MOTIVATION: Local motifs are patterns of DNA or protein sequences that occur within a sequence interval relative to a biologically defined anchor or landmark. Current protein motif discovery methods do not adequately consider such constraints to identify biologically significant motifs that are only weakly over-represented but spatially confined. Using negatives, i.e. sequences known to not contain a local motif, can further increase the specificity of their discovery. RESULTS: This article introduces the method DLocalMotif that makes use of positional information and negative data for local motif discovery in protein sequences. DLocalMotif combines three scoring functions, measuring degrees of motif over-representation, entropy and spatial confinement, specifically designed to discriminatively exploit the availability of negative data. The method is shown to outperform current methods that use only a subset of these motif characteristics. We apply the method to several biological datasets. The analysis of peroxisomal targeting signals uncovers several novel motifs that occur immediately upstream of the dominant peroxisomal targeting signal-1 signal. The analysis of proline-tyrosine nuclear localization signals uncovers multiple novel motifs that overlap with C2H2 zinc finger domains. We also evaluate the method on classical nuclear localization signals and endoplasmic reticulum retention signals and find that DLocalMotif successfully recovers biologically relevant sequence properties. AVAILABILITY: http://bioinf.scmb.uq.edu.au/dlocalmotif/

MOTIVATION: Nucleo-cytoplasmic trafficking of proteins is a core regulatory process that sustains the integrity of the nuclear space of eukaryotic cells via an interplay between numerous factors. Despite progress on experimentally characterizing a number of nuclear localization signals, their presence alone remains an unreliable indicator of actual translocation. RESULTS: This article introduces a probabilistic model that explicitly recognizes a variety of nuclear localization signals, and integrates relevant amino acid sequence and interaction data for any candidate nuclear protein. In particular, we develop and incorporate scoring functions based on distinct classes of classical nuclear localization signals. Our empirical results show that the model accurately predicts whether a protein is imported into the nucleus, surpassing the classification accuracy of similar predictors when evaluated on the mouse and yeast proteomes (area under the receiver operator characteristic curve of 0.84 and 0.80, respectively). The model also predicts the sequence position of a nuclear localization signal and whether it interacts with importin-α. AVAILABILITY: http://pprowler.itee.uq.edu.au/NucImport

BACKGROUND: The use of ontologies to control vocabulary and structure annotation has added value to genome-scale data, and contributed to the capture and re-use of knowledge across research domains. Gene Ontology (GO) is widely used to capture detailed expert knowledge in genomic-scale datasets and as a consequence has grown to contain many terms, making it unwieldy for many applications. To increase its ease of manipulation and efficiency of use, subsets called GO slims are often created by collapsing terms upward into more general, high-level terms relevant to a particular context. Creation of a GO slim currently requires manipulation and editing of GO by an expert (or community) familiar with both the ontology and the biological context. Decisions about which terms to include are necessarily subjective, and the creation process itself and subsequent curation are time-consuming and largely manual. RESULTS: Here we present an objective framework for generating customised ontology slims for specific annotated datasets, exploiting information latent in the structure of the ontology graph and in the annotation data. This framework combines ontology engineering approaches, and a data-driven algorithm that draws on graph and information theory. We illustrate this method by application to GO, generating GO slims at different information thresholds, characterising their depth of semantics and demonstrating the resulting gains in statistical power. CONCLUSIONS: Our GO slim creation pipeline is available for use in conjunction with any GO-annotated dataset, and creates dataset-specific, objectively defined slims. This method is fast and scalable for application to other biomedical ontologies.

Most of gene regulatory network (GRN) studies are based on crisp and parametric algorithms, despite inherent fuzzy nature of gene co-regulation. This paper presents adaptive fuzzy evolutionary GRN Reconstruction (AFEGRN) framework for modeling GRNs. The AFEGRN automatically determines model parameters, such as, number of clusters for fuzzy c-means using fuzzy-PBM index and estimation of Gaussian distribution algorithm. The proposed strategy was tested for breast cancer and normal GRNs. The results conformed to biological knowledge and showed that most of cancer related GRN changes were caused by differentially expressed genes. This demonstrates effectiveness of AFEGRN to model any GRN.

