Although this question is found around in some forms, I am wondering about a specific example.
There is an undecided distance around the TSS in which one can look for promoters, and that is usually several kb upstream and some distance in the first exome. I get the difficulty in giving the precise number, but I have run into the paper here which gave a promoter for LAMA4 with CpGs between +5312 to +5669 from TSS.
This seemed too far, so I started snooping around and found the following GeneCards data: https://www.genecards.org/cgi-bin/carddisp.pl?gene=LAMA4
Here we can see there are promotors in both directions at +20.1kb and -44kb. The gene is large but this still seems too much.
I would very much like a meaningful comment on the whole matter:
1) Are the values reasonable? +5K? +20K? -44K?
2) This promoter/enhancer, is there an overlap between those I wasn't aware of? Can a faraway enhancer be in anyway confused with a promotor?
Hello there. I'd say that you're focusing your efforts on a single gene. The question you asked is, how unlikely is the promoter sequence's distance? That is a bioinformatics question certainly. To 'guestimate' whether this anwer is reasonable you'd need data on promoter locations across many genes, not just one. Your question can be re-worded as 'what is the distribution (normal? gamma? mean? median?) of promoter distances in -kbp upstream of the assigned/assembled transcription start site?' A second and related question is 'Are there alternative promoter motifs that have been un-annotated for this gene?' And thus what methodology could you use to find an alternative promoter.
Those questions aside, let us use a hypothetical to dig further. Even if the -5kbp promoter was in the top 10% of distances in the genome of interest, that still seems fairly plausible. But say it was so far outside a plausible range that the likelihood was so low you'd reject it. What does the promoter look like in related organisms? If the structure is different, then perhaps that would suggest alternative promoter motifs that might not be easily annotatable in an automated way. They're just programs with magic numbers for cutoffs. They can't predict everything accurately and you can figure it out.
The last question I'd leave you with is, how comfortable are you with looking for PCR and gel analyses? Is it possible that someone has already done 5' RACE PCR? Or used rtPCR to ask if the region between the -5kbp promoter and the putative TSS is expressed? Or post-transcriptionally modified? Lots of possibilities here. I'd be happy to continue this thread with you, bbut since i'm not an expert in your organisms I can only give you questions.
This sounds more like a (molecular) biology question, you might have more applicable and swift answers on for instance the biology stackexchange forum