2023Activity reportProject-TeamDYLISS

3.2.1 Research topics

Facilitating data integration and querying The quantity and inner complexity of life science data require semantically-rich analysis methods. A major challenge is then to combine data (from local project as well as from reference databases) and symbolic knowledge seamlessly. Semantic Web technologies (RDF for annotating data, OWL for representing symbolic knowledge, and SPARQL for querying) provide a relevant framework, as demonstrated by the success of Linked (Open) Data  28. However, life science end users (1) find it difficult to learn the languages for representing and querying Semantic Web data, and consequently (2) miss the possibility they had to interact with their tabulated data (even when doing so was exceedingly slow and tedious). Our first objective in this axis is to develop accurate abstractions of datasets or knowledge repositories to facilitate their exploration with RDF-based technologies.

Scalability of semantic web queries. A bottleneck in data querying is given by the performance of federated SPARQL queries, which must be improved by several orders of magnitude to allow current massive data to be analyzed. In this direction, our research program focuses on the combination of linked data fragments  67, query properties and dataset structure for decomposing federated SPARQL queries.

Building and compressing static maps of interacting compounds A final approach to handle heterogeneity is to gather multi-scale data knowledge into a functional static map of biological models that can be analyzed and/or compressed. This requires to link genomics, metabolomics, expression data and protein measurement of several phenotypes into unified frameworks. In this direction, our main goal is to develop families of constraints, inspired by symbolic dynamical systems, to link datasets together. We currently focus on health (personalized medicine) and environmental (role of non-coding regulations, graph compression) datasets.

3.2.2 Associated software tools

AskOmics platform AskOmics is an integration and interrogation software for linked biological data based on semantic web technologies1. AskOmics aims at bridging the gap between end user data and the Linked (Open) Data cloud (LOD cloud). It allows heterogeneous bioinformatics data (formatted as tabular files or directly in RDF) to be loaded into a Triple Store system using a user-friendly web interface. It helps end users (1) to take advantage of the information available in the LOD cloud for analyzing their own data, and (2) to contribute back to the linked data by representing their data and the associated metadata in the proper format, as well as by linking them to other resources. An originality is the graphical interface that allows any dataset to be integrated in a local RDF datawarehouse and SPARQL query to be built transparently and iteratively by a non-expert user.

Pax2graphml aims at easily manipulating BioPAX source files as regulated reaction graphs described in graph format. The goal is to be highly flexible and to integrate graphs of regulated reactions from a single BioPAX source or by combining and filtering BioPAX sources. The output graphs can then be analyzed with additional tools developed in the team, such as KeyRegulatorFinder.

FinGoc-tools The FinGoc tools allow filtering interaction networks with graph-based optimization criteria in order to elucidate the main regulators of an observed phenotype. The main added-value of these tools is the functionality allowing to make explicit the criteria used to highlight the role of the main regulators. (1) The KeyRegulatorFinder package searches key regulators of lists of molecules (like metabolites, enzymes or genes) by taking advantage of knowledge databases in cell metabolism and signaling2. (2) The PowerGrasp python package implements graph compression methods oriented toward visualization, and based on power graph analysis3. (3) The iggy package enables the repairing of an interaction graph with respect to expression data4.

3.3.1 Research topics

Genomic level: characterizing functions of protein sequences Precise characterization of functional proteins, such as enzymes or transporters, is a key to better understand and predict the actors involved in a metabolic process. In order to improve the precision of functional annotations, we develop machine learning approaches that take a sample of functional sequences as input and infer a model representing their key syntactical characteristics, including dependencies between residues.

System level: enriching and comparing metabolic networks for non-model organisms

Non-model organisms often lack both complete and reliable annotated sequences, which cause the draft networks of their metabolism to largely suffer from incompleteness. In former studies, the team has developed several methods to improve the quality of eukaryotic metabolic networks, by solving several variants of the so-called Metabolic Network gap-filling problem with logical programming approaches 9, 8. The main drawback of these approaches is that they cannot scale to the reconstruction and comparison of families of metabolic networks. Our main objective is therefore to develop new tools for the comparison of species strains at the metabolic level.

Consortium level: exploring the diversity of community consortia The newly emerging field of system ecology aims at building predictive models of species interactions within an ecosystem, with the goal of deciphering cooperative and competitive relationships between species  46. This field raises two new issues: (1) uncertainty on the species present in the ecosystem and (2) uncertainty about the global objective governing an ecosystem. To address these challenges, our first research focus is the inference of metabolic exchanges and relationships for transporter identification, based on our expertise in metabolic network gap-filling. The second challenging focus is the prediction of transporters families via refined characterization of transporters, which are quite unexplored apart from specific databases  59.

