Abstract: The aim of the present work is to identify a connection network describing the expression profiles of the genes involved in the cellular cycle of the yeast. In order to derive such connections from the microarray data, a mathematical model is needed, whose states represent the expression levels of the genes. However, since the cellular dynamics exhibit a limit cycle behavior [1] (i.e. roughly speaking, a periodic response, which is robust to small perturbations), it is not possible to describe the whole phenomenon by means of only linear models. Therefore, we investigate the suitability of quadratic systems [2], for the description of the genetic network dynamics. A quadratic system is made up of a set of first order differential equations, with linear terms plus quadratic terms (i.e. multiplications between two state variables or between a state and an input variable).

Abstract: Population genetics studies based on the analysis of mtDNA and mitochondrial disease studies have produced a huge quantity of sequence data and related information. These data, classified as RFLPs, mtDNA SNPs, pathogenic mutations, HVS1 and HVS2 sequences, and complete mtDNA sequences, are distributed in databases differently organised:: MITOMAP, HVRBASE, mtSNPs and mtDB. The two latter databases more or less report frequency data associated with the mitochondrial SNPs, while MITOMAP simply associates the mtSNP to the different phenotypes. HmtDB, stores human complete mitochondrial genomes annotated with variability data estimated through the application of specific algorithms implemented in an automatically running Variability Generation Work Flow (VGWF). Another Work Flow, called Classification Work Flow (CWF), is implemented to perform the automatic classification of newly sequenced genomes. The aims of HmtDB are to collect and integrate all human mitochondrial genomes publicly available, to produce and provide the scientific community with site-specific nucleotidic and aminoacidic variability data estimated on all available human mitochondrial genome sequences through the automatic application of VGWF, to allow researchers to analyse their own complete or partial mitochondrial genomes in order to automatically detect the nucleotidic variants respect to the revised Cambridge Reference Sequence (rCRS) and to predict their haplogroup paternity. At present, 1255 genomes classified according to their continental origin are stored in HmtDB.

Abstract: The detection of local similarities between protein structures may give insights for the identification of a common function. Different tools exist which try to elucidate the mechanisms connecting the similarity of local subsets of residues with the biological activity of whole proteins: i) databases of functionally annotated structures and ii) structural comparison algorithms. Both types of tools suffer from two major problems. The first is the low degree of integration among databases containing functional information. The second is the low or absent integration between existing methods for structural comparison and functional annotation resources.

Abstract: A critical step in drug development is the optimization of the efficacy and specificity of candidate therapeutic compounds. Ideally, optimization is carried out using knowledge of the drug’s mode of action, i.e., the gene products with which a drug functionally interacts (drug targets). These drug targets may include genes that mediate the therapeutic effects of the drug, as well as genes that mediate undesirable side-effects. However, for many drug candidates the targets are unknown and difficult to identify among the thousands of genes in a typical genome. Previously, we developed an algorithm to identify drug targets in yeast using multiple perturbations to a cell and by measuring the response at steady-state (di Bernardo et al, Nature Biotechnology, in press). Here, we report a novel computational approach for rapidly identifying drug targets using time-course gene expression profiles. The approach filters expression profiles using a reverse-engineered gene-network model to distinguish the targets of compounds from the genes that exhibit only secondary responses. We tested this approach experimentally in E coli and show that it can overcome some of the experimental and computational limitations of existing chemogenetic approach for identifying a drug’s mode of action.

Abstract: Discovering new subclasses of pathologies and expression signatures related to specific phenotypes are challenging problems in the context of gene expression data analysis. To pursue these objectives, we need to estimate the “natural” number and the stability of the discovered clusters. To this end, new approaches based on random subspaces and bootstrap methods have been recently proposed.

Abstract: The prediction of intrinsically disordered regions in proteins from sequence is increasingly becoming of interest, as the presence of many such regions in the complete genome sequences are discovered. Recent reports also indicate disordered regions to be conserved in evolution and to have a distinct preference for certain functional classes.

Abstract: A Suffix Tree (ST) is a --now classical-- data structure, computable in linear time, which represents the most algorithmically appropriate way to store a string, in order to face problems like the Exact String Matching Problem (ESM) or the Longest Common Exact Substring Problem (LCES). Even if very efficient in solving these problems, the ST data structure suffers from an important drawback, when dealing with an Approximate String Matching Problem (ASM) or with the harder Longest Common Approximate Substring Problem (LCAS), as only exact matching can be used in visiting a ST. In the approximate cases, a suitable notion of distance (most frequently Hamming or Levenshtein distances) must come into play. However, in the literature, there is no universally accepted data structure capable to deal with approximate searches just by performing algorithmic manipulations similar to ST’s. This makes necessary, when using ST's in an approximate context, taking into account errors by using unnatural and complicate strategies, inevitably leading to cumbersome algorithms.

