I am sure there is a bias in this analysis, likely due to the fact as to what tissues are primarily analyzed. For instance, I would assume keratins to be highly abundant as well. It does rarely make sense for a proteomics researchers to shave off and digest the fur of a mouse to look for keratins (except for when one looks for some trichomegaly or anything else related to hair growth). So, these data are not reported and should be missing. Nevertheless, it is probably the most accurate information available out there at this point.
I would point you to the article "Initial Quantitative Proteomic Map of 28 Mouse Tissues Using the SILAC Mouse" :
http://www.mcponline.org/content/12/6/1709.long
In this work they did proteomic analysis if 28 different mouse tissues. You can use the data to extract the highest abundant proteins in each tissue rather that the whole mouse (probably more powerful information for you). For this you can use the Supplementary table 2 (Top-100 proteome):
As a proxy for protein abundance the authors use the mass spectrometric signal of each proteins. This measurement has been shown to reflect well (differences of) protein abundance.
Also to add to Sonja's answer, do you consider the mouse with its blood content? Albumin would be a very good candidate in this case!