3.3.2 Associated software tools

Protomata5 is a machine learning suite for the inference of automata characterizing (functional) families of proteins at the sequence level. It provides programs to build a new kind of sequence alignments (characterized as partial and local), learn automata, and search for new family members in sequence databases. By enabling to model local dependencies between positions, automata are more expressive than classical tools (PSSMs, Profile HMMs, or Prosite Patterns) and are well suited to predict new family members with a high specificity. This suite is for instance embedded in the cyanolase database  35 to automate its updade and was used for refining the classification of HAD enzymes 6 or identify shared conservations in the core proteome of extracellular vesicles produced by human and animal S. aureus strains  64.

PPSuite6 is one of the first frameworks taking into account coevolutionary dependencies between residues for the comparison of protein sequences. It proposes a complete workflow enabling to infer direct couplings between the positions of a sequence of interest by a Potts model with the help of the sequence close homologs and to score the similarity of the sequences by alignment of the inferred Potts models, as well as tools to visualize the models and their alignments  63, 62.

AuReMe and AuCoMe workspaces is designed for tractable reconstruction of metabolic networks7. The toolbox allows for the Automatic Reconstruction of Metabolic networks based on the combination of multiple heterogeneous data and knowledge sources 1. The main added values are the inclusion of graph-based tools relevant for the study of non-model organisms (Meneco and Menetools packages), the possibility to trace the reconstruction and curation procedures (Padmet package), and the exploration of reconstructed metabolic networks with wikis (wiki-export package, see: aureme.genouest.org/wiki.html)  11. It also generates outputs to explore the resulting networks with Askomics. It has been used for reconstructing metabolic networks of micro and macro-algae  57, extremophile bacteria  39 and communities of organisms 4.

Mpwt, emmapper2gbk is a Python package for running Pathway Tools8 on multiple genomes using multiprocessing. Pathway Tools is a comprehensive systems biology software system that is associated with the BioCyc database collection9. Pathway Tools is frequently used for reconstructing metabolic networks. In order to allow the output of the eggnoggmapper annotation tool to be used by Mpwt, we also developed emmaper2gbk to create relevant genome files.

Metage2metabo is a Python tool to perform graph-based metabolic analysis starting from annotated genomes (reference genomes or metagenome-assembled genomes)  26. It uses Mpwt to reconstruct metabolic networks for a large number of genomes. The obtained metabolic networks are then analyzed individually and collectively in order to get the added value of metabolic cooperation in microbiota over individual metabolism and to identify and screen interesting organisms among all.

3.4.1 Research topics

Genomic level: characterizing gene structure with grammatical languages and conservation information The goal here is to accurately represent gene structure, including intron/exon structure, for predicting the products of genes, such as isoform transcripts, and comparing the expression potential of a eukaryotic gene according to its context (e.g. tissue) or according to the species. Our approach consists in designing grammatical and comparative-genomics based models for gene structures able to detect heterogeneous functional sites (splicing sites, regulatory binding sites...), functional regions (exons, promotors...) and global constraints (translation into proteins)  30. Accurate gene models are defined by identifying general constraints shaping gene families and their structures conserved over evolution. Syntactic elements controlling gene expression (transcription factor binding sites controlling transcription; enhancers and silencers controlling splicing events...), i.e. short, degenerated and overlapping functional sequences, are modeled by relying on the high capability of SVG grammars to deal with structure and ambiguity  60.

System level: extracting causal signatures of complex phenotypes with systems biology frameworks Our main challenge is to set up a generic formalism to model inter-layer interactions in large-scale biological networks. To that goal, we have developed several types of abstractions: multi-experiments framework to learn and control signaling networks 10, multi-layer reactions in interaction graphs  31, and multi-layer information in large-scale Petri nets  25. Our main issues are to scale these approaches to standardized large-scale repositories by relying on the interoperable Linked Open Data (LOD) resources and to enrich them with ad-hoc regulations extracted from sequence-based analysis. This will allow us to characterize changes in system attractors induced by mutations and how they may be included in pathology signatures.

3.4.2 Associated software tools

Logol software is designed for complex pattern modeling and matching10. It is a swiss-army-knife for pattern matching on DNA/RNA/Protein sequences, based on expressive patterns which consist in a complex combination of motifs (such as degenerated strings) and structures (such as imperfect stem-loop ou repeats) 2. Logol key features are the possibilities (i) to divide a pattern description into several sub-patterns, (ii) to model long range dependencies, and (iii) to enable the use of ambiguous models or to permit the inclusion of negative conditions in a pattern definition. Therefore, Logol encompasses most of the features of specialized tools (Vmatch, Patmatch, Cutadapt, HMM) and enables interplays between several classes of patterns (motifs and structures), including stem-loop identification in CRISPR.