Abstract: Neoplastic transformation is a multistep process in which gene products of specific regulatory pathways are involved at each stage. Identification of these over-expressed or mutated gene products provides an unprecedented opportunity to address the immune system against defined antigens and eliminate transformed cells. Mice transgenic for these oncogenes (e.g. HER-2/neu, a prototype of deregulated oncogenic protein kinase membrane receptors) are ideal experimental models for assessing the potential of active immunization. The demonstration that vaccines can cure HER-2/neu transplantable tumors, prevent their onset and delay the progression of preneoplastic lesions in mice at risk suggests that efficient immunological inhibition of HER- 2/neu carcinogenesis can be achieved by specific vaccination. Since an important issue of mammary cancer immunoprevention is the definition of a set target genes other that HER-2/neu in order to make broader the coverage of the efficacy of the vaccination we decided to use part of the data previously described by us (Quaglino et al. J. Clical. Invest. 2004; Astolfi et al. Cancer Immunol., Immunot.2005) to identify a set of new putative targets to be used as TAA.

Abstract: Computational analysis of genomic sequence databases has recently revealed a striking abundance of Natural Antisense Transcripts within the mouse and human genomes. Since antisense transcription is increasingly recognized as a molecular mechanism involved in the regulation of gene expression, a significant proportion of human disease genes could potentially display antisense-mediated abnormal patterns of gene expression. Cancer is a pathological phenotype that could represent a potential target for such gene regulation mechanism, given the high number of genes governing cancer-related cellular functions such as proliferation, differentiation and apoptosis. Preliminary experimental evidence has been reported recently for the occurrence of natural antisense transcript for several genes whose function has been implicated in cancer pathogenesis. Therefore, a targeted in silico survey of antisense transcription, coupled with a detailed inspection of annotated gene features, could represent a novel tool for the identification of candidate cancer-related genes. We have performed an in silico search for “sense-antisense gene clusters” within two regions from human chromosome 6 (6q21 and 6q27) that have long been reported to carry cancer-associated deletions and rearrangements, but for which no tumor suppressor genes has been unambiguously identified. Experimental validation of each sense-antisense cluster detected in this study, followed by definition of bona fide tumor suppressor candidates based on the available annotation features, confirmed the feasibility of this approach to better define candidate cancer-associated genes.

Abstract: Motivation:Similar to what has been achieved by comparing genome structures and protein sequences, we hope to obtain valuable information about systems evolution by comparing the organization of interaction networks stored in protein interaction databases and by analyzing their variation and conservation. Equally significantly we can learn whether and how to extend the network information obtained experimentally in well-characterized model systems to different organisms.

Abstract: Hidden Markov models (HMM) are powerful machine learning tools successfully applied to problems of computational Molecular Biology. In a predictive task, the HMM is endowed with a decoding algorithm in order to assign the most probable state path, and in turn the class labeling, to an unknown sequence. The Viterbi and the posterior decoding algorithms are the most common. The former is very efficient when one path dominates, while the latter, even though does not guarantee to preserve the automaton grammar, is more effective when several concurring paths have similar probabilities. A third good alternative is 1-best, which was shown to perform equal or better than Viterbi.

Abstract: One of the major challenges in the post-genomic era is the improvemnt of the capability to recognize the disease related molecular targets. Even if the new experimental methods have greatly enhanced the analytical capability,the application of NLP (natural language processing) play a relevant role in the selection and extraction of relevant information. The published paper still constitute the main entry point to bridge togheter molcular biology, chemical and medical information. The text processing gives a great effort in the speed up the data warehousing. In this paper we present an interactive tool that allow to conceptually organize organize the Medline.

Abstract: The increasing availability and maturity of DNA microarray technology has led to an explosion of cancer related studies. In order to evaluate, integrate, and cross-validate the accumulating amount of publicly available microarray data, new methods are needed. Statistical methods have been applied to raw microarray data to identify intersections of multiple gene expression signatures from a diverse collection of data sets. However, this analysis is very stringent and time consuming. Here we propose a fast method to detect common patterns in gene expression data by mining published lists of microarray experiments

Abstract: Cryptococcus neoformans is an opportunistic human pathogen of global importance. The two strains of C. neoformans serotype D JEC21 and B-3501A are highly related but markedly different in their pathogenicity. We report their genome sequences as an important step in the elucidation of the molecular basis for Cryptococcus virulence and for its difference with other fungal pathogens.

Abstract: Analysis of inherited diseases and their associated phenotypes is of great importance to gain insight into the underlying genetic interactions and could ultimately give clinically useful insights into disease processes, including complex diseases influenced by multiple genetic loci. Nevertheless, to date few computational contributions have been proposed for this purpose mainly due to lack of controlled clinical information easily accessible and structured for computational genome-wise analyses. To enable performing comprehensive evaluations of gene annotations sparsely available in numerous different databanks accessible via Internet, we previously developed GFINDer, a Web server that dynamically aggregates functional annotations of user uploaded gene lists and allows performing their statistical analysis and mining (http://www.bioinformatics.polimi.it/GFINDer/). Exploiting and structuring information present in textual form in the Online Mendelian Inheritance in Man (OMIM) databank, we developed and made available within GFINDer new original modules specifically devoted to the analysis of inherited disorder related genes. They allow annotating large numbers of user classified biomolecular sequence identifiers with morbidity and clinical information, classifying them according to genetic disease and phenotypic location categories, and statistically analyzing the obtained classifications.

Abstract: Knowledge about the subcellular localization of proteins can provide important information about their function. A reliable automatic classification method for predicting subcellular localization from sequences is therefore a valuable tool to shed light on protein function and may help towards the solution of genomic scale level problems such as the identification of pharmaceutical targets.

