We have shown that counting the domain-to-protein links observed in the protein and protein domain/motifs data sets and analysis of statistical distributions of these counts lead to common probabilistic model of evolution of proteome complexity of the archael, bacterial and eukaryotic organisms (Kuznetsov et al., 2002; Kuznetsov, 2003a, b). In this work, using InterPro data sets, we test several basic assumptions and predictions of our model.