Indexing and integrating multiple features for WWW images

In this paper, we present a novel indexing technique called Multi-scale Similarity Indexing (MSI) to index image's multi-features into a single one-dimensional structure. Both for text and visual feature spaces, the similarity between a point and a local partition's center in individual space is used as the indexing key, where similarity values in different features are distinguished by different scale. Then a single indexing tree can be built on these keys. Based on the property that relevant images have similar similarity values from the center of the same local partition in any feature space, certain number of irrelevant images can be fast pruned based on the triangle inequity on indexing keys. To remove the dimensionality curse existing in high dimensional structure, we propose a new technique called Local Bit Stream (LBS). LBS transforms image's text and visual feature representations into simple, uniform and effective bit stream (BS) representations based on local partition's center. Such BS representations are small in size and fast for comparison since only bit operation are involved. By comparing common bits existing in two BSs, most of irrelevant images can be immediately filtered. To effectively integrate multi-features, we also investigated the following evidence combination techniques-Certainty Factor, Dempster Shafer Theory, Compound Probability, and Linear Combination. Our extensive experiment showed that single one-dimensional index on multi-features improves multi-indices on multi-features greatly. Our LBS method outperforms sequential scan on high dimensional space by an order of magnitude. And Certainty Factor and Dempster Shafer Theory perform best in combining multiple similarities from corresponding multiple features.

Kalata B8, a novel antiviral circular protein, exhibits conformational flexibility in the cystine knot motif

The cyclotides are a family of circular proteins with a range of biological activities and potential pharmaceutical and agricultural applications. The biosynthetic mechanism of cyclization is unknown and the discovery of novel sequences may assist in achieving this goal. In the present study, we have isolated a new cyclotide from Oldenlandia affinis, kalata B8, which appears to be a hybrid of the two major subfamilies (Mobius and bracelet) of currently known cyclotides. We have determined the three-dimensional structure of kalata B8 and observed broadening of resonances directly involved in the cystine knot motif, suggesting flexibility in this region despite it being the core structural element of the cyclotides. The cystine knot motif is widespread throughout Nature and inherently stable, making this apparent flexibility a surprising result. Further-more, there appears to be isomerization of the peptide backbone at an Asp-Gly sequence in the region involved in the cyclization process. Interestingly, such isomerization has been previously characterized in related cyclic knottins from Momordica cochinchinensis that have no sequence similarity to kalata B8 apart from the six conserved cysteine residues and may result from a common mechanism of cyclization. Kalata B8 also provides insight into the structure-activity relationships of cyclotides as it displays anti-HIV activity but lacks haemolytic activity. The 'uncoupling' of these two activities has not previously been observed for the cyclotides and may be related to the unusual hydrophilic nature of the peptide.

Molecular recognition of an RNA trafficking element by heterogeneous nuclear ribonucleoprotein A2

Heterogeneous nuclear ribonucleoprotein (hnRNP) A2 is a multitasking protein involved in RNA packaging, alternative splicing of pre-mRNA. telomere maintenance, cytoplasmic RNA trafficking, and translation. It binds short segments of single-stranded nucleic acids, including the A2RE11 RNA element that is necessary and sufficient for cytoplasmic transport of a subset of rnRNAs in oligodendrocytes and neurons. We have explored the structures of hnRNP A2, its RNA recognition motifs (RRMs) and Gly-rich module, and the RRM complexes with A2RE11. Circular dichroism spectroscopy showed that the secondary structure of the first 189 residues of hnRNP A2 parallels that of the tandem beta alpha beta beta alpha beta RRMs of its paralogue, hnRNP A1, previously deduced from X-ray diffraction studies. The unusual GRD was shown to have substantial beta-sheet and beta-turn structure. Sedimentation equilibrium and circular dichroism results were consistent with the tandem RRM region being monomeric and supported earlier evidence for the binding of two A2RE11 oligoribonucleotides to this domain, in contrast to the protein dimer formed by the complex of hnRNP A1 with the telomeric ssDNA repeat. A three-dimensional structure for the N-terminal, two-RRM-containing segment of hnRNP A2 was derived by homology modeling. This structure was used to derive a model for the complex with A2RE11 using the previously described interaction of pairs of stacked nucleotides with aromatic residues on the RRM beta-sheet platforms, conserved in other RRM-RNA complexes, together with biochemical data and molecular dynamics-based observations of inter-RRM mobility.