Abstract: Data integration is a fundamental process in Bioinformatics because the enormous quantity of information available is often difficult to interpret. However, using a high performance platform such as GRID, it will be possible to complete important studies to improve the understanding of the biological process. Through facing this challenge, the importance of creating a data management system that guarantees efficiency on a distributed platform has emerged. This study concerns the definition of an innovative tool for the databases management in GRID technology and the implementation of a concrete case of use in integrating biological data. The core software is a Web Service that allows the execution of SQL query on a series of distributed databases, available on different computer, through the SOAP protocol. In this way, through the client, that can be run from a GRID Computing Element, it is possible to interact with the database. The data extracted from each local database are then integrated by the Web Server and sent to the application that asks for them, optimizing the communication times. Through this software it is possible to perform elaborations that involve data access and integration on GRID easily. This Web Service has been tested during the development of an integration pipeline among two important biological databases as UNIPROT and ENSEMBL in order to coordinate different information about a certain protein sequences. Thanks to this distributed system of data access it has been possible to conduct a systematic GRID analysis of the kinase sequence references in these important databases.

Abstract: One of the greatest challenges in modern molecular biology is the identification and characterization of the functional elements regulating gene expression. Two of the most important elements are transcription factors (TFs), and the sites of the genome where they can bind (TFBSs). The TF-DNA interactions, that are responsible for the modulation of gene transcription, are at the basis of many critical cellular processes, and their malfunction often involves the onset of genetic diseases. TFBSs are located either near the transcription start site of a gene (usually within 500-1000 bps), or alternatively at very large distance (often several kilobases) from it, either upstream or downstream. When the regulation of a single gene is investigated, the idea is to increase the signal/noise ratio by comparing its flanking regions (upstream and/or downstream) with homologous genome regions of the same or other organisms at different evolutionary distances. Those parts of the regions that are more conserved throughout the different species are more likely to have been preserved by evolution for their function, and thus could be (or contain) TFBSs. Most of the methods introduced so far first build a global alignment of the sequences (some pairwise, some multiple), and report the most conserved parts of the alignment (with or without further processing, for examples by looking for known TFBSs instances in them). While this approach can produce good results, since a highly conserved region can be a good candidate for a regulatory activity, some experiments have shown that real TFBSs are often mis-aligned, and fall outside the “best regions” of the alignment (that, anyway, becomes computationally problematic for long regions, especially in the case of multiple comparisons). In this work we present an algorithm that does not require a global alignment of the sequences, nor needs to be supported by matrices or instances of known TFBSs in order to detect potential regulatory motifs.

Abstract: Novel and more efficient structural prediction methods are needed in order to reduce the everincreasing gap between the number of known protein sequences and structures. Better predictions of non-local contacts between residues could help improve fold recognition and de novo modeling. In the present study, we introduce PROFcon, a novel contact prediction method based on a neural network that combines information from various sources. PROFcon predictions are then used to estimate experimental folding rates of two-state proteins (i.e. proteins that fold without intermediate states).

Abstract: Coordinate uncertainties in crystal structures of biological macromolecules can vary from 0.01 Angstroem for a well determined atom in a high-resolution structure to 1 Angstroem for a mobile atom in a structure determined at low resolution. Given this wide range, it is essential to take into account the coordinate uncertainties in the comparison of structural models in order to discriminate significant differences from noise. Furthermore, given the rapid growth of available structure data, new methods for the analysis of large numbers of structural models need to be implemented.

Abstract: The recent developments in sequencing genomes have given a new direction to bioinformatic research. Actually, the possibility of sequencing the whole genome has raised the question of discovering common features between biological sequences corresponding to different species, reflecting on common evolutionary and functional mechanisms. This reason has led researchers to look for a definition of a distance measure on sequences able to capture these common mechanisms. Most of the traditional methods for comparing biological sequences were based on the technique of sequence alignment. Nevertheless, sequence alignment considers only local mutations of the genome, therefore it is not suitable to measure events like segment rearrangements, that involve longer genomic sequences. For this reason some alignment free distance measures have been recently introduced (see [VinAl] for a survey) and most of them are based on the concept of information theory and data compression (cf. [OS, LV, CV, EMS, BCL]). Such measures are more suitable to deal with the problem of whole genome phylogeny. The intuitive idea is that the more similar two sequences are, the more effective their joint compression is than their independent compression. We introduce a new alignment free method for comparing sequences that, differently from other ones, is combinatorial by nature and does not make use of any compressor nor any information theoretic notion.

Abstract: The advent of the technology of DNA microarrays constitutes an epochal change in the study, treatment, analysis, classification and discovery of different types of cancer. The information provided by DNA microarrays allows of approaching to the problem of cancer diagnosis and treatment from a quantitative rather than qualitative point of view.

Abstract: Pathway models or protein-protein interaction networks are excellent tools for the drug discovery process. They can be used to identify and select relevant targets to test a disease hypothesis. Combining information from diverse sources (in house experiments as well as literature) potentially transforms protein-protein interaction networks into detailed descriptions of cellular pathways. Interactive diagrams allow the linking of data directly onto the pathway and in this way can be used to integrate all relevant data regarding a protein or pathway entry into one framework. This not only leads to a better understanding of the biological mechanisms of normal and disease processes but also enable scientists to compare data from diverse experimental areas.

