Antoine Mahul

Metavir 2: new tools for viral metagenome comparison and assembled virome analysis

BackgroundMetagenomics, based on culture-independent sequencing, is a well-fitted approach to provide insights into the composition, structure and dynamics of environmental viral communities. Following recent advances in sequencing technologies, new challenges arise for existing bioinformatic tools dedicated to viral metagenome (i.e. virome) analysis as (i) the number of viromes is rapidly growing and (ii) large genomic fragments can now be obtained by assembling the huge amount of sequence data generated for each metagenome.ResultsTo face these challenges, a new version of Metavir was developed. First, all Metavir tools have been adapted to support comparative analysis of viromes in order to improve the analysis of multiple datasets. In addition to the sequence comparison previously provided, viromes can now be compared through their k-mer frequencies, their taxonomic compositions, recruitment plots and phylogenetic trees containing sequences from different datasets. Second, a new section has been specifically designed to handle assembled viromes made of thousands of large genomic fragments (i.e. contigs). This section includes an annotation pipeline for uploaded viral contigs (gene prediction, similarity search against reference viral genomes and protein domains) and an extensive comparison between contigs and reference genomes. Contigs and their annotations can be explored on the website through specifically developed dynamic genomic maps and interactive networks.ConclusionsThe new features of Metavir 2 allow users to explore and analyze viromes composed of raw reads or assembled fragments through a set of adapted tools and a user-friendly interface.

Metavir: a web server dedicated to virome analysis

SUMMARY Metavir is a web server dedicated to the analysis of viral metagenomes (viromes). In addition to classical approaches for analyzing metagenomes (general sequence characteristics, taxonomic composition), new tools developed specifically for viral sequence analysis make it possible to: (i) explore viral diversity through automatically constructed phylogenies for selected marker genes, (ii) estimate gene richness through rarefaction curves and (iii) perform cross-comparison against other viromes using sequence similarities. Metavir is thus unique as a platform that allows a comprehensive virome analysis. AVAILABILITY Metavir is freely available online at: http://metavir-meb.univ-bpclermont.fr. CONTACT simon.roux@univ-bpclermont.fr. SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.

Annotation of microsporidian genomes using transcriptional signals

High-quality annotation of microsporidian genomes is essential for understanding the biological processes that govern the development of these parasites. Here we present an improved structural annotation method using transcriptional DNA signals. We apply this method to re-annotate four previously annotated genomes, which allow us to detect annotation errors and identify a significant number of unpredicted genes. We then annotate the newly sequenced genome of Anncaliia algerae. A comparative genomic analysis of A. algerae permits the identification of not only microsporidian core genes, but also potentially highly expressed genes encoding membrane-associated proteins, which represent good candidates involved in the spore architecture, the invasion process and the microsporidian-host relationships. Furthermore, we find that the ten-fold variation in microsporidian genome sizes is not due to gene number, size or complexity, but instead stems from the presence of transposable elements. Such elements, along with kinase regulatory pathways and specific transporters, appear to be key factors in microsporidian adaptive processes.

Evolutionary history of hepatitis C virus genotype 5a in France, a multicenter ANRS study

The epidemic history of HCV genotype 5a is poorly documented in France, where its prevalence is very low, except in a small central area, where it accounts for 14.2% of chronic hepatitis C cases. A Bayesian coalescent phylogenetic investigation based on the E1 envelope gene and a non-structural genomic segment (NS3/4) was carried out to trace the origin of this epidemic using a large sample of genotype 5a isolates collected throughout France. The dates of documented transmissions by blood transfusion were used to calibrate five nodes in the phylogeny. The results of the E1 gene analysis showed that the best-fitting population dynamic model was the expansion growth model under a relaxed molecular clock. The rate of nucleotide substitutions and time to the most recent common ancestors (tMRCA) of genotype 5a isolates were estimated. The divergence of all the French HCV genotype 5a strains included in this study was dated to 1939 [95% HPD: 1921-1956], and the tMRCA of isolates from central France was dated to 1954 [1942-1967], which is in agreement with epidemiological data. NS3/4 analysis provided similar estimates with strongly overlapping HPD values. Phylodynamic analyses give a plausible reconstruction of the evolutionary history of HCV genotype 5a in France, suggesting the concomitant roles of transfusion, iatrogenic route and intra-familial transmission in viral diffusion.

