todos.md

Todos

final checklist

• index / description
• outline
• day1
• day2
• day3
• day4
• summary / template

day 1

• RColorBrewer example
• recommendation windows -> workbench
• no git -> download repo zip as alternative.

day 2

• phrasing of question,
• write_tsv instead of write_table
• DESeq workflow -> a,b,d,e
• before writing table, include resultsNames to get the coefficients printed out
• possibly skip relevel() --> extensive contrast discussion

day 3

• celltype + condition + celltype:condition --> fix manual explanation of beta (JJ<->BM)
• use "real" sample names from data, rather than sample1,2,...
• after break (heatmap of significant genes): - subset(padj < 0.05) is optional - simplify filter (%in%) - simplify heatmap (keep only scale = "row", everything else is optional)
• skip results() in favour of lfcShrink()
• mod_mat trick: - move up? - have a task/poll to create complex contrast: BM_DMSO vs JJ_TG
• GENEID --> SYMBOL task too challenging.
  • needs more time or hints
  • columns selection: dplyr::select(8,9,10,14,15,16) --> select by pattern ?
• reduce to only one prior gene set apoptosis or spliceosome

general todo

• continuous variable in design ?
• give example with own sizefactors (e.g. from spike-ins)?
• enforce consistency in sample naming (DMSO or CTRL)