Abstract: The neuronal cell represents a very fascinating and complex system. In addition to basic processes common to all types of eukaryotic cells such as gene transcription, protein synthesis, and metabolism, neurons are electrically excitable and able to receive and propagate excitation via thousands of synaptic contacts. Investigating neuronal functioning requires a crossdisciplinary approach, involving on the one hand quantitative experimental methods to study excitatory processes, large scale molecular networks and the kinetics of protein-protein interactions, and on the other hand computational modeling of intra-cellular processes and, as far as the synaptic transmission is concerned, inter-cellular communication. Thus, an integration of different experimental and modeling approaches is crucial for a comprehensive description of the cell and for a complete biological understanding of neuronal behaviour. Several kinds of expertise are required in order to cope with the great heterogeneity of cellular events that must be investigated and described by computational models in such a comprehensive view: molecular biology, biophysics, mathematics. But it is bioinformatics that plays a pivotal role in extracting information from the huge amounts of data stemming from recent “-omics” research, and in devising ways to integrate such diverse bits of knowledge in order to attain a truly systemic view of cells. Several Systems Biology initiatives are underway worldwide. These are usually large consortia based in the U.S.A. or Japan. A few projects have recently been established in Europe, but none devoted to the study of neurons. The European project SYMBIONIC is a Specific Support Action aimed at establishing a European-wide initiative in the field of the Systems Biology of the neuronal cell. The long-term aim of the project is to contribute to exhaustive in silico models of the neuron. Currently, the activity of SYMBIONIC is mainly focused on the training and dissemination area and on the coordination of this project with other European initiatives in the field of Systems Biology. The project partners are Lay Line Genomics (a biotech company based in Rome that is also the project coordinator), SISSA in Trieste and the University of Barcelona. About 20 other European institutions and industries actively collaborate with the project. The main objectives of SYMBIONIC are the following: 1) To disseminate ideas and techniques among young scientists, also through a training program, in order to create a new generation of specialists in the area of Neuronal Systems Biology; 2) To consolidate among researchers a view of the neuronal cell as a complex ensemble in terms of Systems Biology; 3) To increase the consciousness of the European scientific community that there is a relevant void to be filled, both in terms of scientific topics and available crossdisciplinary expertise, in the study of neuronal cells as a complex system, and in the development of relevant research strategies; 4) To establish contacts and collaborations with other European groups which are carrying out projects in the field of Systems Biology, even if related to other cellular systems. Currently, the main collaboration is with EUSYSBIO, another EU-funded initiative on Systems Biology.

Abstract: Being the prediction of protein structure a very complex task, most methodologies concentrate on the simplified task of predicting secondary structures. In this paper, we illustrate a technique based on multiple experts, aimed at predicting protein secondary structures. The prediction activity results from the interaction of a population of experts, each integrating genetic and neural technologies. Roughly speaking, an expert of this kind embodies a genetic classifier designed to control the activation of a feedforward artificial neural network for performing a locally-scoped prediction activity. Genetic and neural components (i.e., guard and embedded predictor, respectively) are devoted to perform different tasks and are supplied with different information: Each guard is aimed at (soft-)partitioning the input space, insomuch assuring both the diversity and the specialization of the corresponding embedded predictor, which in turn is devoted to perform the actual prediction.

Abstract: Divining the functional role of proteins in life processes, starting from their structural features, is the challenging purpose of structural biology. In this context, SCOP and CATH databases are the starting point for scientist aiming at discovering the relationships between function and molecular structure. The classification criteria adopted are based on the topological architecture of the molecule, thus the proteins are clustered by finding local folding motifs which are repeatedly encountered in the protein data bank. It is an assumption largely accepted that the overall structure deserves the appropriate distribution of chemical properties on the surface that the protein presents to its molecular target. Thanks to its “appearance” the protein could have the correct approach with the target or not. However, two proteins with similar structures may be divergent in their sequence, thus playing different functions. The structural classification as a tool for the individuation of a common physiological role is misleading in this case and a surface classification is more and more reputed necessary. Up to now, a successful strategy has not yet developed to reach such a general goal: the most exploited approach consists in adopting a surface patch already recognized to play a key role for a certain function and use it as a probe to explore the surfaces of all the proteins with known structure. This method, based on local features, will fail in all cases characterized by highly adaptable surfaces and when portions of surface far from the active site have a dominant allosteric effect on the functional region of the protein. In this paper we propose an original approach to classification of proteins based on their surface characteristics, which has the advantage of being not based on local surface features neither on already known functional meanings.

Abstract: The maintenance of biological databases is at present a problem of great interest since the progress made in many experimental procedures has led to an ever increasing amount of data. These data need to be structured and stored in databases and made accessible to the biological community in user-friendly ways. Although both the interest and the need of accessing biological databases are high, the mechanisms to fund their maintenance are unclear. Funding agencies cannot support data annotation in terms of labour costs and hence the development of new tools based on “data miming” technologies could greatly contribute to keep biological databases updated. Here we present a new approach aimed to contribute to the annotation in the HmtDB resource (http://www.hmdb.uniba.it/) of variability data associated to clinical phenotypes [1]. These data are prevalently available in literature where they are reported in a completely free style. Thus, we suggest the construction of a knowledge base derived from browsing papers on web and to be used in the retrieval phase. Nevertheless, problems in extracting data from literature come not only from the heterogeneity of presentation styles but mainly from the unstructured format (i.e. the natural language) in which they are represented. In this scenario, the goal is to feed a knowledge base by identifying occurrences of specific biological entities and their features as well as the particular method and experimental setting of the scientific study adopted in the publication. In this work, we describe some solutions to the problem of structuring information contained in scientific literature in digital (i.e., pdf) or paper format.

