search

QuackenbushLab / NetworkDataCompanion

An R library of utilities for performing analyses on TCGA and GTEx data using the Network Zoo
GNU General Public License v3.0
4 stars 0 forks source link

Implement flexible slicing in methylation probe mapping #14

Open katehoffshutta opened 1 year ago

katehoffshutta commented 1 year ago

In mapProbesToGenes, the options for slicing are rangeUp and rangeDown, indicating how far up and down from the TSS to search. Both have to be non-negative numbers. It would be good to fix this so a user can slice regions that don't include the TSS (e.g., -1500 to -200bp).

soelmicheletti commented 1 year ago

Before I implement this I have a question.

In the manifest there is "probe_strand"="+" (and possibly, I guess, "probe_strand"="-"). However, we don't look at it in our implementation.

I remember that Jonas explained me that based on strand, the genome is read in a certain direction. Shouldn't we switch rangeUp and rangeDown based on this?

katehoffshutta commented 1 year ago

Good catch, Soel. So, rangeUp and rangeDown interact with the distToTSS column in the manifest, which I believe Jonas checked is calculated based on strand, so -200 means 200 base pairs upstream, whether on the + or - strand. However, it would be great to spot check this using the genome browser on a few rows of the manifest to make sure we are correct.