Autoantibody recognition mechanisms of MUC1

The most cost-effective blood-based, noninvasive molecular cancer biomarkers are based on p53 epitopes and MUC1 tandem repeats. Here we use dimensionally compressed bioinformatic fractal scaling analysis to compare the two distinct and comparable probes, which examine different sections of the autoantibody population, achieving combined sensitivities of order 50%. We discover a promising MUC1 epitope in the SEA region outside the tandem repeats. Introduction The value of p53 printed 15-mer epitopes as biomarkers for seromic autoantibody paratopes, much more sensitive than complete 393 amino acid (aa) p53, was established in [1], using the ~ 50,000 London patient database. In subsequent papers similar methods identified central MUC1 tandem repeats as also effective, and complementary [2,3]. We have developed thermodynamic scaling methods that utilize only parameter-free bioinformatic scales derived from the geometry of thousands of protein structures recently added to the Protein Data Base [4]. In [5] these scaling methods were tested against the p53 epitopes discovered in [1], with strongly positive results. Here we extend the analysis to MUC1 tandem repeats. Because p53 is the most studied protein, it is an attractive subject for testing scaling methods, which appear to have universal features, including quantitative aspects of evolution not accessible to phylogenetic trees [4]. Between human and mouse, p53 has evolved with ~ 80% BLAST similarity. Mucin forms multiple families, the ones most relevant to humans being MUC1 and MUC4. Much less is known about mucin than about p53, so we begin by reviewing its salient feature, the 20 aa tandem strict and peripheral repeats near the N terminal of MUC1 and the more variable 16 aa tandem repeats near the N terminal of MUC4. The number of these repeats varies usually from 30 to 90 in human individuals, but is only 16 in mouse [6].