Abstract: Collecting, analyzing and extracting useful information from a very large amount of biomedical texts is a difficult task for researchers in biomedicine who need to keep up with scientific advances. Nowadays several domains in medical practice, drug development, and health care require support for such actives such as bioinformatics, medical informatics, clinical genomics, and many other sectors. Moreover, for this particular task, the data to be examined (i.e. textual data) are generally unstructured as in the case of Medline abstracts and the available resources (e.g. PubMed) and as many other textual resources such as medical records, patents etc. and they do not still provide adequate mechanisms for retrieving the required information as well as to help humans in “deeply analyse” very large amount of content. In this work we present a Text-Mining framework aiming to support biomedical researchers in the task of disease-genes relationships identification from scientific abstracts retrieved by querying Medline.

Abstract: In prokaryotes are known two kinds of transcription terminators that are distinguished by their mechanisms and DNA sequences. When RNA polymerase encounters an intrinsic terminator (RIT), it can release the nascent RNA spontaneously, but when it encounters a Rho dependent terminator (RDT), the release of the RNA depends on the action of a protein factor called Rho. RDT are involved in the gene expression as attenuators in the leader or intra operon and as terminators at the end of operons. A RDT consists of three distinct parts, which together extend over 150-200bp of DNA (figure 1). The upstream part, called the Rho utilization (rut) site, encodes a segment of the nascent transcript to which Rho can bind and is essential for starting termination. The central part, that we called Rho activity (rac) sequence, is the second mRNA binding site to which Rho can bind and is essential for helicase/traslocation activity. The downstream part, called the transcription stop point (tsp) region, is where RNA polymerase pauses during elongation in the absence of Rho. In the literature is present only a small number of studies of single RDT, very little is know about their structure and sequence and is not existing any in silico predictive method.

Abstract: Improvements of bio-nano-technologies and biomolecular techniques have led to increasing production of high-throughput experimental data. Spotted cDNA microarray is one of the most diffuse technologies, used in single research laboratories and in biotechnology service facilities. Although they are routinely performed, spotted microarray experiments are complex procedures entailing several experimental steps and actors with different technical skills and roles. During an experiment, involved actors, who can also be located in a distance, need to access and share specific experiment information according to their roles. Furthermore, complete information describing all experimental steps must be orderly collected to allow subsequent correct interpretation of experimental results. To satisfy such requirements, we developed MicroGen, a Web based system for managing information and workflow in the production pipeline of spotted microarray experiments. Our aim was to realize a multi-database system able to store all data completely characterizing different spotted microarray experiments according to the Minimum Information About Microarray Experiments (MIAME) standard, and to support the collaborative work required among multidisciplinary actors and roles involved in microarray experiment production.

Abstract: A set of over 3 million putative single Nucleotide Polymorphism (SNPs) are now available for disease association studies, eventually replacing the current RFLP and microsatellite (STRP) linkage analysis screening sets. Many different technological platforms are available for allelic discrimination and, among them, Taq Man chemistry employing 5’ nuclease allelic discrimination is one of the most widely diffused. Each Taq Assay employs two allelic specific probes for each SNPs, commonly VIC or FAM dyes. Although technological advancements in assay design and application allows monitoring SNPs at a high-throughput rate with this technology, the allele calling procedure of any single sample still requires the manual intervention of an expert operator for adjustment of SNP-dependent thresholds, signal selection, and quality revision. Usually the genotyping is calculated at the end of the amplification process resulting in an end point analysis. To enable the genotyping of multiple SNPs in several different samples we developed an algorithm for the automatic allele-calling from real time data.

Abstract: The identification of repeats is an essential step for genome analysis and annotation, but is not easy because repeats tend to be little conserved during evolution. This particular aspect of repeats makes very difficult the identification of homologous sequences that diverged significantly, both within the same genome and between genomes of different organisms.

Abstract: The ESTree db represents a collection of Prunus persica expressed sequenced tags (ESTs) and is intended as a resource for peach functional genomics. With this aim, the db has been structured to be a repository of information and links related to the sequences and to provide a userfriendly interface to allow easy querying of all the db fields. Within the month of March 2005 the third release of ESTree will be online and will include 18630 sequences, encompassing 8 libraries at different peach fruit developmental stages. A second version of the db, including only the 6155 sequences produced by the FPTP-CERSA group will also be online. The major data resources that are included in the ESTree db are: annotation both with BLASTx versus the NCBI nr database and with BLASTx versus the GO viridiplantae subset of sp-trembl; contig assembly and display; SNP analysis with AutoSNP and links to the KEGG metabolic pathways and the enzyme entries of NiceZyme (Expasy). Gene Ontology statistics are also presented both for the whole set of sequences that are included in the db and for library-specific subsets.

Abstract: Within the Biosapiens network of excellence (EC Framework VI), the Biocomputing Group of the Bologna University installed a DAS server in a pipeline connected to the EBI. Our task in collaboration with Gunnar von Hejne (Stockholm Bioinformatics Center, SCFAB, Stockholm University, Sweden), Gert Vriend (CMBI University of Nijmegen, the Netherlands) and David Jones (Bioinformatics Unit, University College London, United Kingdom) is the large scale screening of the human genome in order to annotate membrane proteins based on topology prediction of chains.

