Minimal functionality: Input checking and basic analysis

[frhl]

We need the following for minimal functionality:

Input checking: Input format checking for four different types of input.

We should have a function that checks format. E.g. separate replicates, we call a moderated t-test. We should get a list of 1) input data.frame, 2) and a boolean vector indicating/string indicating what further functions should be called to process the data. In summary, read input, check input, map input, moderated t.test and identify enriched interactors.

Include, verbose error message that we can use for debugging and potentially for users in the future.

R error handling: stopifnot('FDR' %nin% colnames(df))

Function details

1. Reading. Input: filename/path. output: list(data.frame, booleans: needmap, boolean: needtest)
2. Read input. Input: input data.frame. output: list(data.frame in which first column contains mapped genes, data.frame: accession number to gene mapping).
3. moderated t.test: input: data.frame. output: data.frame with three collumns attached, LogFC, FDR, P-value)
4. identify enriched interactions. Input: df, logFc cutoff, logfc_direction, pvalue (NULL), fdr. output: list( (3) but with attach boolean significant column, data.frame with interactors and non-interactors that can be downloaded by the user)

[yuhanhsu]

DONE on March 12: (1) check_input (2) read_input

TO DO on March 12: (3) map_geneid (FHL) (4) calc_mod_ttest (YHH) (5) id_enriched_proteins (YHH)

[yuhanhsu]

YHH TO DO (Mar 13):

• [x] modify map_gene_id function. Input: df containing accession_number. Output: list of (df with access_number col replaced by gene col, df with accession_number -> gene mapping)
• [x] modify id_enriched_proteins function. Input: df containing logFC, pvalue, FDR, + significance cutoff parameters. Output: df + boolean significant column.

[yuhanhsu]

TO DISCUSS:

• map_gene_id function: instead of returning "NA" for accession_number with isoforms, could remove the isoform ("-X") suffix before mapping
• related to above but also in general, do we want to allow duplicate gene names in the input df throughout our pipeline?

[yuhanhsu]

• [x] modify map_gene_id function. Check for "-" and "." in acccession_number, remove subsequent substring before mapping, save original isoform accession in output mapping results.

*** maybe in the future, show isoform information in plots

Allow duplicate gene names, only check for unique names if doing overlap enrichment type tests.