Why don't you include all of the peptides? It's going to be more representative of the protein as a whole anyway. Most of the work I've seen settles on the 10-20 peptides per protein level, I'm sure if you do some more digging you'll find some statistical reasoning for that selection. You'll still get plenty of proteins with less than that number of peptides. That's were your statistical testing will need to adequately remove any noise and ensure what you are seeing is real change and not random.

Peter, thanks for the suggestion. We are finding that the peptide fragments for a protein are different in different samples. Since the number of peptides is different for same protein, we are not sure how to rank and look at a specific number (10, 20, 30 or ?) from the available peptides. 

I see your point, in an ideal world you would use the same peptides for each protein and each sample. SWATH **should** let you compare the same peptides as it is far more reproducible than DDA methods.

If you haven't already done so, I'd suggest you have a look at the following;

Teo, G.; Kim, S.; Tsou, C.-C.; Collins, B.; Gingras, A.-C.; Nesvizhskii, A. I.; Choi, H. mapDIA: Preprocessing and Statistical Analysis of Quantitative Proteomics Data from Data Independent Acquisition Mass Spectrometry. Journal of Proteomics.

Goeminne, L. J. E.; Argentini, A.; Martens, L.; Clement, L. Summarisation vs Peptide-Based Models in Label-Free Quantitative Proteomics: Performance, Pitfalls, and Data Analysis Guidelines. J. Proteome Res. 2015.

Keller, A.; Bader, S. L.; Shteynberg, D.; Hood, L.; Moritz, R. L. Automated Validation of Results and Removal of Fragment Ion Interferences in Targeted Analysis of Data Independent Acquisition MS using SWATHProphet. Mol Cell Proteomics 2015, mcp.O114.044917.

Li, Y.; Zhong, C.-Q.; Xu, X.; Cai, S.; Wu, X.; Zhang, Y.; Chen, J.; Shi, J.; Lin, S.; Han, J. Group-DIA: analysing multiple data-independent acquisition mass spectrometry data files. Nat Meth 2015, advance online publication.

https://www.bioconductor.org/packages/release/bioc/html/SWATH2stats.html

Peter, thanks for the references.