PhylOPDb: a 16S rRNA oligonucleotide probe database for prokaryotic identification

In recent years, high-throughput molecular tools have led to an exponential growth of available 16S rRNA gene sequences. Incorporating such data, molecular tools based on target-probe hybridization were developed to monitor microbial communities within complex environments. Unfortunately, only a few 16S rRNA gene-targeted probe collections were described. Here, we present PhylOPDb, an online resource for a comprehensive phylogenetic oligonucleotide probe database. PhylOPDb provides a convivial and easy-to-use web interface to browse both regular and explorative 16S rRNA-targeted probes. Such probes set or subset could be used to globally monitor known and unknown prokaryotic communities through various techniques including DNA microarrays, polymerase chain reaction (PCR), fluorescent in situ hybridization (FISH), targeted gene capture or in silico rapid sequence identification. PhylOPDb contains 74 003 25-mer probes targeting 2178 genera including Bacteria and Archaea. Database URL: http://g2im.u-clermont1.fr/phylopdb/

Queueing network modelling with distributed neural networks for service quality estimation in B-ISDN networks

We discuss a scheme based on distributed feedforward neural networks (NN), aimed at modelling several queueing systems in cascade fed with bursty traffic. For each queueing system, a neural network is trained to anticipate the average number of waiting packets, the packet loss rate and the coefficient of variation of the packet inter-departure time, given the mean rate, the peak rate and the coefficient of variation of the packet inter-arrival time. The latter serves for the calculation of the coefficient of variation of the cell inter-arrival time of the aggregated traffic which is fed as input to the next NN along the path. The potential of this method was sucessfully illustrated on several single server FIFO queues in Aussem et al. (1999). We now apply this technique to model a small queueing network made up from a combination of queues in tandem and in parallel fed by a superimposition of OnOff sources. Our long-term goal is the design of preventive control strategy in a multiservice communication network.

Designing generic algorithms for operations research

Design solutions have been proposed to implement generic data structures, however such techniques dedicated to algorithms are not well known. This article discusses various recurrent problems encountered when designing reusable, extensible algorithms for operations research. It explains how to use object‐oriented concepts and the notion of genericity to design algorithms that are independent of the data structures and the algorithms they use, but that can still interact deeply with them. An object‐oriented design is sometimes considered to be less efficient than a classical one, and operations research is one of these scientific fields where efficiency really matters. Hence, the main goal of this article is to explain how to design algorithms that are both generic and efficient. It also discusses specific recurring design issues for operations research software and proposes solutions that improve the genericity of the algorithms. Copyright

Distributed Neural Networks for Quality of Service Estimation in Communication Networks

We study an original scheme based on distributed feedforward neural networks, aimed at modelling several queueing systems in cascade fed with bursty traffic. For each queueing system, a neural network is trained to anticipate the average number of waiting packets, the packet loss rate and the coefficient of variation of the packet inter-departure time, given the mean rate, the peak rate and the coefficient of variation of the packet inter-arrival time. The latter serves for the calculation of the coefficient of variation of the cell inter-arrival time of the aggregated traffic which is fed as input to the next neural network along the path. The potential of this method is successfully illustrated on several single server FIFO (First In, First Out) queues and on small queueing networks made up from a combination of queues in tandem and in parallel fed by a superposition of ideal sources. Our long-term goal is the design of preventive control strategy in a multiservice communication network.

Probe Design Strategies for Oligonucleotide Microarrays.

Oligonucleotide microarrays have been widely used for gene detection and/or quantification of gene expression in various samples ranging from a single organism to a complex microbial assemblage. The success of a microarray experiment, however, strongly relies on the quality of designed probes. Consequently, probe design is of critical importance and therefore multiple parameters should be considered for each probe in order to ensure high specificity, sensitivity, and uniformity as well as potentially quantitative power. Moreover, to assess the complete gene repertoire of complex biological samples such as those studied in the field of microbial ecology, exploratory probe design strategies must be also implemented to target not-yet-described sequences. To design such probes, two algorithms, KASpOD and HiSpOD, have been developed and they are available via two user-friendly web services. Here, we describe the use of this software necessary for the design of highly effective probes especially in the context of microbial oligonucleotide microarrays by taking into account all the crucial parameters.

Large scale explorative oligonucleotide probe selection for thousands of genetic groups on a computing grid: application to phylogenetic probe design using a curated small subunit ribosomal RNA gene database.

Phylogenetic Oligonucleotide Arrays (POAs) were recently adapted for studying the huge microbial communities in a flexible and easy-to-use way. POA coupled with the use of explorative probes to detect the unknown part is now one of the most powerful approaches for a better understanding of microbial community functioning. However, the selection of probes remains a very difficult task. The rapid growth of environmental databases has led to an exponential increase of data to be managed for an efficient design. Consequently, the use of high performance computing facilities is mandatory. In this paper, we present an efficient parallelization method to select known and explorative oligonucleotide probes at large scale using computing grids. We implemented a software that generates and monitors thousands of jobs over the European Computing Grid Infrastructure (EGI). We also developed a new algorithm for the construction of a high-quality curated phylogenetic database to avoid erroneous design due to bad sequence affiliation. We present here the performance and statistics of our method on real biological datasets based on a phylogenetic prokaryotic database at the genus level and a complete design of about 20,000 probes for 2,069 genera of prokaryotes.