Summary: pcaMethods is a Bioconductor compliant library for computing principal component analysis (PCA) on incomplete data sets. The results can be analyzed directly or used to estimate missing values to enable the use of missing value sensitive statistical methods. The package was mainly developed with microarray and metabolite data sets in mind, but can be applied to any other incomplete data set as well.

The outstanding mechanical toughness of silk fibers is thought to be caused by embedded crystalline units acting as cross links of silk proteins in the fiber. Here, we examine the robustness of these highly ordered beta-sheet structures by molecular dynamics simulations and finite element analysis. Structural parameters and stress-strain relationships of four different models, from spider and Bombyx mori silk peptides, in antiparallel and parallel arrangement, were determined and found to be in good agreement with x-ray diffraction data. Rupture forces exceed those of any previously examined globular protein many times over, with spider silk (poly-alanine) slightly outperforming Bombyx mori silk ((Gly-Ala)(n)). All-atom force distribution analysis reveals both intrasheet hydrogen-bonding and intersheet side-chain interactions to contribute to stability to similar extent. In combination with finite element analysis of simplified beta-sheet skeletons, we could ascribe the distinct force distribution pattern of the antiparallel and parallel silk crystalline units to the difference in hydrogen-bond geometry, featuring an in-line or zigzag arrangement, respectively. Hydrogen-bond strength was higher in antiparallel models, and ultimately resulted in higher stiffness of the crystal, compensating the effect of the mechanically disadvantageous in-line hydrogen-bond geometry. Atomistic and coarse-grained force distribution patterns can thus explain differences in mechanical response of silk crystals, opening up the road to predict full fiber mechanics.

Summary. Background: To avoid pathological platelet aggregation by von Willebrand factor (VWF), VWF multimers are regulated in size and reactivity for adhesion by ADAMTS13-mediated proteolysis in a shear flow dependent manner. Objective and methods: We examined whether tensile stress in VWF under shear flow activates the VWF A2 domain for cleavage by ADAMTS13 using molecular dynamics simulations. We generated a full length mutant VWF featuring a homologous disulfide bond in A2 (N1493C and C1670S), in an attempt to lock A2 against unfolding. Results: We indeed observed stepwise unfolding of A2 and exposure of its deeply buried ADAMTS13 cleavage site. Interestingly, disulfide bonds in the adjacent and highly homologous VWF A1 and A3 domains obstruct their mechanical unfolding. We find this mutant A2 (N1493C and C1670S) to feature ADAMTS13-resistant behavior in vitro. Conclusions: Our results yield molecular-detail evidence for the force-sensing function of VWF A2, by revealing how tension in VWF due to shear flow selectively exposes the A2 proteolysis site to ADAMTS13 for cleavage while keeping the folded remainder of A2 intact and functional. We find the unconventional ÔknottedÕ Rossmann fold of A2 to be the key to this mechanical response, tailored for regulating VWF size and activity. Based on our model we discuss the pathomechanism of some natural mutations in the VWF A2 domain that significantly increase the cleavage by ADAMTS13 without shearing or chemical denaturation, and provide with the cleavage-activated A2 conformation a structural basis for the design of inhibitors for VWF type 2 diseases.