DNA Protein Interactions Problems

[biolearner]

I have done a lot of DNA protein interaction assay by gel mobility shift assay (GMSA) for my project, but I just found the results were not repeatable. There is a big difference with the same experimental procedure with the same sample and reagents. I wonder if other methods are better. Filter binding, footprinting, ChIP? Any advice is welcome. Here are some protocols on DNA protein interaction protocols I found.
http://www.e-biotek.com/ebiotek_283.htm

[blcr11]

I don't think I can offer much by way of advice. Gel shift and filter binding assays are the most common types of assay for DNA binding. They are also notoriously tricky to get to work reliably. If there are any changes to the protein that occur over time, it is quite possible that a sample, even one that worked perfectly once, won't work the same way twice. When you have a difficult interaction - which usually means a weak one - you have to almost be an anal-retentive with your conditions. Other possibilies might be a flourescence quenching assay where either the DNA or the protein is labeled with a fluorescent tag which is quenched by close approach to tryptophans usually. In fact, it is sometimes possible to use the flourescence quenching of the protein's own tryptophan residuse as a measure of association. Another more exotic method would be isothermal titration calorimetry, but that requires a greater amount of protein and DNA than does "simple" gel shift or filter binding. They are getting better with DNA arrays, you might look into those kinds of things, of course, I don't know if you're most interested in characterizing the type of protein that binds to a known DNA sequence, or going the other way around.

[biolearner]

Thank you for your tips.

I had the exact problem as you mentioned. I try to probe the interactions between the gene PPAR-gamma and the antigen beta-catenin.

I did not have a lot of experience with other method to probe the interactions, but I have to find a protocol that can give me a repeatable results.