Caspo Cell ASP Optimizer (Caspo) software constitutes a pipeline for automated reasoning on logical signaling networks (learning, classifying, designing experimental perturbations, identifying controllers, take time-series into account)11. The software handles inherent experimental noise by enumerating all different logical networks which are compatible with a set of experimental observations 10. The main advantage is that it enables a complete study of logical network without requiring any linear constraint programs.

Cadbiom package aims at building and analyzing the asynchronous dynamics of enriched logical networks12. It is based on Guarded transition semantic and allows synchronization events to be investigated in large-scale biological networks  25. For example, it allowed to analyze controler of phenotypes in a large-scale knowledge database (PID)  5.

Recently, we have significantly refactored Cadbiom package towards a framework that allows the identification of causal regulators in large-scale models, formalized in the BioPAX language and automatically interpreted as guarded transitions. The Cadbiom framework was applied to the BioPAX version of two ressources (PID, KEGG) of the PathwayCommons database and to the Atlas of Cancer Signalling Network (ACSN). As a case-study, it was used to characterize the causal signatures of markers of the epithelial-mesenchymal transition.

7.2.1 AskOmics

• Name:
  Convert tabulated data into RDF and create SPARQL queries intuitively and "on the fly".
• Keywords:
  RDF, SPARQL, Querying, Graph, LOD - Linked open data
• Functional Description:
  AskOmics aims at bridging the gap between end user data and the Linked (Open) Data cloud. It allows heterogeneous bioinformatics data (formatted as tabular files) to be loaded in a RDF triplestore and then be transparently and interactively queried. AskOmics is made of three software blocks: (1) a web interface for data import, allowing the creation of a local triplestore from user's datasheets and standard data, (2) an interactive web interface allowing "à la carte" query-building, (3) a server performing interactions with local and distant triplestores (queries execution, management of users parameters).
• URL:
  https://­askomics.­org/
• Authors:
  Charles Bettembourg, Xavier Garnier, Anthony Bretaudeau, Fabrice Legeai, Olivier Dameron, Olivier Filangi, Yvanne Chaussin, Mateo Boudet
• Contact:
  Olivier Dameron
• Partners:
  Université de Rennes 1, CNRS, INRA

7.2.2 Metage2Metabo

• Keywords:
  Metabolic networks, Microbiota, Metagenomics, Workflow
• Scientific Description:
  Flexible pipeline for the metabolic screening of large scale microbial communities described by reference genomes or metagenome-assembled genomes. The pipeline comprises several main steps. (1) Automatic and parallel reconstruction of metabolic networks. (2) Computation of individual metabolic potentials (3) Computation of collective metabolic potential (4) Calculation of the cooperation potential described as the set of metabolites producible by species only in a cooperative context (5) Computation of minimal-sized communities sastifying a metabolic objective (6) Extraction of key species (essential and alternative symbionts) associated to a metabolic function
• Functional Description:
  Metabolic networks are graphs which nodes are compounds and edges are biochemical reactions. To study the metabolic capabilities of microbiota, Metage2Metabo uses multiprocessing to reconstruct metabolic networks at large-scale. The individual and collective metabolic capabilities (number of compounds producible) are computed and compared. From these comparisons, a set of compounds only producible by the community is created. These newly producible compounds are used to find minimal communities that can produce them. From these communities, the keytstone species in the production of these compounds are identified.
• URL:
  https://­github.­com/­AuReMe/­metage2metabo
• Publication:
  hal-02395024
• Contact:
  Clemence Frioux
• Participants:
  Clemence Frioux, Arnaud Belcour, Anne Siegel

7.2.3 CADBIOM

• Name:
  Computer Aided Design of Biological Models
• Keywords:
  Health, Biology, Biotechnology, Bioinformatics, Systems Biology
• Functional Description:
  The Cadbiom software provides a formal framework to help the modeling of biological systems such as cell signaling network with Guarder Transition Semantics. It allows synchronization events to be investigated in biological networks among large-scale network in order to extract signature of controllers of a phenotype. Three modules are composing Cadbiom. 1) The Cadbiom graphical interface is useful to build and study moderate size models. It provides exploration, simulation and checking. For large-scale models, Cadbiom also allows to focus on specific nodes of interest. 2) The Cadbiom API allows a model to be loaded, performing static analysis and checking temporal properties on a finite horizon in the future or in the past. 3) Exploring large-scale knowledge repositories, since the translations of the large-scale PID repository (about 10,000 curated interactions) have been translated into the Cadbiom formalism.
• News of the Year:
  We have significantly refactored Cadbiom package towards a framework that allows the identification of causal regulators in large-scale models, formalized in the BioPAX language and automatically interpreted as guarded transitions.
• URL:
  http://­cadbiom.­genouest.­org
• Publications:
  inserm-00978313, hal-01242893, hal-01559249, hal-03693653
• Contact:
  Anne Siegel
• Participants:
  Geoffroy Andrieux, Michel Le Borgne, Nathalie Theret, Nolwenn Le Meur, Pierre Vignet, Anne Siegel

