Pascal Neveu

PHENOPSIS DB: an information system for Arabidopsis thaliana phenotypic data in an environmental context.

BackgroundRenewed interest in plant × environment interactions has risen in the post-genomic era. In this context, high-throughput phenotyping platforms have been developed to create reproducible environmental scenarios in which the phenotypic responses of multiple genotypes can be analysed in a reproducible way. These platforms benefit hugely from the development of suitable databases for storage, sharing and analysis of the large amount of data collected. In the model plant Arabidopsis thaliana, most databases available to the scientific community contain data related to genetic and molecular biology and are characterised by an inadequacy in the description of plant developmental stages and experimental metadata such as environmental conditions. Our goal was to develop a comprehensive information system for sharing of the data collected in PHENOPSIS, an automated platform for Arabidopsis thaliana phenotyping, with the scientific community.DescriptionPHENOPSIS DB is a publicly available (URL: http://bioweb.supagro.inra.fr/phenopsis/) information system developed for storage, browsing and sharing of online data generated by the PHENOPSIS platform and offline data collected by experimenters and experimental metadata. It provides modules coupled to a Web interface for (i) the visualisation of environmental data of an experiment, (ii) the visualisation and statistical analysis of phenotypic data, and (iii) the analysis of Arabidopsis thaliana plant images.ConclusionsFirstly, data stored in the PHENOPSIS DB are of interest to the Arabidopsis thaliana community, particularly in allowing phenotypic meta-analyses directly linked to environmental conditions on which publications are still scarce. Secondly, data or image analysis modules can be downloaded from the Web interface for direct usage or as the basis for modifications according to new requirements. Finally, the structure of PHENOPSIS DB provides a useful template for the development of other similar databases related to genotype × environment interactions.

Bayesian nonlinear model selection and neural networks: a conjugate prior approach

In order to select the best predictive neural-network architecture in a set of several candidate networks, we propose a general Bayesian nonlinear regression model comparison procedure, based on the maximization of an expected utility criterion. This criterion selects the model under which the training set achieves the highest level of internal consistency, through the predictive probability distribution of each model. The density of this distribution is computed as the model posterior predictive density and is asymptotically approximated from the assumed Gaussian likelihood of the data set and the related conjugate prior density of the parameters. The use of such a conjugate prior allows the analytic calculation of the parameter posterior and predictive posterior densities, in an empirical-Bayes-like approach. This Bayesian selection procedure allows us to compare general nonlinear regression models and in particular feedforward neural networks, in addition to embedded models as usual with asymptotic comparison tests.

Milk acidification by Lactococcus lactis is improved by decreasing the level of dissolved oxygen rather than decreasing redox potential in the milk prior to inoculation

Although redox potential is very rarely taken into account in food fermentation it could be as influential as pH on bacterial activities. Lactococcus lactis is already known to exhibit a powerful reducing activity in milk but its reduction activity was shown to occur prior to its acidification activity with a potential interaction between these two lactococcal activities. Therefore, acidification lag-type phase could be shortened by decreasing the redox potential of milk before inoculation. As the redox potential is highly dependent on the dissolved oxygen level, our objective was to study their separate and combined influences on acidification and growth kinetics of pure L. lactis strains in milk. Results showed that high level of dissolved oxygen is significantly more influential on growth, and even more on acidification kinetics, than initial decreased redox potential of milk. Reduction of milk was drastic and mostly due to bacterial activity. The redox potential of milk only dropped when dissolved oxygen was entirely consumed. When there was no dissolved oxygen from the beginning, L. lactis immediately decreased the redox potential of milk and acidified afterwards. When the level of dissolved oxygen was initially high, acidification and reduction of milk occurred at the same time. Acidification kinetics was then biphasic with a slower rate during the aerobic stage and a faster rate during the anaerobic stage. The seven strains tested demonstrated diversity in both their acidification kinetics and their adaptation to high level of dissolved oxygen, independent of their growth kinetics. To conclude, we have shown that the level of dissolved oxygen in milk has a dramatic influence on acidification kinetics and could be used to control acidification kinetics in dairy industries.

Measures for interoperability of phenotypic data: minimum information requirements and formatting

BackgroundPlant phenotypic data shrouds a wealth of information which, when accurately analysed and linked to other data types, brings to light the knowledge about the mechanisms of life. As phenotyping is a field of research comprising manifold, diverse and time-consuming experiments, the findings can be fostered by reusing and combining existing datasets. Their correct interpretation, and thus replicability, comparability and interoperability, is possible provided that the collected observations are equipped with an adequate set of metadata. So far there have been no common standards governing phenotypic data description, which hampered data exchange and reuse.ResultsIn this paper we propose the guidelines for proper handling of the information about plant phenotyping experiments, in terms of both the recommended content of the description and its formatting. We provide a document called “Minimum Information About a Plant Phenotyping Experiment”, which specifies what information about each experiment should be given, and a Phenotyping Configuration for the ISA-Tab format, which allows to practically organise this information within a dataset. We provide examples of ISA-Tab-formatted phenotypic data, and a general description of a few systems where the recommendations have been implemented.ConclusionsAcceptance of the rules described in this paper by the plant phenotyping community will help to achieve findable, accessible, interoperable and reusable data.

