BACKGROUND: Machine learning is a powerful approach for describing and predicting classes in microarray data. Although several comparative studies have investigated the relative performance of various machine learning methods, these often do not account for the fact that performance (e.g. error rate) is a result of a series of analysis steps of which the most important are data normalization, gene selection and machine learning. RESULTS: In this study, we used seven previously published cancer-related microarray data sets to compare the effects on classification performance of five normalization methods, three gene selection methods with 21 different numbers of selected genes and eight machine learning methods. Performance in term of error rate was rigorously estimated by repeatedly employing a double cross validation approach. Since performance varies greatly between data sets, we devised an analysis method that first compares methods within individual data sets and then visualizes the comparisons across data sets. We discovered both well performing individual methods and synergies between different methods. CONCLUSION: Support Vector Machines with a radial basis kernel, linear kernel or polynomial kernel of degree 2 all performed consistently well across data sets. We show that there is a synergistic relationship between these methods and gene selection based on the T-test and the selection of a relatively high number of genes. Also, we find that these methods benefit significantly from using normalized data, although it is hard to draw general conclusions about the relative performance of different normalization procedures.

BACKGROUND: Cluster analysis, and in particular hierarchical clustering, is widely used to extract information from gene expression data. The aim is to discover new classes, or sub-classes, of either individuals or genes. Performing a cluster analysis commonly involve decisions on how to; handle missing values, standardize the data and select genes. In addition, pre-processing, involving various types of filtration and normalization procedures, can have an effect on the ability to discover biologically relevant classes. Here we consider cluster analysis in a broad sense and perform a comprehensive evaluation that covers several aspects of cluster analyses, including normalization. RESULT: We evaluated 2780 cluster analysis methods on seven publicly available 2-channel microarray data sets with common reference designs. Each cluster analysis method differed in data normalization (5 normalizations were considered), missing value imputation (2), standardization of data (2), gene selection (19) or clustering method (11). The cluster analyses are evaluated using known classes, such as cancer types, and the adjusted Rand index. The performances of the different analyses vary between the data sets and it is difficult to give general recommendations. However, normalization, gene selection and clustering method are all variables that have a significant impact on the performance. In particular, gene selection is important and it is generally necessary to include a relatively large number of genes in order to get good performance. Selecting genes with high standard deviation or using principal component analysis are shown to be the preferred gene selection methods. Hierarchical clustering using Ward's method, k-means clustering and Mclust are the clustering methods considered in this paper that achieves the highest adjusted Rand. Normalization can have a significant positive impact on the ability to cluster individuals, and there are indications that background correction is preferable, in particular if the gene selection is successful. However, this is an area that needs to be studied further in order to draw any general conclusions. CONCLUSIONS: The choice of cluster analysis, and in particular gene selection, has a large impact on the ability to cluster individuals correctly based on expression profiles. Normalization has a positive effect, but the relative performance of different normalizations is an area that needs more research. In summary, although clustering, gene selection and normalization are considered standard methods in bioinformatics, our comprehensive analysis shows that selecting the right methods, and the right combinations of methods, is far from trivial and that much is still unexplored in what is considered to be the most basic analysis of genomic data.

Recent years have seen a wealth of computational methods applied to problems stemming from molecular biology. In particular, with the completion of many new full genome sequences, great advances have been made in studying the role of non-protein-codi

MOTIVATION: Correct prediction of residue-residue contacts in proteins that lack good templates with known structure would take ab initio protein structure prediction a large step forward. The lack of correct contacts, and in particular long-range contacts, is considered the main reason why these methods often fail. RESULTS: We propose a novel hidden Markov model (HMM)-based method for predicting residue-residue contacts from protein sequences using as training data homologous sequences, predicted secondary structure and a library of local neighborhoods (local descriptors of protein structure). The library consists of recurring structural entities incorporating short-, medium- and long-range interactions and is general enough to reassemble the cores of nearly all proteins in the PDB. The method is tested on an external test set of 606 domains with no significant sequence similarity to the training set as well as 151 domains with SCOP folds not present in the training set. Considering the top 0.2 x L predictions (L = sequence length), our HMMs obtained an accuracy of 22.8% for long-range interactions in new fold targets, and an average accuracy of 28.6% for long-, medium- and short-range contacts. This is a significant performance increase over currently available methods when comparing against results published in the literature. AVAILABILITY: http://predictioncenter.org/Services/FragHMMent/.