Predicting residue-wise contact orders in proteins by support vector regression

Background: The residue-wise contact order (RWCO) describes the sequence separations between the residues of interest and its contacting residues in a protein sequence. It is a new kind of one-dimensional protein structure that represents the extent of long-range contacts and is considered as a generalization of contact order. Together with secondary structure, accessible surface area, the B factor, and contact number, RWCO provides comprehensive and indispensable important information to reconstructing the protein three-dimensional structure from a set of one-dimensional structural properties. Accurately predicting RWCO values could have many important applications in protein three-dimensional structure prediction and protein folding rate prediction, and give deep insights into protein sequence-structure relationships. Results: We developed a novel approach to predict residue-wise contact order values in proteins based on support vector regression (SVR), starting from primary amino acid sequences. We explored seven different sequence encoding schemes to examine their effects on the prediction performance, including local sequence in the form of PSI-BLAST profiles, local sequence plus amino acid composition, local sequence plus molecular weight, local sequence plus secondary structure predicted by PSIPRED, local sequence plus molecular weight and amino acid composition, local sequence plus molecular weight and predicted secondary structure, and local sequence plus molecular weight, amino acid composition and predicted secondary structure. When using local sequences with multiple sequence alignments in the form of PSI-BLAST profiles, we could predict the RWCO distribution with a Pearson correlation coefficient (CC) between the predicted and observed RWCO values of 0.55, and root mean square error (RMSE) of 0.82, based on a well-defined dataset with 680 protein sequences. Moreover, by incorporating global features such as molecular weight and amino acid composition we could further improve the prediction performance with the CC to 0.57 and an RMSE of 0.79. In addition, combining the predicted secondary structure by PSIPRED was found to significantly improve the prediction performance and could yield the best prediction accuracy with a CC of 0.60 and RMSE of 0.78, which provided at least comparable performance compared with the other existing methods. Conclusion: The SVR method shows a prediction performance competitive with or at least comparable to the previously developed linear regression-based methods for predicting RWCO values. In contrast to support vector classification (SVC), SVR is very good at estimating the raw value profiles of the samples. The successful application of the SVR approach in this study reinforces the fact that support vector regression is a powerful tool in extracting the protein sequence-structure relationship and in estimating the protein structural profiles from amino acid sequences.

The BAR domain proteins: Molding membranes in fission, fusion, and phagy

The Bin1/amphiphysin/Rvs167 (BAR) domain proteins are a ubiquitous protein family. Genes encoding members of this family have not yet been found in the genomes of prokaryotes, but within eukaryotes, BAR domain proteins are found universally from unicellular eukaryotes such as yeast through to plants, insects, and vertebrates. BAR domain proteins share an N-terminal BAR domain with a high propensity to adopt alpha-helical structure and engage in coiled-coil interactions with other proteins. BAR domain proteins are implicated in processes as fundamental and diverse as fission of synaptic vesicles, cell polarity, endocytosis, regulation of the actin cytoskeleton, transcriptional repression, cell-cell fusion, signal transduction, apoptosis, secretory vesicle fusion, excitation-contraction coupling, learning and memory, tissue differentiation, ion flux across membranes, and tumor suppression. What has been lacking is a molecular understanding of the role of the BAR domain protein in each process. The three-dimensional structure of the BAR domain has now been determined and valuable insight has been gained in understanding the interactions of BAR domains with membranes. The cellular roles of BAR domain proteins, characterized over the past decade in cells as distinct as yeasts, neurons, and myocytes, can now be understood in terms of a fundamental molecular function of all BAR domain proteins: to sense membrane curvature, to bind GTPases, and to mold a diversity of cellular membranes.

