BETASCAN: Probable β-amyloids Identified by Pairwise Probabilistic Analysis

Amyloids and prion proteins are clinically and biologically important β-structures, whose supersecondary structures are difficult to determine by standard experimental or computational means. In addition, significant conformational heterogeneity is known or suspected to exist in many amyloid fibrils. Recent work has indicated the utility of pairwise probabilistic statistics in β-structure prediction. We develop here a new strategy for β-structure prediction, emphasizing the determination of β-strands and pairs of β-strands as fundamental units of β-structure. Our program, BETASCAN, calculates likelihood scores for potential β-strands and strand-pairs based on correlations observed in parallel β-sheets. The program then determines the strands and pairs with the greatest local likelihood for all of the sequence's potential β-structures. BETASCAN suggests multiple alternate folding patterns and assigns relative a priori probabilities based solely on amino acid sequence, probability tables, and pre-chosen parameters. The algorithm compares favorably with the results of previous algorithms (BETAPRO, PASTA, SALSA, TANGO, and Zyggregator) in β-structure prediction and amyloid propensity prediction. Accurate prediction is demonstrated for experimentally determined amyloid β-structures, for a set of known β-aggregates, and for the parallel β-strands of β-helices, amyloid-like globular proteins. BETASCAN is able both to detect β-strands with higher sensitivity and to detect the edges of β-strands in a richly β-like sequence. For two proteins (Aβ and Het-s), there exist multiple sets of experimental data implying contradictory structures; BETASCAN is able to detect each competing structure as a potential structure variant. The ability to correlate multiple alternate β-structures to experiment opens the possibility of computational investigation of prion strains and structural heterogeneity of amyloid. BETASCAN is publicly accessible on the Web at http://betascan.csail.mit.edu. Author Summary: Amyloid is a highly ordered form of protein aggregation that a wide variety of proteins can form. While the earliest discovered amyloids were associated with systemic and neurodegenerative diseases, recent findings indicate amyloids may have myriad roles and functions ranging from learning and memory, to yeast epigenetics, to biofilm and melanin production. In this study, we expand the range and flexibility of our ability to understand how amyloid properties arise from their polypeptide sequence. By taking advantage of the intrinsic properties of a characteristic amyloid structure—parallel β-strands—and data from available protein structures, we construct and test an algorithm to predict the probability that particular portions of a protein will form amyloid. Our method has the advantage of more accurate detection of the edges of such zones, as well as the ability to consider and evaluate the likelihood of multiple folding patterns.