Build a promoter finding/prediction tool for Microbacterium (Specifically for MTB)

Question

[nkuo]

To all the bioinformatician out here.

our lab is studying tuberculosis and want to have a promoter prediction software.

We look for few promoter predicting software for prokaryotic but they are either too specific or too general. (As in not accurate enough for microbacterium)

Therefore, we think we want to build our own. But we really have no experience on this sorts of things and want to have suggestion.

Currently we found PePPER is the closest to what we think should work https://bmcgenomics.biomedcentral.com/articles/10.1186/1471-2164-13-299

What we think is we implement the same model of PePPER but with the information from Microbacterium such as sigma factor consensus sequence, TF motif.

Since we are really new to this, I was wondering if anyone on this forum have experience on building a promoter prediction system and have any suggestion on the step toward this.

Thanks

Answer 1

[natasha.sernova]

There was an old paper (2004) with many citations. I would look at it.

https://academic.oup.com/femsle/article/238/1/57/547826/Identification-of-a-promoter-motif-regulating-the

More recent papers:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4867592/

see Figure 1 there:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4867592/figure/f1/

and references within:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC538824/

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2871449/

etc..

This paper below gives consensus recognition sequence of sigma-factor::

http://jb.asm.org/content/190/3/1128.full.pdf

[nkuo]

Hi Natasha

Thanks for the input.

While haven't read specificity, I think all these are great resource for finding the consensus sequence for specific promoters in MTB. Especially the last one in which they provide which sigma factor factor to which genes.

We will definitely look into this and if you have more to share please don't be shy.

Thanks

Answer 2

[Petr Ponomarenko]

There are many approaches. I only have experience with vertibrate and plant genomes promoter prediction. What kind of supporting experimental evidence you have? Or do you want to predict promoters solely on the genome sequence?

[nkuo]

Hi Petr, thanks for the response

while not in final, I think our goal is the latter one which is to predict promoters solely on the genome sequence.

I believe our ultimate goal is to be able to see under what kind of mutation will the certain promoter to be expressed that cause the drug resistance.

But as for now, what we want is a promoter prediction of the genome that will catch all the possible promoters that is or should be expressed. Such as with the information sigma factor consensus sequence and the transcription motif we will be able to tell the following region is likely to be the promoter of the certain genes. We are expecting this will give us low specificity but that is the initial plan.

Then from that, we can use the isolates we had for different mutation to use as training set and lower our error rate.

However, all this seems to be immature and I might have missed out some important step in our planning. If there is any question or things we should be focus on that we missed entirely please let us know and we appreciate all the helps.