International Journal of Computational Biology and Drug Design最新文献

Functional classification aims at grouping genes according to their molecular function or the biological process they participate in. Evaluating the validity of such unsupervised gene classification remains a challenge given the variety of distance measures and classification algorithms that can be used. We evaluate here functional classification of genes with the help of reference sets: KEGG (Kyoto Encyclopaedia of Genes and Genomes) pathways and Pfam clans. These sets represent ground truth for any distance based on GO (Gene Ontology) biological process and molecular function annotations respectively. Overlaps between clusters and reference sets are estimated by the F-score method. We test our previously described IntelliGO semantic distance with hierarchical and fuzzy C-means clustering and we compare results with the state-of-the-art DAVID (Database for Annotation Visualisation and Integrated Discovery) functional classification method. Finally, study of best matching clusters to reference sets leads us to propose a set-difference method for discovering missing information.

Comparative Molecular Field Analysis (CoMFA) and Comparative Molecular Similarity Indices Analysis (CoMSIA) were performed on a series of 103 azole dione derivatives, as selective anti-cancer inhibitors. The atom and shape based root mean square alignment yielded the best predictive CoMFA model q² = 0.923, r² = 0.980, when compared with the CoMSIA model. Docking studies were employed to position the inhibitors into active site of Crystal Structure of Delta (4)-3-ketosteroid 5-beta-reductase (PDB id: 3BUR). Results that indicate steric, electrostatic, hydrophobic, hydrogen bond donor and acceptor substituents play a significant role in design novel, potent and selective anti-cancer activity of the compounds.

With the heightened interest in Bacillus anthracis as a potential biological threat agent, novel drug targets identification is of great importance in drug discovery. This study considered a genome-wide approach to identify 270 non-redundant, non-human homologous genes and 103 essential genes of the bacteria as putative drug targets. Sub-cellular localisation of each drug target was annotated using PSORTb 3.0 and confirmation by a hybrid support vector machine analysis identified 16 membrane-bound genes with reliability index ≥4. SPAAN analysis predicted 3 adhesion-like proteins and BLAST against the MEROPS database identified 7 peptidases with inhibitors. As a case study, a homology model was built for the ptsG gene using Modeller 9v8. The work reported here identified a small subset of potential drug targets involved in vital aspects of the metabolism of pathogen, persistence, virulence and cell wall biosynthesis. Thus, this manifold workflow can speed up the process of drug target discovery.

Recent studies have suggested the common disease-rare variant (CD-RV) hypothesis in the mapping of disease-related genetic variants and have proposed a number of statistical methods to detect associations between rare variants and human inherited diseases. However, most of these methods take the selection of functional variants as a preliminary step in order to maximise the power of statistical tests. To meet this end, we put forward a filtration approach to identify genetic variants that are potentially associated with a query disease of interest from the perspective of one-class novelty learning. We propose to prioritise candidate non-synonymous single nucleotide polymorphisms (nsSNPs) relying on the integrated use of two sequence conservation properties of amino acids calculated from multiple sequence alignment of protein sequences and one functional similarity measure derived from domain-domain interaction data. We show the power of this approach in the detection of disease-related nsSNP via large-scale leave-one-out cross-validation experiments.

Stemming from the need to score relations between functional groups of genes and multiple clinical types associated with a tumour, this study proposes to use contingency-based measures to quantify such relations. It aims at reflecting a relative measure of association within a specific set of functional groups, and a specific set of clinical statuses. The proposed methodology is based on extracting features (scores) from expression sets that relate genes to multiple cancer subtypes (clinical statuses), and use those features (scores) to associate cancer subtypes with functional groups. It proposes combining t-test scores at several levels of cancer statuses' differentiation to calculate such gene features. It also proposes using contingency based measures as Jaccard and F-measure to associate gene functional groups to multiple cancer subtypes/statuses. Variations from the original Jaccard measure are proposed to reflect scores of genes' relations to classes/groups rather than using binary relations. The core objective of the experimental study is to identify the functional categories of genes that mark the change in lymph node status under each of oestrogen receptor positive and negative statuses in breast cancer expression sets.

Multitemporal Association Tracking (MAT) is a new graph-based method for multitarget tracking in biological applications that reduces the error rate and implementation complexity compared to approaches based on bipartite matching. The data association problem is solved over a window of future detection data using a graph-based cost function that approximates the Bayesian a posteriori association probability. MAT has been applied to hundreds of image sequences, tracking organelle and vesicles to quantify the deficiencies in axonal transport that can accompany neurodegenerative disorders such as Huntington's Disease and Multiple Sclerosis and to quantify changes in transport in response to therapeutic interventions.

Zebrafish is a valuable vertebrate model in life science research. In the zebrafish research, segmentation of zebrafish embryo images is a prerequisite for subsequent processing and analysis. Specifically, shape analysis of zebrafish's two important organs, the dorsal diencephalon and the ventral midbrain, is significant for studying the mutants caused by gene expression. Nevertheless, due to non-uniform intensity distribution and weak boundaries in dorsal diencephalon and ventral midbrain microscopic images, classical segmentation methods are unable to determine the precise boundaries. In this paper, a novel segmentation technique for zebrafish embryo images is proposed based on active contour model, which includes region based active contour model and geodesic active contour. Finally, the effectiveness of this approach is confirmed by the experimental results that the agreement between the algorithm and manual segmentation is more than 90%.

Wnt signalling is a critically important signalling pathway regulating embryogenesis and differentiation, and is broadly conserved amongst multicellular animals. In addition, dysregulation of Wnt signalling contributes to the pathogenesis of many human cancers, in particular colorectal cancer. Core members of the Wnt signalling pathway are quite well defined, although it has become apparent that a much broader network of interacting proteins regulates Wnt signalling activity. The goal of this paper is first to identify novel members of the Wnt regulatory network; and second, to identify sub-networks of the larger Wnt signalling network that are active in different biological contexts. We address these two questions using complementary computational approaches and show how these approaches may identify potentially novel Wnt signalling proteins as well as defining Wnt sub-networks active in different stages of colorectal cancer.