Accurate promoter prediction - (Jun/01/2009 )

Hi,

I'm an undergraduate volunteering in a lab for the summer. We're working on isolating a novel promoter region in the rat gnome. I was wondering if anybody had any information on promoter prediction software. I have read a paper on promoter prediction software http://www.nature.com/nbt/journal/v22/n11/full/nbt1032.html, but i had some trouble using some of the programs.

CpGproD- I'm having trouble with RepeatMasker due to unfamiliarity with Unix(I dont have it installed). Tips appreciated.

Eponine- The webinterface seems to be offline and I am having trouble executing the .jar file.

FirstEF- I'm not getting any results from this program. Are there specific parameters I should be setting for the query?

McPromoter- I got results from this, however the data output is somewhat hard to interpret for someone with little experience in this. This is the output image http://img188.imageshack.us/my.php?image=rnb1.png. I'm guessing if the sequence crosses the threshold it is like there is a promoter there.

Neural Network Promoter Prediction 2.2- I got 11 hits with this program. Each hit being exactly 50bp long. Is this reliable?

Promoter 2.0- The webinterface also seems to be down and the software install is also in Unix.

Genomatix trial- The grad student that is doing the research got 4 hits here.

PROSCAN- 1 hit

I was wondering if anybody has any information on the reliability of these or if there has been any new software developed in recent times.

Thanks,
Alan

The program I like most is Neural Network Promoter Prediction which gives me most reliable prediction than other program. The 50 bp sequence it gives is just the core promoter with the TSS in bold face. If your purpose is to isolate, clone and characterize a promoter sequence, you should include sequence at lease 1 kb upstream and some sequence downstream.

Does the promoter contain any CpG island? if yes, prediction is more accurate by most program.

pcrman on Jun 1 2009, 10:01 PM said:

Thanks for replying We're doing follow up research on a gene. Previous studies have suggested that there is a secondary promoter within the gene sequence. However, the promoter region was not characterized and not much is know about it. We're unsure if the promoter contains CpG islands. We do plan on cloning the sequence, but we wanted to narrow down the possible promoter region as much as possible before we do any wet lab work. Iv only been in the lab for a week, so I'm not exactly an expert on it yet . NNPP gave some hits that overlapped within 1kb. Would you suggest cloning each hit independently or clustering the predicted promoters?

hi there,

i am trying to find promoter regions in human chr#22, when i try to use online web server promoter prediction programs
i too face similar hurdles,
i wish to know from ur experience how u solved this problem.
which are the best promoter prediction programs do u suggest?

Thanks,
Swarna

hi,

this is an area that people often have issues with

do not rely on bioinformatics for promoter prediction.

J

drjcroberts on Aug 5 2009, 10:45 AM said:

I like Transfac professional for its transcription factor binding site analyses and related tools, (it is subscription based), but I am not sure it works for what you have in mind.

Hi

I personally rely on Neural Network Promoter Prediction program, to carry out my experiments. Till date it has given me satisfactory results.

Biointelligence
biointelligence.wordpress.com