If you had mass spectrometry data available you could base your choice on this.

Otherwise, the number of gels that is required can be reduced by making western blot 'strips' of two lanes, a control lane with recombinant histones and your 'mouse model' lane. That way you can operationally determine the specificity of your antibody directed against the modified histone because it should not bind to the recombinant histone proteins.

A very different approach is to simply use anti-H3K27ac antibodies alongside anti-H3K4me3 and -H3K4me1 antibodies. The H3K27ac modification is found on most active gene promoters and distal elements, which you can then stratify in promoters (H3K4me3 present) or distal elements (H3K4me1 present + H3K4me3 absent). This latter set-up is certainly powerful when you aim to determine an epigenome from a functional point of view, as you look at the transcription program itself rather than its output.

If however, you want to define important functional roles for other histone post-translational modifications, the shortest route is to screen mass-spec data sets or to perform the western blot screen, as you indicated in your question. Two lanes per modification would mean that you should need at most 100 lanes (there are ~ 50 histone PTM against which an antibody has been produced) which means 10 blots, which can be generated in less than one week. Perhaps start with the 10 most interesting candidate histone modifications for which you have an antibody? Assuming your 'mouse model' has two states and you take the recombinant histone lane along, that would mean 10 x 3 lanes, i.e.; 3 gels with 12 lanes each, including room for size markers.

Hello Nicolas. This may be a very late response to your inquiry, but I stumbled across your question and thought that our histone modification multiplex ELISAs would be perfect for you, if you still have such a need.

http://www.epigentek.com/catalog/epiquik-histone-h3-modification-multiplex-assay-kit-colorimetric-p-3606.html

Great! Sounds really interesting!

I saw the 21 modification in 96 well plate, how many "conditions" can you test within the same plate?

Well, it's a total of 96 wells. 4 wells are controls, 4 wells are for total H3, and 4 wells for 21 different methylated, acetylated, and phosphorylated modifications. We would recommend doing duplicates, so you could test two "conditions" per modification, I suppose. You can see the complete layout at http://www.epigentek.com/catalog/images/products/p3000/p3100plate.gif

There's also a H4 multiplex kit as well, cat # P-3102.