Abstract: The explosive growth of the biological data, stimulated by genome projects, has generated a parallel development of efficient computational approaches suitable for several biological research projects. In this area the need of high performance computing is growing, though usually not affordable by computational resources of a single research laboratory. Grid computing addresses this problem by coordinating and unificating several computational resources. To face the problem of searching "Conserved Sequence Tags" (CSTs) between an input DNA sequence, and several whole model genomes a grid framework can provide high performance, high availability and can fairly handle hundreds of concurrent request. Because the size of several whole genomes now exceed the memory capacity of a single machine, it is necessary to spread the search across multiple distributed working hosts to achieve high performance. This also improves the high availability, since the redundancy of the services increases the tolerance to both machine and network failures. This system also guarantees that the same services can be completed by many machines, reaching the ability to perform more requests that a single machine can handle.

Abstract: The incredible explosion of “knowledge production” in Biology in the past two decades has created a critical need for bioinformatic instruments able to manage data and facilitate their retrieval and analysis. Molecular sequences and biological data on nuclear mitochondrial genes and their products are publicly available from a wide variety of mitochondrial specialized databases. Some are species-specific, mainly human dedicated, others contain only few species and for most of them the only sequences data reported concern proteins. We have developed MitoRes database to collect and integrate information on nuclearly encoded proteins and genes targeting the mitochondrion for all metazoan species and to provide a flexible and efficient tool for the export of bio-sequences in support of researchers interested in functional characterization of gene, transcript and amino acid sequences related to biogenesis, metabolism and pathological dysfunctions of mitochondria.

Abstract: Estrogen replacement therapy is one of the most efficient treatment of menopausal symptoms. Unfortunately, despite the benefits of estrogenic therapy evidence exists of increasing number of cases of reproductive tissue cancer. This lead to a strong interest in discovering new compounds that display the benefits of estrogens avoiding such risks. We decided to apply a virtual high throughput screening, based on docking simulations, for the identification of new possible selective receptor compounds.

Abstract: Orthologous genes that maintain a single copy status in a broad range of species may indicate a selection against gene duplication. If this is the case, then duplicates of such genes which do survive may have escaped the dosage control by rapid and sizable changes in their function.

Abstract: The organic polychlorinated compounds like dichlorodiphenyltrichloroethane (DDT) with its metabolites and polychlorinated biphenyls (PCBs) are present in atmospheric particulate as persistent contaminants. They have been recognized to have detrimental health effects both on wildlife and humans acting as endocrine disrupters (EDC) due to their ability of mimicking the action of the steroid hormone thus interfering with hormone response. There are several experimental evidences that they bind and activate human steroid receptors. Despite the growing concern about the toxicological activity of EDC, molecular data of the interaction of this compounds with biological targets are still lacking. In order to better understand the ability of EDC to bind in the receptor hormone binding pocket, we have simulated by docking approach the molecular models of the complexes of estrogen, progesterone and androgen receptors with DDT and PCB family compounds.

Abstract: The vast amount of expression sequence tags (EST) data in the public databases provides an important resource for comparative and functional genomic studies and, moreover, represents a useful information for a reliable annotation of genomic sequences. ESTs are short and errorprone sequences often including artifactual contamination, nevertheless they represent the framework from which fundamental information for both bioinformatic and experimental analysis is derived. Therefore, EST data need suitable preprocessing to provide high quality information. Because of the advances in biotechnologies, large EST datasets are daily released to the scientific community and fast and efficient bioinformatic approaches are the first step to support these data management and analysis. Therefore the requirements for this kind of analysis are suitable computational tools based on advanced technologies. We describe here a pipeline analysis for ESTs clustering, assembling and annotation by parallel computing that optimizes execution time for the processing of large data sets.

Abstract: Mitochondrial disorders are clinical phenotypes associated with abnormalities of oxidative phosphorylation (OXPHOS), the primary energy-producing process of all aerobic organisms. Disorders of OXPHOS are recognized as the most common inborn errors of metabolism affecting at least one in 5000 newborn children. Except for complex II, which is composed of proteins all encoded by nuclear genes, the other OXPHOS complexes are built up of both mitochondrial and nuclear DNA encoded proteins; so, assembling the OXPHOS complexes and fine tuning their activity require specialized regulatory mechanisms to optimize the cross-talk between the two genomes and ensure the coordinated expression of their relevant products. In this context, the characterization of nuclear genes encoding for mitochondrial proteins and of functional elements regulating their expression is of crucial importance to clarify real genetic causes of mitochondrial diseases, to assess the correct diagnosis and set up new and effective therapies. Despite the long evolutionary divergent time, many key pathways that control development and cellular physiology are conserved between Drosophila and humans, and about 70% of the genes associated with human diseases have direct counterpart in the Drosophila genome. To investigate on the functional constraints acting on the evolution and on the regulatory mechanism coordinating the expression of OXPHOS genes we have identified and characterized sequence and structure of these genes in three species of diptera, D. melanogaster, D. pseudoobscura and A. gambiae, and compared them with their human counterparts. Data obtained from this study have been annotated in the MitoDrome2 database. The availability of data produced by our study in MitoDrome2 is expected to be particularly useful for biologists and clinicians interested in studies of functional genomics related to mitochondrial biogenesis, metabolism and to their pathological dysfunctions.

