Biclustering of gene expression data is used to discover groups of genes that are co-expressed over a subset of tested conditions. The objective is to maximize the detection of significant biclusters; to do so, most approaches employ a heuristic approximation in order to avoid a non-polynomial computational complexity. Previous algorithms have focused on enabling the discovery of biologically relevant results within the scope of single studies, where data size and complexity are limited. New methods and algorithms are required in order to enable applications of biclustering to larger scale data sets that can span multiple experiments and that are potentially far more heterogenous. The BiMax [ 1 ] algorithm uses a binary representation of the gene expression matrix that has been proven to discover enriched modules of biologically relevant genes in gene expression data. This model of biclustering allows for exact solutions, however, the BiMax algorithm performs best on a restricted size of input data. We can view the biclustering formulation of BiMax as the search for all maximal bicliques in a bipartite graph; where the nodes are genes or experiments and a connection between a gene and an experiment exists if the gene was significantly expressed in that experiment. We propose a new algorithm capable of enumerating all biclusters on such a graph. In order to solve the maximal biclique enumeration problem, we make use of the backtracking Bron-Kerbosch algorithm [ 2 ] for maximal clique enumeration. We have developed and successfully tested a new algorithm, the Bipartite Bron-Kerbosch algorithm, which uses similar principles to Bron-Kerbosch but traverses the bicliques on bipartite graphs. This approach enables the algorithm to explore all maximal bicliques without visiting branches of the search tree that contain previously discovered biclusters. Our results, see Table 1 , conclude that the new algorithm is significantly faster at bicluster exploration than BiMax, demonstrating a factor n improvement in running time (where n is proportional to the input data size). For instance, with input data of 800 genes and 800 experiments, BiMax solved for the over 500 thousand biclusters in just over three minutes whereas the Bipartite Bron-Kerbosch algorithm takes approximately 3 seconds.

Abstract Motivation: Tandem mass spectrometry allows for high-throughput identification of complex protein samples. Searching tandem mass spectra against sequence databases is the main analysis method nowadays. Since many peptide variations are possible, including them in the search space seems only logical. However, the search space usually grows exponentially with the number of independent variations and may therefore overwhelm computational resources. Results: We provide fast, cache-efficient search algorithms to screen large peptide search spaces including non-tryptic peptides, whole genomes, dozens of posttranslational modifications, unannotated point mutations and even unannotated splice sites. All these search spaces can be screened simultaneously. By optimizing the cache usage, we achieve a calculation speed that closely approaches the limits of the hardware. At the same time, we control the size of the overall search space by limiting the combinations of variations that can co-occur on the same peptide. Using a hypergeometric scoring scheme, we applied these algorithms to a dataset of 1 420 632 spectra. We were able to identify a considerable number of peptide variations within a modest amount of computing time on standard desktop computers. Availability: PepSplice is available as a C++ application for Linux, Windows and OSX at www.ti.inf.ethz.ch/pw/software/pepsplice/. It is open source under the revised BSD license. Contact: [email protected] or [email protected] Supplementary information: Supplementary data are available at Bioinformatics online.

MOTIVATION: In recent years, there have been various efforts to overcome the limitations of standard clustering approaches for the analysis of gene expression data by grouping genes and samples simultaneously. The underlying concept, which is often referred to as biclustering, allows to identify sets of genes sharing compatible expression patterns across subsets of samples, and its usefulness has been demonstrated for different organisms and datasets. Several biclustering methods have been proposed in the literature; however, it is not clear how the different techniques compare with each other with respect to the biological relevance of the clusters as well as with other characteristics such as robustness and sensitivity to noise. Accordingly, no guidelines concerning the choice of the biclustering method are currently available. RESULTS: First, this paper provides a methodology for comparing and validating biclustering methods that includes a simple binary reference model. Although this model captures the essential features of most biclustering approaches, it is still simple enough to exactly determine all optimal groupings; to this end, we propose a fast divide-and-conquer algorithm (Bimax). Second, we evaluate the performance of five salient biclustering algorithms together with the reference model and a hierarchical clustering method on various synthetic and real datasets for Saccharomyces cerevisiae and Arabidopsis thaliana. The comparison reveals that (1) biclustering in general has advantages over a conventional hierarchical clustering approach, (2) there are considerable performance differences between the tested methods and (3) already the simple reference model delivers relevant patterns within all considered settings.

BACKGROUND: Gene function analysis often requires a complex and laborious sequence of laboratory and computer-based experiments. Choosing an effective experimental design generally results from hypotheses derived from prior knowledge or experimentation. Knowledge obtained from meta-analyzing compendia of expression data with annotation libraries can provide significant clues in understanding gene and network function, resulting in better hypotheses that can be tested in the laboratory. DESCRIPTION: Genevestigator is a microarray database and analysis system allowing context-driven queries. Simple but powerful tools allow biologists with little computational background to retrieve information about when, where and how genes are expressed. We manually curated and quality-controlled 3110 mouse Affymetrix arrays from public repositories. Data queries can be run against an annotation library comprising 160 anatomy categories, 12 developmental stage groups, 80 stimuli, and 182 genetic backgrounds or modifications. The quality of results obtained through Genevestigator is illustrated by a number of biological scenarios that are substantiated by other types of experimentation in the literature. CONCLUSION: The Genevestigator-Mouse database effectively provides biologically meaningful results and can be accessed at https://www.genevestigator.ethz.ch.

De novo Sequencing of peptides is a challenging task in proteome re- search. While there exist reliable DNA-sequencing methods, the high- throughput de novo sequencing of proteins by mass spectrometry is still an open problem. Current approaches suffer from a lack in precision to detect mass peaks in the spectrograms. In this paper we present a novel method for de novo peptide sequencing based on a hidden Markov model. Experiments effectively demonstrate that this new method signif- icantly outperforms standard approaches in matching quality.