I don't think this is possible on a global scale.
In case I understood your question right, that you would like to finde motifs which not completly turn off transcription but reduce transcription of certain transcripts, than I would say, this is not possible at the moment.
When you sequence RNA, you get more reads from a transcript, when you have more RNA fragments of that transcript to begin with. The amount of transcripts correlates with the transcriptional activity at that locus. However, what makes this very complex is, that even if the activation of the promotor would always be the same and would not be regulated by for example methylation, than it would still not be the only determining factor for RNA-transcript abundance. Processes like splicing, nonsense-mediated decay and GC content among others will all influence the amount of RNA-transcripts you get.
Using reporter constructs could may be a different story.