Mapping to promoters for all genes on array

[katehoffshutta]

get_gene_level_methylation.r relies on the NDC function probeToMeanPromoterMethylation, which takes in a list of input genes of interest. This input list must be one gene per line. The all_genes list found here: https://github.com/QuackenbushLab/tcga-data-nf/blob/main/bin/r/get_gene_level_methylation.r/#L51 will include some lines with semicolons and multiple genes, and probes will not map correctly to these genes.

To solve this, all_genes needs to be split to long form - one gene per line - prior to being reduced to a unique gene list.

I believe this does not affect the current examples in the manuscript, which rely on the input TF list instead.

We should also add some NDC documentation clarifying that the input gene list has to have the form of one gene per line. I have added this to an outstanding issue in the NDC repo.