Hi,

both your questions are somewhat tricky. Generally, most of the regulation occurs in the proximal promoter - typically in the region between 0-500 (or 0-1000). However, it is not rare, that regions over 10 kbp upstream from transcription start play important role.

Regardint the prediction - it is just prediction, that might help you during the initial screening, but unless you really prove given transcription factor at given site, you can not believe it.

ok. thank you so much.

It is not clear what you wish to do. In the absence of ChIP data you will find all sequences matching the TF binding sites regardless of the promoter length used. One possibility would be to see if the matching sequences are more abundant close to the transcription start sites (TSSs) than further away, and then make some inference. UCSC gives the promoter sequences for 1kb, 2kb and 3kb from the TSSs, which you can download.  

2kb.

roughly for predictions we start with 500 bp upstream and then extend if required.

Usually, less than 200 bp upstream the TSS  is the proximal promoter (meaning the minimum promoter needed to have a basal activity), then the distal promoter goes through 2-5 kb (modulation by basic transcription factors) , then enhancer and silencer regions that may be very far away (it can be on other chromosome!).

But all is depending of your species (here is for human), and what kind of experiments you need to do...

i wanted to predict regulators of my gene of interest so i was looking for transcription factors that might bind to my gene of interest. i had another question that is how do proteins binding n negative strand modulate the activity of gene of interest?cis mediated regulation s suppose to be done by proteins bnding on the same strand as that of gene i guess...