7.2.4 pax2graphml

• Name:
  pax2graphml - Large-scale Regulation Network in Python using BIOPAX and Graphml
• Keyword:
  Bioinformatics
• Functional Description:
  PAX2GRAPHML is an open source python library that allows to easily manipulate BioPAX source files as regulated reaction graphs described in .graphml format. PAX2GRAPHML is highly flexible and allows generating graphs of regulated reactions from a single BioPAX source or by combining and filtering BioPAX sources. Supporting the graph exchange format .graphml, the large-scale graphs produced from one or more data sources can be further analyzed with PAX2GRAPHML or standard python and R graph libraries.
• URL:
  https://­pax2graphml.­genouest.­org/
• Publication:
  hal-03265223
• Contact:
  Francois Moreews
• Partner:
  INRAE

7.2.5 Protomata

• Keywords:
  Proteins, Machine learning, Pattern discovery, Grammatical Inference, Bioinformatics
• Scientific Description:
  Inference of automata modelling protein sequences by partial local alignment
• Functional Description:

  This tool is a grammatical inference framework suitable for learning the specific signature of a functional protein family from unaligned sequences by partial and local multiple alignment and automata modelling. It performs a syntactic characterization of proteins by identification of conservation blocks on sequence subsets and modelling of their succession. Possible fields of application are new members discovery or study (for instance, for site-directed mutagenesis) of, possibly non-homologous, functional families and subfamilies such as enzymatic, signalling or transporting proteins.

  Given a sample of sequences belonging to a structural or functional family of proteins, Protomata-Learner infers an automaton characterizing the family by partial local alignment of the sequences. Automata are graphical models representing a (potentially infinite) set of sequences. Able to express alternative local dependencies between the positions, automata offer a finer level of expressivity than classical sequence patterns (such as PSSM, Profile HMM, or Prosite Patterns) and can model more than homologous sequences. They are well suited to get new insights into a family or to search for new family members in the sequence data banks, especially when approaches based on classical multiple sequence alignments are insufficient.

  The three main modules integrated in the Protomata-learner workflow are available as well as stand-alone programs: 1) paloma builds partial local multiple alignments, 2) protobuild infers automata from these alignements and 3) protomatch and protoalign scans, parses and aligns new sequences with learnt automata. The suite is completed by tools to handle or visualize data and can be used online by the biologists via a web interface on Genouest Platform.

• News of the Year:
  paloma-2 was renamed paloma-d to avoid confusion with version numbering.
• URL:
  http://­tools.­genouest.­org/­tools/­protomata/
• Contact:
  François Coste
• Participant:
  François Coste
• Partners:
  Université de Rennes 1, CNRS, Inria

7.2.6 PPsuite

• Keywords:
  Proteins, Sequence alignment, Bioinformatics, Machine learning, Homology search
• Scientific Description:
  Comparison of protein sequences using coevolutionary dependencies between residues.
• Functional Description:
  This suite contains the following tools : - MakePotts infers a Potts model from a sequence or a multiple sequence alignment - PPalign aligns Potts models and corresponding sequences - VizPotts allows to visualize inferred Potts models and VizContacts allows to visualize inferred couplings with respect to actual contacts in a 3D protein structure.
• URL:
  https://­www-dyliss.­irisa.­fr/­ppalign/
• Publications:
  hal-02402646, hal-02862213, hal-03264248, hal-03926272
• Contact:
  François Coste
• Participants:
  François Coste, Hugo Talibart, Mathilde Carpentier

7.2.7 Transformer Framework for Protein Characterization

• Keywords:
  Deep learning, Transformer, Functional annotation, Proteins, Biological sequences
• Scientific Description:
  A generic framework for the specialization of a pre-trained transformer protein language model for classification or regression tasks.
• Functional Description:
  Given examples of annotated sequences, this tool allows to train and analyse resulting models with respect to evaluation metrics (accuracy, correlation) plots and the importance of the residues for the inference. The process is fully automated and the whole operation can be done by modifying a JSON configuration file and providing a JSON data set. No code skills are thus required.
• News of the Year:
  Added instructions and scripts for reproducibility of the experiments from our publication on the prediction of the enzymatic function of sequence with EnzBert.
• URL:
  https://­gitlab.­inria.­fr/­nbuton/­tfpc
• Publications:
  tel-04347632, hal-04382475
• Contact:
  Nicolas Buton
• Participants:
  Nicolas Buton, François Coste, Yann Le Cunff