Abstract: A crucial issue in microarray studies is the elucidation of how genes change expression and interact as a consequence of external/internal stimuli such as illness, drug assumption, hormone stimulation. To do so one has to reconstruct the regulatory network by describing activation/inhibition and cause-effect relationships among expression profiles. Different approaches are available in literature, but the small number of available samples with respect to the number of genes constitutes a major drawback to apply these methods to real microarray data. At present, a realistic aim is the identification of modules of gene regulation, i.e. sets of genes that are possibly regulated by the same transcription factors, or potential inhibitors or activators of a group of co-expressed genes.

Abstract: This work presents a computational analysis of the molecular characteristics shared by the A domains from traditional nonribosomal peptide synthetases (NRPSs) and the group of the freestanding homologous enzymes: a-aminoadip ate semialdehyde dehydrogenase, a- aminoadipate reductase and the protein Ebony.

Abstract: The usual procedure employed in the analysis of microarray data typically involves little more than a desktop workstation and a spreadsheet program. However, managing data in the form of spreadsheet files is not convenient, particularly when submitting results to public databases prior to publication. In addition, many researchers are discovering that data coming from the latest generation of microarray chips, which may contain many tens of thousands of gene probes, cannot be processed even with the most powerful personal computer. A further deficiency is that normally there is no provision for the systematic recording of information related to the experiment themselves, e.g. platform design, sample hybridization or protocols; such data are critical in checking reproducibility and for comparing with other experiments. Motivation for a more sophisticated and rigorous approach to microarray analysis data has come from researchers in the Hormone Responsive Breast Cancer Network (HRBC, http://www.hrbcgenomics. net/ ), where the analysis and sharing of microarray data with other members of the project, as well as comparison with relevant data in public repositories, are essential requirements

Abstract: The antigen receptor repertoire is generated through rearrangements of Immunoglobulin (Ig) and T-cell receptor (TCR) gene segments into functional genes by a mechanism named V(D)J recombination. This is a two-step process: a cleavage step in which double-strand DNA cuts are made at specific sequences, followed by a joining step to repair the breaks. The coding segments of Ig and TCR genes are flanked by short recombination signal sequences (RSS). The RSS are recognized by a complex of the lymphocyte specific recombination proteins RAG1 and RAG2, which cleave the DNA between the coding sequence and the RSS. The broken coding strands are then rejoined to produce a rearranged gene. Mistakes in this process can generate chromosomal translocations that are involved in acute lymphoid leukemias and non-Hodgkin lymphomas. These errors include cleavage of DNA sequences, termed cryptic RSS, similar to functional RSS but located outside the Ig or TCR loci. We have screened the human and mouse genomes for the presence of putative cryptic RSS. To provide an initial list enriched in cryptic RSS, we used an original search algorithm. This primary set was then further filtered for biologically functional sequences using a published method. We have created a web-accessible database containing these putative recombination signals in the genome context. This is the first repository containing a genome-wide collection of RSS sequences. These sequence tags can be retrieved from a number of starting points including RSS type (with 12- or 23 bp spacer), chromosomal region, cytoband and gene identity. For visualization of our RSS search results, we have chosen to rely on an existing genome annotation knowledgebase and correlate our results with the gene structure, analysis, annotation and browsing features of the UCSC Genome Browser. Sequences of interest may also be searched for the occurrence of RSS and the corresponding tracks searched from within the genome database.

Abstract: The recent advent of high-throughput technology and the exponential increase in computer power have moved biology into a revolutionary mode, shifting the focus of molecular biologists from single genes to whole genomes The possibility of exploring gene function is extremely attractive in such a context of high-throughput data generation and computational inference based on similarities in gene expression has been proved to be a valuable tool for functional characterization. The modern theory of networks offers a new conceptual framework for the analysis of gene expression both at the transcriptomic and proteomic levels: genome-scale data sets can infact be conveniently visualized as networks of gene/protein co-occurrences where genes/proteins are represented by nodes and the relationships between them are represented by connections. This paper reports just an “in silico” approach to gene expression analysis based on a gene co-expression network.

Abstract: The transport of various metabolites across the mitochondrial membranes is essential for eukaryotic metabolism. Specific transport through the inner mitochondrial membrane is achieved by nuclear encoded carriers which form a large transport family, the mitochondrial carrier family. The structure of the ADP/ATP carrier in complex with its inhibitor carboxyatractyloside (CATR) has been recently solved by X-ray crystallography providing for the first time an insight into one conformation of the protein. In order to shed light on the possible conformation sampled by the protein and on the effect of CATR on constraining a definite configuration we have carried out two 10 ns molecular dynamics simulation of the protein embedded in a lipid bilayer of palmitoyl-oleoyl-phosphatidyl-choline (POPC) with and without its co-crystallized inhibitor CATR.

Abstract: Estrogen Receptor belongs to the Nuclear receptor family, there are two different currently known subtypes Era and Erb. Levels and proportion of the two subtypes differs in different target cells. We can hypotheses a different pharmacological activity upon ligand binding. The aim of this work is the investigation through molecular dynamics simulations, on both the subtypes a and b of the estrogen receptor, of possible differences among them and for the study of their dynamical behaviour.