MOTIVATION: Genome-scale 'omics' data constitute a potentially rich source of information about biological systems and their function. There is a plethora of tools and methods available to mine omics data. However, the diversity and complexity of different omics data types is a stumbling block for multi-data integration, hence there is a dire need for additional methods to exploit potential synergy from integrated orthogonal data. Rough Sets provide an efficient means to use complex information in classification approaches. Here, we set out to explore the possibilities of Rough Sets to incorporate diverse information sources in a functional classification of unknown genes. RESULTS: We explored the use of Rough Sets for a novel data integration strategy where gene expression data, protein features and Gene Ontology (GO) annotations were combined to describe general and biologically relevant patterns represented by If-Then rules. The descriptive rules were used to predict the function of unknown genes in Arabidopsis thaliana and Schizosaccharomyces pombe. The If-Then rule models showed success rates of up to 0.89 (discriminative and predictive power for both modeled organisms); whereas, models built solely of one data type (protein features or gene expression data) yielded success rates varying from 0.68 to 0.78. Our models were applied to generate classifications for many unknown genes, of which a sizeable number were confirmed either by PubMed literature reports or electronically interfered annotations. Finally, we studied cell cycle protein-protein interactions derived from both tandem affinity purification experiments and in silico experiments in the BioGRID interactome database and found strong experimental evidence for the predictions generated by our models. The results show that our approach can be used to build very robust models that create synergy from integrating gene expression data and protein features. AVAILABILITY: The Rough Set-based method is implemented in the Rosetta toolkit kernel version 1.0.1 available at: http://rosetta.lcb.uu.se/

MOTIVATION: Copy number profiling methods aim at assigning DNA copy numbers to chromosomal regions using measurements from microarray-based comparative genomic hybridizations. Among the proposed methods to this end, Hidden Markov Model (HMM)-based approaches seem promising since DNA copy number transitions are naturally captured in the model. Current discrete-index HMM-based approaches do not, however, take into account heterogeneous information regarding the genomic overlap between clones. Moreover, the majority of existing methods are restricted to chromosome-wise analysis. RESULTS: We introduce a novel Segmental Maximum A Posteriori approach, SMAP, for DNA copy number profiling. Our method is based on discrete-index Hidden Markov Modeling and incorporates genomic distance and overlap between clones. We exploit a priori information through user-controllable parameterization that enables the identification of copy number deviations of various lengths and amplitudes. The model parameters may be inferred at a genome-wide scale to avoid overfitting of model parameters often resulting from chromosome-wise model inference. We report superior performances of SMAP on synthetic data when compared with two recent methods. When applied on our new experimental data, SMAP readily recognizes already known genetic aberrations including both large-scale regions with aberrant DNA copy number and changes affecting only single features on the array. We highlight the differences between the prediction of SMAP and the compared methods and show that SMAP accurately determines copy number changes and benefits from overlap consideration.

BACKGROUND: We present an approach designed to identify gene regulation patterns using sequence and expression data collected for Saccharomyces cerevisae. Our main goal is to relate the combinations of transcription factor binding sites (also referred to as binding site modules) identified in gene promoters to the expression of these genes. The novel aspects include local expression similarity clustering and an exact IF-THEN rule inference algorithm. We also provide a method of rule generalization to include genes with unknown expression profiles. RESULTS: We have implemented the proposed framework and tested it on publicly available datasets from yeast S. cerevisae. The testing procedure consists of thorough statistical analyses of the groups of genes matching the rules we infer from expression data against known sets of co-regulated genes. For this purpose we have used published ChIP-Chip data and Gene Ontology annotations. In order to make these tests more objective we compare our results with recently published similar studies. CONCLUSION: Results we obtain show that local expression similarity clustering greatly enhances overall quality of the derived rules, both in terms of enrichment of Gene Ontology functional annotation and coherence with ChIP-Chip binding data. Our approach thus provides reliable hypotheses on co-regulation that can be experimentally verified. An important feature of the method is its reliance only on widely accessible sequence and expression data. The same procedure can be easily applied to other microbial organisms.

Comparative modeling methods can consistently produce reliable structural models for protein sequences with more than 25% sequence identity to proteins with known structure. However, there is a good chance that also sequences with lower sequence identity have their structural components represented in structural databases. To this end, we present a novel fragment-based method using sets of structurally similar local fragments of proteins. The approach differs from other fragment-based methods that use only single backbone fragments. Instead, we use a library of groups containing sets of sequence fragments with geometrically similar local structures and extract sequence related properties to assign these specific geometrical conformations to target sequences. We test the ability of the approach to recognize correct SCOP folds for 273 sequences from the 49 most popular folds. 49% of these sequences have the correct fold as their top prediction, while 82% have the correct fold in one of the top five predictions. Moreover, the approach shows no performance reduction on a subset of sequence targets with less than 10% sequence identity to any protein used to build the library.

MOTIVATION: Microarray technology enables large-scale inference of the participation of genes in biological process from similar expression profiles. Our aim is to induce classificatory models from expression data and biological knowledge that can automatically associate genes with novel hypotheses of biological process. RESULTS: We report a systematic supervised learning approach to predicting biological process from time series of gene expression data and biological knowledge. Biological knowledge is expressed using gene ontology and this knowledge is associated with discriminatory expression-based features to form minimal decision rules. The resulting rule model is first evaluated on genes coding for proteins with known biological process roles using cross validation. Then it is used to generate hypotheses for genes for which no knowledge of participation in biological process could be found. The theoretical foundation for the methodology based on rough sets is outlined in the paper, and its practical application demonstrated on a data set previously published by Cho et al. (Nat. Genet., 27, 48-54, 2001). AVAILABILITY: The Rosetta system is available at http://www.idi.ntnu.no/~aleks/rosetta. SUPPLEMENTARY INFORMATION: http://www.lcb.uu.se/~hvidsten/bioinf_cho/