7.2.8 Emapper2GBK

• Keywords:
  Bioinformatics, Metabolic networks, Functional annotation
• Functional Description:
  Starting from FASTA and Eggnog-mapper annotation files, Emapper2GBK builds a GBK file that is suitable for metabolic network reconstruction with Pathway Tools, and adds the GO terms and EC numbers annotations in the GenBank file.
• URL:
  https://­github.­com/­AuReMe/­emapper2gbk
• Publication:
  hal-02395024
• Contact:
  Clemence Frioux
• Participants:
  Clemence Frioux, Arnaud Belcour, Anne Siegel

7.2.9 AuCoMe

• Name:
  Automatic Comparison of Metabolisms
• Keywords:
  Bioinformatics, Workflow, Metabolic networks, Omic data, Data analysis
• Functional Description:
  AuCoMe is a Python package that aims at reconstructing homogeneous metabolic networks and pan-metabolism starting from genomes with heterogeneous levels of annotations. Four steps are composing AuCoMe. 1) It automatically infers annotated genomes from draft metabolic networks thanks to Pathway Tools and MPWT. 2) The Gene-Protein-Reaction (GPR) associations previously obtained are propagated to protein orthogroups in using Orthofinder and, an additional robustness criteria. 3) AuCoMe checking the presence of supplementary GPR associations by finding missing annotation in all genomes. In this step, the tools BlastP, TblastN and, Exonerate are called. 4) It adding spontaneous reactions to metabolic pathways that were completed by the previous steps. AuCoMe generates several outputs to facilitate the analysis of results: tabuled files, SBML files, PADMET files, supervenn and a dendogram of reactions.
• URL:
  https://­github.­com/­AuReMe/­aucome
• Publication:
  hal-03778267
• Contact:
  Anne Siegel
• Participants:
  Arnaud Belcour, Jeanne Got, Meziane Aite, Ludovic Delage, Jonas Collen, Clemence Frioux, Catherine Leblanc, Simon M. Dittami, Samuel Blanquart, Gabriel V. Markov, Anne Siegel

7.2.10 mpwt

• Keywords:
  Metabolic networks, Multi-processor
• Functional Description:
  mpwt is a Python package for running Pathway Tools on multiple genomes using multiprocessing. More precisely, it launches one PathoLogic process for each organism. This allows to increase the speed of draft metabolic network reconstruction when working on multiple organisms.
• Publication:
  hal-02395024
• Contact:
  Anne Siegel
• Participants:
  Arnaud Belcour, Anne Siegel, Clemence Frioux, Meziane Aite

9.1.1 Associate Teams in the framework of an Inria International Lab or in the framework of an Inria International Program

Participants: Anne Siegel.

9.2.1 Visits to international teams

9.3.1 Other european programs/initiatives

DeepImpact : Deciphering plant-microbiome interactions to enhance crop defense to bioagressors

Participants: Samuel Blanquart, Olivier Dameron, Jeanne Got, Victor Mataigne, Pauline Hamon-Giraud, Anne Siegel.

DEEP IMPACT is a multidisciplinary consortium-based project that aims at combining ecology, biology, plant genetics and mathematics to identify, characterize and validate the microbial communities, plant communities and abiotic factors (including agricultural managements) explaining variation in Brassica napus and Triticum aestivum resistance to several pests. For this, we will start from an in situ approach by characterizing 100 fields (50 for each crop species) for both habitat (climatic and edaphic variables) and biotic (microbiota, virome, weed communities, pest attacks and pathobiota prevalence) features. Information from this broad characterization will be integrated into sparse and correlative statistical models to describe the relative part of the variance explained by both habitat and biotic features and correlated with a reduction of pest's attacks. This analysis will allow us to identify a combination of microbial species and soils, correlated with an increase of crop's resistance to pests. These microbial consortia will be isolated by taking advantages of newly developed culturomics methods and characterized by both whole genome sequencing and biochemical assays. Synthetic Consortia (SynComs) will be reconstructed to test their efficacy on a broad range of pests attacking both crops. 2021–2026. Dyliss grant: 176k€.

SEABIOZ : Potential microbial origins of the biostimulant properties of extracts from a brown algae holobinte

Participants: Samuel Blanquart, Olivier Dameron, Jeanne Got, Anne Siegel.

For sustainable agriculture, new bio-based solutions include biocontrol and the use of plant biostimulants such as aqueous seaweed extracts. The most widely exploited biomass for biostimulant production is the brown seaweed Ascophyllum nodosum and its commercial extracts, including products from the Roullier Group, have demonstrated their ability to improve plant growth and mitigate certain abiotic and biotic stresses. A unique feature of the alga is its mutualistic association with the fungal endophyte Mycophycias ascophylli and other microbes constituting an holobiont. Many questions remain as to the nature and origin of the active compounds in algal extracts. Are these bioactive metabolites produced by the host or by its microbiota? The main objective of SEABIOZ is to answer these questions by combining a multi-omics approach and systems biology. 2021–2024. Dyliss grant: 120k€.

