Incorporate "read noise" as a quantification parameter
lweasel opened this issue · comments
Investigation of real RNA-seq data alongside MAQC TaqMan qPCR measurements shows that reads originating from introns and (to a lesser extent?) intergenic regions are a significant confounding factor in transcript quantification.
One way I can see of incorporating some kind of simulation of this into piquant is to add a "noise read depth percentage" quantification parameter, and specify a "noise reads" GTF file (e.g. that consists of "transcripts" consisting of a single exon spanning the whole of each original transcript, i.e. including the intronic regions). The noise read depth percentage controls the depth of reads simulated from this second GTF file, as a fraction of the read depth of the main transcript GTF file.
Last comment referenced wrong issue.
Fixed, needs documenting before merging to master.