.t. fdr) pathways
pathways may have weird names, so resulting filenames might not be valid, i.e. saving to file failed in some cases; to fix that, replace unwanted characters with underscore
the diffexp-heatmap plot now uses gene names instead of transcript names (as suggested in the TODO)
do not load sleuth object twice in plot-pca.R
spia: if there are no significant genes, make an empty plot instead of failing
common.smk:
allow # comments in samples.tsv + units.tsv
remove singular columns from samples dataframe (since every column is used in the analysis scripts at some point, and singular ones just make everything fail)
This PR aims to fix some issues I encountered when using the pipeline:
rules/common.smk::is_single_endwas somewhat backwardsplot-pca.R#comments in samples.tsv + units.tsv