IDEALG (ANR/PIA-Biotechnology and Bioresource)

Participants: Arnaud Belcour, François Coste, Jeanne Got, Anne Siegel.

The project gathers 18 partners from Station Biologique de Roscoff (coordinator), CNRS, IFREMER, UEB, UBO, UBS, ENSCR, University of Nantes, INRA, AgroCampus, and the industrial field in order to foster biotechnology applications within the seaweed field. Dyliss is co-leader of the WP related to the establishment of a virtual platform for integrating omics studies on seaweed and the integrative analysis of seaweed metabolism. Major objectives are the building of brown algae metabolic maps, metabolic flux analysis and the selection of symbiotic bacteria for brown algae. We will also contribute to the prediction of specific enzymes (sulfatases and haloacid dehalogenase)14. 2012–2021. Total grant: 11M€. Dyliss grant: 534k€.

ENDOVIRE (ANR)

Participants: Emmanuelle Becker, Olivier Dameron, Yael Tirlet-Greiner.

The whole ANR project gathers 4 partners : the BIPAA platform (INRAe), the DGIMI laboratory, the BF2I laboratory and the Dyliss team of IRISA. The project is focused about the understanding of how genes of a endogeneized viral genome in a parasitoid wasp are the activated and regulated. The available data produced by the consortium will cover genomics, epigenomics, pathways, regulation and orthology. We will contribute to identify the key actors involved in the activation of parasitoids genes, to propose a data and knowledge integration framework for the data of the global project, and to develop integrative data analysis methods for elucidating the mechanism involving the key actors identified in the first point. It will consist in proposing a library of queries (which contains a reasoning part), and further to propose regulation mechanisms based on heterogeneous -omics data across interacting organisms. To tackle the different challenges, our appoach will be based on (1) adequate statistical analysis workflows or methods, (2) Semantic Web technologies and AskOmics developed within the team, (3) knowledge-guided traversal strategies across multiplex graphs.

2023–2025. Total grant: 630k€. Dyliss grant: 176k€.

9.4.1 Programs funded by Inria

10.1.1 Scientific events: organisation

10.1.2 Scientific events: selection

10.1.3 Journal

10.1.4 Invited talks

• Seminar of the CNB-CSIC lab, In silico screening of metabolic functions of large-scale genomes and/or OTU : towards comparative metabolism?, Madrid, December 2023 [A. Siegel]
• École d’été interdisciplinaire en numérique de la santé (EINS 2023), 10–14th July Univ. Sherbrooke Canada. « Intégration et interrogation avancées de données et de connaissances grâce au Web Sémantique » [O. Dameron]
• Journée de lancement du PEPR Atlasea Des génomes marins à la synthèse in silico et in vivo de molécules, Paris, Janvier 2023 [A. Siegel]
• Colloque Ferment IA. Étude Et Comparaison De Cartes Métaboliques De Centaines De Souches Bactériennes. Vers La Prédiction De Cocktails Produisant Des Composés Cibles ?, Saclay, Septembre 2023 [A. Siegel]
• Journée du GDR BIOSS, Étude Et Comparaison De Cartes Métaboliques : s’adapter au triptyque connaissances – données – passage à l’échelle, Lille, Novembre 2023 [A. Siegel]
• Institut Français de Bioinformatique : École thématique sur la Bioinformatique Intégrative (ETBII, 16th–20th January, Fréjus) : animateur du thème « Web Sémantique » [O. Dameron]
• Journées du Réseau Inrae PEPI IBIS (15 septembre 2023, Paris): Predicting enzymatic function of protein sequences with attention [F. Coste]