Abstract: SH3 domains bind polyproline II peptides characterized by the PxxP consensus (P is proline and x in any amino acid). Single domain specificities display a preference for peptides within a range of variability on the common structural theme and different domains may interact with common peptides. We defined a new neural strategy to extract information from interacting partner sequences to improve the identification of SH3 domains specificity.

Abstract: Membrane proteins play key role in cell biology e.g. as ion channels, drug receptors, and solute transporters. It has been estimated that ~25% of genes code for membrane proteins, and that ca. 50% of potential new drug targets are membrane proteins. Despite the central importance of membrane proteins, the number of high resolution structures (from X-ray diffraction and more recently from NMR) remains small but the literature about experimental data available is huge. Literature gives a large amount of disjointed information about this essential group of proteins that needs to be organized to give a direct access to the researcher. In order to ease the browse of experimental data we are preparing the “membranome” site. Membranome site will select, store and efficiently organize literature data about: - classification; - genomic and protein sequences; - expression, purification, crystallization and structure determination; - structure and function; - transmembrane regions predictions (if the structure is not available); - interactions between membrane proteins and the rest of cell components: ions, lipids, sugars, ligands, substrates, solvent, a variety of molecules and other proteins; - mutants, mutation technique, altered functionality and pathological consequences of mutations; - publication references.

Abstract: Histone folds are structural elements that are able to form dimers by means of tight interactions between hydrophobic surfaces. Normally, a histone fold is composed by a long alpha-helix flanked by two or three shorter helices. In the nucleosome core particle, two pairs of H2a-H2b and H3-H4 histone heterodimers assemble together, giving rise to a disk-like octamer upon which DNA rolls up. The publication of the X-ray structure of the prokaryotic histone from Methanopyrus kandleri highlighted a novel protein fold, which is originated by the assembly of two consecutive histone folds included in the same peptide chain. More recently, the publication of the X-ray structure of the amino-terminal domain of hSos1 showed that also this protein module assumes a similar fold, dubbed the histone pseudodimer and here also referred to as “double histone fold”. In fact, the evolutionary relationship between the H2a histone and the domain spanning the protein sequence 96-190 of hSos1 had been already disclosed, due to high sequence similarity between the two domains. However, the first histone-like domain of hSos1, spanning the protein portion 6-95, does not show evident sequence similarity with histones (Sondermann et al, 2003). Moreover, it is unknown whether the double histone fold can be found in other protein families. In view of this, we initiated an in silico study aiming at the identification of other proteins characterized by a sequence that is compatible with the double histone fold.

Abstract: Much effort has been devoted to the detection, from multiple sequence alignments of homologous proteins, of pairs or groups of amino acid positions that evolve in a nonindependent – or compensatory – manner. It is expected that such clusters of positions might either tend either to be proximal in the mature, folded protein, or to be involved in similar aspects of protein function. Several such methods have been shown to be reasonably effective in the detection of intra-protein contacts. However, all of the most successful published algorithms rely on pairwise comparisons between aligned sequences. We wished to investigate whether evolutionary information (the topology and branchlengths of the phylogenetic tree describing relationships between the sequences under study) can allow an improvement in the prediction of intra protein contacts from correlated substitutions.

Abstract: Extracting the complete functional information encoded in a genome, including genic, regulatory and structural elements, is a central challenge in biological research. Prediction of non-protein-coding functional regions, such as regulatory elements, is especially difficult because they are usually short (6-15 bp for S.cerevisiae and many other eukaryotic genomes), often degenerate, and can reside on either strand of DNA at variable distances from the genes they control. Since functional sequences tend to be conserved through evolution, they can appear as ‘phylogenetic footprints’ in alignments of genome sequences of different species. Recently, two groups sequenced several Saccharomyces genomes. The main goal of these studies was to identify the regulatory sites in Saccharomyces spp. using multiple whole-genome alignment or and multiple alignments of gene upstream regions. Results were represented as two lists of predicted binding motifs. Our comparison of these lists shows a rather moderate intersection. This prompted us to analyze the conservation rate for known and predicted binding sites in Saccharomyces genomes in more detail.

Abstract: An increasing number of eukaryotic and prokaryotic genes are being found to have natural antisense transcripts (NATs). Also, there is a growing evidence to suggest that antisense transcription might have a key role in a range of human diseases. Consequently, there have been several recent attempts to set up computational procedures aimed at identifying novel NATs. Our group has developed the AntiHunter program for the identification of expressed sequence tag (EST) antisense transcripts from BLAST output.

Abstract: Since transcriptional control is the result of complex networks that interpret a variety of inputs, analyzing dynamical states of gene expression is of paramount importance to detect the multivariate nature of complex biological mechanisms. Although hundreds of studies fully demonstrated the relevancy of microarrays in describing the transcriptional status of different physiological conditions, to access and reconstruct complex interaction pathways it is necessary to analyze the temporal evolution of transcriptional states. However, an appropriate experimental design to accurately identify differentially expressed genes over a meaningful temporal window would require large amounts of microarrays and computational procedures able to assess the correlation structure among data at different time points. Unfortunately, replicates for each time point and experimental condition are not always available, because of cost limitations and/or biological sample scarcity, while common data analysis tools, e.g., ANOVA, do require replicates.