Genetic microarray expression data often contains multiple missing values that can significantly affect the performance of statistical and machine learning algorithms. This paper presents an innovative missing value estimation technique, called collateral missing value estimation (CMVE) which has demonstrated superior estimation performance compared with the K-nearest neighbour (KNN) imputation algorithm, the least square impute (LSImpute) and Bayesian principal component analysis (BPCA) techniques. Experimental results confirm that CMVE provides an improvement of 89%, 12% and 10% for the BRCA1, BRCA2 and sporadic ovarian cancer mutations, respectively, compared to the average error rate of KNN, LSImpute and BPCA imputation methods, over a range of randomly selected missing values. The underlying theory behind CMVE also means that it is not restricted to bioinformatics data, but can be successfully applied to any correlated data set.

MOTIVATION: Microarray data are used in a range of application areas in biology, although often it contains considerable numbers of missing values. These missing values can significantly affect subsequent statistical analysis and machine learning algorithms so there is a strong motivation to estimate these values as accurately as possible before using these algorithms. While many imputation algorithms have been proposed, more robust techniques need to be developed so that further analysis of biological data can be accurately undertaken. In this paper, an innovative missing value imputation algorithm called collateral missing value estimation (CMVE) is presented which uses multiple covariance-based imputation matrices for the final prediction of missing values. The matrices are computed and optimized using least square regression and linear programming methods. RESULTS: The new CMVE algorithm has been compared with existing estimation techniques including Bayesian principal component analysis imputation (BPCA), least square impute (LSImpute) and K-nearest neighbour (KNN). All these methods were rigorously tested to estimate missing values in three separate non-time series (ovarian cancer based) and one time series (yeast sporulation) dataset. Each method was quantitatively analyzed using the normalized root mean square (NRMS) error measure, covering a wide range of randomly introduced missing value probabilities from 0.01 to 0.2. Experiments were also undertaken on the yeast dataset, which comprised 1.7% actual missing values, to test the hypothesis that CMVE performed better not only for randomly occurring but also for a real distribution of missing values. The results confirmed that CMVE consistently demonstrated superior and robust estimation capability of missing values compared with other methods for both series types of data, for the same order of computational complexity. A concise theoretical framework has also been formulated to validate the improved performance of the CMVE algorithm. AVAILABILITY: The CMVE software is available upon request from the authors.

An optimal genetic mutation diagnosis requires proper selection of mutation classifier. This work investigates the performance of different classification, missing value estimation (MVE) and data dimension reduction techniques for the classification of gene expression data for BRCA1, BRCA2 and Sporadic mutations of epithelial ovarian cancer. Bayesian MVE and zero imputation techniques were employed to deal with missing values. Our study showed the better performance of the Bayesian technique. A novel approach is introduced to use generalized regression neural network (GRNN) as genetic mutation classifier which completely outperformed both well established support vector machine and probabilistic neural network.

This paper presents decision-based fusion models to classify BRCA1, BRCA2 and Sporadic genetic mutations for breast and ovarian cancer. Different ensembles of base classifiers using the stacked generalization technique have been proposed including support vector machines (SVM) with linear, polynomial and radial base function kernels. A generalized regression neural network (GRNN) is then applied to predict the mutation type based on the outputs of base classifiers, and experimental results show that the new proposed fusion methodology for selecting the best and removing weak classifiers outperforms single classification models.

Microarray data often contains multiple missing genetic expression values that degrade the performance of statistical and machine learning algorithms. This paper presents a K ranked diagonal covariance-based missing value estimation algorithm (KRCOV) that has demonstrated significantly superior performance compared to the more commonly used K-nearest neighbour (KNN) imputation algorithm when it is applied to estimate missing values of BRCA1, BRCA2 and Sporadic genetic mutation samples present in ovarian cancer. Experimental results confirm KRCOV outperformed both KNN and zero imputation techniques in terms of their classification accuracies when used toimpute randomly missing values from 1% to 5%.The classifier used for this purpose was the Generalized Regression Neural Network.The paper also provides a hypothesis for why KRCOV performs better than KNN not only for bioinformatics data but also for other data types having strong correlated values.