10.1.5 Leadership within the scientific community

10.1.6 Scientific expertise

10.1.7 Research administration

10.2.1 Teaching

• Master : E. Becker, "Rappels de Statistiques", 15h, Master 1 in Bioinformatics, Univ. Rennes, France
• Master : E. Becker, "Short Introduction to R", 31h, Master 1 in Bioinformatics, Master 1 in Ecology and Envirronment, Univ. Rennes, France
• Master : E. Becker, "Data Visualisation and Manipulation", 12h, Master 1 in Bioinformatics, Univ. Rennes, France
• Master : E. Becker, "Object oriented programming", 30h, Master in Bioinformatics, Univ. Rennes 1, France
• Master : E. Becker, "Method", 15h, Master 2 in Computer Sciences, Univ. Rennes, France
• Master : E. Becker, "Advanced Python Programming", 38h, Master 2 in Bioinformatics, Univ. Rennes, France
• Master : E. Becker, "Biological netorks", 28h, Master 2 in Bioinformatics, Univ. Rennes, France
• Master : E. Becker, "Introduction to Bioinformatics", 3h, Master MEEF Biology, Univ. Rennes, France.
• Master : E. Becker, "Python for Life Sciences 1", 21h, Master minor in Bioinformatics, Univ. Rennes, France
• Master : E. Becker, "Insertion Professionnelle et tables rondes", 4h, Master 1 and Master 2 in Bioinformatics, Univ. Rennes, France
• Master: C. Belleannée, 18h, responsable des stages, Master 1 Ingenierie logicielle et Master 1 Cloud et Réseaux, Univ. Rennes, France
• Licence: C. Belleannée, responsable de la 2eme année du parcours Oui-si du portail ISTN, Univ. Rennes, France
• Licence: C. Belleannée, responsable de la Journée Portes Ouvertes pour l'Istic, Univ. Rennes, France
• Licence: C. Belleannée, Enseignante référente, 10h, L1 informatique, Univ. Rennes, France
• Licence: C. Belleannée, Langages formels, 17h, L3 informatique, Univ. Rennes, France.
• Licence: C. Belleannée, Projet professionnel et communication, 16h, L1 informatique, Univ. Rennes, France
• Licence: C. Belleannée, Spécialité informatique : Functional and immutable programming , 44h, L1 informatique, Univ. Rennes, France
• Master: C. Belleannée, Algorithmique du texte et bioinformatique, 10h, M1 informatique, Univ. Rennes, France
• Master: C. Belleannée, Programmation logique et contraintes, 32h, M1 informatique, Univ. Rennes, France
• Licence: C. Belleannée, Données structurées et bases de données, 20h, L2 informatique, Univ. Rennes, France
• Licence: C. Belleannée, Outils formels pour l'informatique, 24h, L2 informatique, Univ. Rennes, France
• Licence: C. Belleannée, Fondements mathématiques, 49h, L1 informatique, Univ. Rennes, France
• Master: F. Coste, Supervised machine learning, 10h, M2 Science Informatique, Univ. Rennes, France
• Licence: O. Dameron, "Programmation 1", 98h, Licence 1 informatique, Univ. Rennes, France
• Licence: O. Dameron, "Complément informatique", 24h, Licence 1 informatique, Univ. Rennes, France
• Licence: O. Dameron, "Data analysis and statistics", 24h, Licence 2 informatique, Univ. Rennes, France
• Licence: O. Dameron, "Graph Modeling and Algorithms", 21h, Licence 2 informatique, Univ. Rennes, France
• Licence: O. Dameron, "Programmation", 36h, Licence 3 miage, Univ. Rennes, France
• Master: O. Dameron, "Semantic Web", 10h, Master 1 miage, Univ. Rennes 1, France
• Master: O. Dameron, 2h, "Internship", Master 1 in bioinformatics, Univ. Rennes 1, France
• Master: O. Dameron, 20h, "Data Engineering in Life Science", Master 2 in bioinformatics, Univ. Rennes 1, France
• Master: O. Dameron, 28h, "Internship", Master 2 in bioinformatics, Univ. Rennes 1, France
• Doctoral level: A. Siegel, Ecole jeunes chercheurs "Modélisation des systèmes biologiques", Porquerolles, France.
• Licence : Y. Le Cunff "Modélisation des phénomènes du vivant", 30h, L2 Biologie, Univ. Rennes 1, France
• Master: Y. Le Cunff, "Apprentissage statistique", 110h, Master 1 in Bioinfortmatics Univ. Rennes 1, France
• Master: Y. Le Cunff, "Biologie aux interfaces", 25h, Master 1 in Biology, Univ. Rennes 1, France
• Master: Y. Le Cunff,"Simulating dynamic systems in biology", Master 2 in bioinformatics, 20h, Univ. Rennes 1, France
• Master: Y. Le Cunff, "Applied Interdisciplinarity", 20h, Master 2 in biology, Univ. Rennes 1, France
• PhD program: Y. Le Cunff, "Introduction to Machine Learning", 20h, FdV PhD Program, Sorbonne Paris Université, Paris, France

10.2.2 Supervision

10.2.3 Doctoral advisory committees (CSID)

• Yvon Awuklu, Université de Bordeaux [O. Dameron]
• Tiphaine Casy, Université de Rennes [O. Dameron]
• Guénolé Dande, Université de Rennes [O. Dameron]
• Simon Brocard, Nantes Université [F. Coste]
• Guillaume Doré, Université de Rennes [E. Becker]

10.2.4 Juries

10.3.1 Articles and contents

• Trousse à projet, Territoire du numérique éducatif "le numérique, c’est que pour les garçons?" tne.trousseaprojets.fr/professionnel-education-nationale/idees-recues-sur-le-numerique

