ms_workflow: Improvemnts
-
separate protein groups -
report 0-proportions similar to NAs -
expose renaming scheme as an argument -
protein_acc
extraction depends on study (with/witnout name, w/o separator). Need to generifyprotein_acc=str_split_fixed(protein_ids, "[|]", 3)
. One way:--extract extrac_acc.R
which defines extractor function ->protein_acc=extract_fun(protein_ids)
-
detect presence/absence of identifcation types --> conditionaed executino of corresponding bits -
try to postpone annotation handling to end of data-prep workflow/analysis -
fix result table links #' [identSummary]( r add_prefix("ident_types_summary.txt")
) -
How to auto-adjust imputation proportion? -
We just need imputation because we want an abundance matrix for limma. for t-tests neither na->0 nor impuations are required because we can work with long data.
limma
- also show batch-corrected qc plots (clustering, pca)
- man muss das mal alles durchschauen
- why does the voom with a
~condition
design fix the sample clustering infile:///Volumes/project-stepien/stepien_ms_fractions/data/limma/p_vs_s/dge_limma.html
- add condition/sample colors to voom vs raw plot
- externalize annotation of results
- why do ma plots look so weired?