journal menu
abstracts
Amyloid diseases, including Alzheimer's, Parkinson's, and the prion conditions, are each associated with a particular protein in fibrillar form. At the morphological level, these fibers appear similar and are termed "amyloid." We found that the adhesive segments of amyloid fibers are short protein sequences which form pairs of interdigitated, in-register beta sheets. These amyloid fibrils were long suspected to be the disease agents, but evidence suggests that in the neurodegenerative diseases, smaller, often transient and polymorphic oligomers are the toxic entities. We have identified a segment of the amyloid-forming protein, alphaB crystallin, which forms an oligomeric complex exhibiting properties of other amyloid oligomers: beta-sheet-rich structure, cytotoxicity, and recognition by an anti-oligomer antibody. The X-ray-derived atomic structure of the oligomer reveals a cylindrical barrel, formed from six anti-parallel, out-of-register protein strands, which we term a cylindrin. The cylindrin structure is compatible with sequence segments from the Abeta protein of Alzheimer's disease and from other amyloid proteins. Cylindrins offer models for the hitherto elusive structures of amyloid oligomers, and are distinct in structure from amyloid fibrils.
abstracts
Solving molecular structures from X-ray diffraction data without knowledge of the phases is an example of an inverse problem. Solving such problems requires searching a potentially high dimensional parameter space (e.g. of phases) and recognizing when a correct solution is encountered. Computers are good at performing rapid searches, whereas teaching a computer to recognize subtle patterns is generally challenging. Humans are naturally better at the latter task, but lack the search speed required for many scientific problems. We have begun experiments to `crowd-source' the ab-initio phase problem in crystallography by harnessing the pattern recognition capabilities of a large group of human game players to drive a computational search for good phase sets given only low resolution structure factor magnitudes [1]. The search framework is based on a genetic algorithm; each `individual' in a population has a genome that encodes a set of (initially random) phases for the given diffraction magnitudes. In an iterative procedure of Darwinian evolution (i.e. survival of the fittest), human gamers use a multi-player web-interface (CrowdPhase) to select the most fit individuals for survival based on the qualities of their corresponding electron density maps. Through mutation, recombination, and selection, the population then evolves over many generations toward a final solution. Preliminary experiments with groups of users numbering around 20 show that good phases can be obtained for very low resolution test problems. New results on more challenging test cases will be discussed.