10.3.2 Interventions

• Trousse à projet, Territoire du numérique éducatif "le numérique, c’est que pour les garçons?"
• A la découverte de la recherche, "Informatique pour la biologie et la santé" (Lycée Assomption Rennes, 5 avril 2023) [E. Becker]
• "How will we learn with next-gen AI ?" (Learning Planete Institute, Paris, France) [Y. Le Cunff]
• Table ronde "Intelligence artificielle" (CSIESR, Toulon, France) [Y. Le Cunff]
• "Impact de l'IA sur l'enseignement", Interview in Journal du Dimanche [Y. Le Cunff]
• Conference "Les nouvelles écritures et l'IA" (Journées nationales de l'ANSTIA, Rennes, France) [Y. Le Cunff]
• "enjeux sociétaux de l'IA" (SciencesPo Rennes et élus du département) [Y. Le Cunff]

International journals

• 11 articleA.Arnaud Belcour, J.Jeanne Got, M.Méziane Aite, L.Ludovic Delage, J.Jonas Collén, C.Clémence Frioux, C.Catherine Leblanc, S. M.Simon M Dittami, S.Samuel Blanquart, G. V.Gabriel V. Markov and A.Anne Siegel. Inferring and comparing metabolism across heterogeneous sets of annotated genomes using AuCoMe.Genome Research33June 2023, 972 - 987HALDOIback to textback to textback to text
• 12 articleN.Nicolas Buton, F.François Coste and Y.Yann Le Cunff. Predicting enzymatic function of protein sequences with attention.Bioinformatics3910October 2023, 1-10HALDOIback to textback to textback to text
• 13 articleA.Anca Caranfil, Y. L.Yann Le Cunff and C.Charles Kervrann. BayesTICS: Local temporal image correlation spectroscopy and Bayesian simulation technique for sparse estimation of diffusion in fluorescence imaging.Biological Imaging32023, 1-12HALDOIback to text
• 14 articleO.Olivier Dennler, F.François Coste, S.Samuel Blanquart, C.Catherine Belleannée and N.Nathalie Théret. Phylogenetic inference of the emergence of sequence modules and protein-protein interactions in the ADAMTS-TSL family.PLoS Computational Biology198August 2023, e1011404HALDOIback to textback to textback to text
• 15 articleI.Ibrahim Fakih, J.Jeanne Got, C. E.Carlos Eduardo Robles-Rodriguez, A.Anne Siegel, E.Evelyne Forano and R.Rafael Munoz Tamayo. Dynamic genome-based metabolic modeling of the predominant cellulolytic rumen bacterium Fibrobacter succinogenes S85.mSystems83June 2023, 1-23HALDOIback to textback to text
• 16 articleC.Camille Juigné, E.Emmanuelle Becker and F.Florence Gondret. Small networks of expressed genes in the whole blood and relationships to profiles in circulating metabolites provide insights in inter-individual variability of feed efficiency in growing pigs.BMC Genomics232023, 1-20HALDOIback to textback to textback to text
• 17 articleC.Camille Juigné, O.Olivier Dameron, F.François Moreews, F.Florence Gondret and E.Emmanuelle Becker. Fixing molecular complexes in BioPAX standards to enrich interactions and detect redundancies using Semantic Web Technologies.Bioinformatics3952023, btad257HALDOIback to textback to textback to text

International peer-reviewed conferences

• 18 inproceedingsC.Camille Juigné, E.Emmanuelle Becker and F.Florence Gondret. Combined transcriptomics and metabolomics in the whole blood to depict feed efficiency in pigs.Book of abstracts74. Annual meeting of the european federation of animal science (EAAP)29Book of abstracts of the 74st annual meeting of the european federation of animal scienceLyon, FranceWageningen Academic Publishers2023, 516HALback to textback to text

Doctoral dissertations and habilitation theses

• 19 thesisM.Matthieu Bouguéon. Kappa modeling of hepatic stellate cell dynamics during fibrosis development and reversion.Université de Rennes (2023-....)December 2023HALback to textback to text
• 20 thesisN.Nicolas Buton. Modèle Transformer pour l’interprétabilité et les prédictions multi-niveaux des fonctions des protéines à partir de leurs séquences.Université de RennesOctober 2023HALback to textback to text
• 21 thesisC.Camille Juigné. Integration and analysis of heterogeneous biological data through multilayer graph exploitation to gain deeper insights into feed efficiency variations in growing pigs.Institut Agro - Rennes AngersDecember 2023HALback to textback to textback to text

Reports & preprints

• 22 miscC.Camille Juigné, E.Emmanuelle Becker and F.Florence Gondret. Small networks of expressed genes in the whole blood and relationships to profiles in circulating metabolites provide insights in inter-individual variability of feed efficiency in growing pigs.March 2023HAL

Other scientific publications

• 23 thesisP.Pablo Espana Gutierrez. Learning canonical Potts models.Université de rennesJune 2023HALback to textback to text
• 24 thesisA.Alix Regnier. Identification systématique de modules dans une collection de requêtes SPARQL.Université de RennesJune 2023HALback to textback to text