Structural ambiguity of the Chinese version of the Hospital Anxiety and Depression Scale in patients with coronary heart disease

Background The Hospital Anxiety and Depression Scale (HADS) is a widely used screening tool designed as a case detector for clinically relevant anxiety and depression. Recent studies of the HADS in coronary heart disease (CHD) patients in European countries suggest it comprises three, rather than two, underlying sub-scale dimensions. The factor structure of the Chinese version of the HADS was evaluated in patients with CHD in mainland China. Methods Confirmatory factor analysis (CFA) was conducted on self-report HADS forms from 154 Chinese CHD patients. Results Little difference was observed in model fit between best performing three-factor and two-factor models. Conclusion The current observations are inconsistent with recent studies highlighting a dominant underlying tri-dimensional structure to the HADS in CHD patients. The Chinese version of the HADS may perform differently to European language versions of the instrument in patients with CHD.

Edge-coloured cube decompositions

An edge-colored graph is a graph H together with a function f:E(H) → C where C is a set of colors. Given an edge-colored graph H, the graph induced by the edges of color c C is denoted by H(c). Let G, H, and J be graphs and let μ be a positive integer. A (J, H, G, μ) edge-colored graph decomposition is a set S = {H 1,H 2,...,H t} of edge-colored graphs with color set C = {c 1, c 2,..., c k} such that Hi ≅ H for 1 ≤ i ≤ t; Hi (cj) ≅ G for 1 ≤ i ≤ t and ≤ j ≤ k; and for j = 1, 2,..., k, each edge of J occurs in exactly μ of the graphs H 1(c j ), H 2(c j ),..., H t (c j ). Let Q 3 denote the 3-dimensional cube. In this paper, we find necessary and sufficient conditions on n, μ and G for the existence of a (K n ,Q 3,G, μ) edge-colored graph decomposition. © Birkhäuser Verlag, Basel 2007.

Anisotropic viscous models of large-deformation Mohr-Coulomb failure

We have developed a way to represent Mohr-Coulomb failure within a mantle-convection fluid dynamics code. We use a viscous model of deformation with an orthotropic viscoplasticity (a different viscosity is used for pure shear to that used for simple shear) to define a prefered plane for slip to occur given the local stress field. The simple-shear viscosity and the deformation can then be iterated to ensure that the yield criterion is always satisfied. We again assume the Boussinesq approximation, neglecting any effect of dilatancy on the stress field. An additional criterion is required to ensure that deformation occurs along the plane aligned with maximum shear strain-rate rather than the perpendicular plane, which is formally equivalent in any symmetric formulation. We also allow for strain-weakening of the material. The material can remember both the accumulated failure history and the direction of failure. We have included this capacity in a Lagrangian-integration-point finite element code and show a number of examples of extension and compression of a crustal block with a Mohr-Coulomb failure criterion. The formulation itself is general and applies to 2- and 3-dimensional problems.

Indexing text and visual features for WWW images

In this paper, we present a novel indexing technique called Multi-scale Similarity Indexing (MSI) to index imagersquos multi-features into a single one-dimensional structure. Both for text and visual feature spaces, the similarity between a point and a local partitionrsquos center in individual space is used as the indexing key, where similarity values in different features are distinguished by different scale. Then a single indexing tree can be built on these keys. Based on the property that relevant images haves similar similarity values from the center of the same local partition in any feature space, certain number of irrelevant images can be fast pruned based on the triangle inequity on indexing keys. To remove the ldquodimensionality curserdquo existing in high dimensional structure, we propose a new technique called Local Bit Stream (LBS). LBS transforms imagersquos text and visual feature representations into simple, uniform and effective bit stream (BS) representations based on local partitionrsquos center. Such BS representations are small in size and fast for comparison since only bit operation are involved. By comparing common bits existing in two BSs, most of irrelevant images can be immediately filtered. Our extensive experiment showed that single one-dimensional index on multi-features improves multi-indices on multi-features greatly. Our LBS method outperforms sequential scan on high dimensional space by an order of magnitude.