InfraPhenoGrid: A scientific workflow infrastructure for Plant Phenomics on the Grid

Plant phenotyping consists in the observation of physical and biochemical traits of plant genotypes in response to environmental conditions. Challenges , in particular in context of climate change and food security, are numerous. High-throughput platforms have been introduced to observe the dynamic growth of a large number of plants in different environmental conditions. Instead of considering a few genotypes at a time (as it is the case when phenomic traits are measured manually), such platforms make it possible to use completely new kinds of approaches. However, the data sets produced by such widely instrumented platforms are huge, constantly augmenting and produced by increasingly complex experiments, reaching a point where distributed computation is mandatory to extract knowledge from data. In this paper, we introduce InfraPhenoGrid, the infrastructure we designed and deploy to efficiently manage data sets produced by the PhenoArch plant phenomics platform in the context of the French Phenome Project. Our solution consists in deploying scientific workflows on a Grid using a middle-ware to pilot workflow executions. Our approach is user-friendly in the sense that despite the intrinsic complexity of the infrastructure, running scientific workflows and understanding results obtained (using provenance information) is kept as simple as possible for end-users.

Towards an open grapevine information system

Viticulture, like other fields of agriculture, is currently facing important challenges that will be addressed only through sustained, dedicated and coordinated research. Although the methods used in biology have evolved tremendously in recent years and now involve the routine production of large data sets of varied nature, in many domains of study, including grapevine research, there is a need to improve the findability, accessibility, interoperability and reusability (FAIR-ness) of these data. Considering the heterogeneous nature of the data produced, the transnational nature of the scientific community and the experience gained elsewhere, we have formed an open working group, in the framework of the International Grapevine Genome Program (www.vitaceae.org), to construct a coordinated federation of information systems holding grapevine data distributed around the world, providing an integrated set of interfaces supporting advanced data modeling, rich semantic integration and the next generation of data mining tools. To achieve this goal, it will be critical to develop, implement and adopt appropriate standards for data annotation and formatting. The development of this system, the GrapeIS, linking genotypes to phenotypes, and scientific research to agronomical and oeneological data, should provide new insights into grape biology, and allow the development of new varieties to meet the challenges of biotic and abiotic stress, environmental change, and consumer demand.

Neural network architecture selection: new Bayesian perspectives in predictive modelling: Application to a soil hydrology problem

The aim of this paper is to present to the community of ecologists concerned with predictive modelling by feedforward neural network, a new statistical approach to select the best neural network architecture (number of layers, number of neurons per layer and connectivity) in a set of several candidate networks. The interest of this approach is demonstrated on a soil hydrology problem.

Ontology-Based Model for Food Transformation Processes - Application to Winemaking

This paper describes an ontology for modeling any food processing chain. It is intended for data and knowledge integration and sharing. The proposed ontology (Onto-FP) is built based on four main concepts: Product, Operation, Attribute and Observation. This ontology is able to represent food product transformations as well as temporal sequence of food processes. The Onto-FP can be easy integrated to other domains due to its consistencies with DOLCE ontology. We detail an application in the domain of winemaking and prove that it can be easy queried to answer questions related to data classification, food process itineraries and incomplete data identification.

Profile diversity in search and recommendation

We investigate profile diversity, a novel idea in searching scientic documents. Combining keyword relevance with popularity in a scoring function has been the subject of different forms of social relevance [2, 6, 9]. Content diversity has been thoroughly studied in search and advertising [4, 11], database queries [16, 5, 8], and recommendations [17, 10, 18]. We believe our work is the first to investigate profile diversity to address the problem of returning highly popular but too-focused documents. We show how to adapt Fagins threshold-based algorithms to return the most relevant and most popular documents that satisfy content and profile diversities and run preliminary experiments on two benchmarks to validate our scoring function.

Using ontologies of software: example of r functions management

In a scientific context, making available scientific resources like computer programs is a real challenge for multidisciplinary research teams. In this paper, we propose an ontology-based approach to manage, share, reuse and promote software programs in a research community. Specifically, we were interested in the capitalization of R functions, R being a language for statistics and graphics. We designed an ontology to annotate R functions. We adopted the Semantic Web models: annotations are represented into the Resource Description Framework and the ontology in the Ontology Web Language. In the so-built semantic repository, R functions can be retrieved by expressing semantic queries in the SPARQL language. As a result, we have developed a new kind of software repository with semantic inferences. It is based upon the Corese semantic search engine and accessible through a Web Service. It has been adopted by a multidisciplinary team in life sciences.