(archives from old symbiose site)
Scope spectral clustering of 5610 prokaryotic metabolic networks to discover environmental phenotypesAdèle Weber (UPMC)Thursday, February 20, 2020 – 10:30 to 12:00Room AurignyTalk abstract:
The scope of metabolic networks, a method of analysis of metabolic graphs, is defined as all metabolites and reactions in an expanded network from a set of seed metabolites. We explore how the scope can be efficiently incorporated into state-of-the-art machine learning methods. In particular, we investigate how the scope can be used to perform dimensionality reduction, clustering, and real data exploration. We run our experiments on most prokaryotic species from KEGG database (5610). We perform the metabolic graph reconstruction using our pipeline, and we compute the metabolic scopes for all species. The spectral clustering applied to the reduced data set, represented by the metabolic scopes, leads to a partitioning of the species which can be linked to various biological, phylogenetic and environmental phenotypes. Furthermore, our method enables us to identify the underlying biological phenomenons characterising the different clusters.
Protein Multiple Alignments: Sequence-based vs Structure-based ProgramsMathilde Carpentier (MNHN)Thursday, January 30, 2020 – 10:30 to 12:00Room AurignyTalk abstract:
Motivation: Multiple sequence alignment programs have proved to be very useful and have already been evaluated in the literature yet, not alignment programs based on structure or both sequence and structure. In the present article we wish to evaluate the added value provided through considering structures. Results: We compared the multiple alignments resulting from 25 programs either based on sequence, structure, or both, to reference alignments deposited in five databases (BALIBASE 2 and 3, HOMSTRAD, OXBENCH and SISY- PHUS). On the whole, the structure-based methods compute more reliable alignments than the sequence-based ones, and even than the sequence+structure-based programs whatever the databases. Two programs lead, MAMMOTH and MATRAS, nevertheless the performances of MUSTANG, MATT, 3DCOMB, TCOFFEE+TM ALIGN and TCOFFEE+SAP are better for some alignments. The advantage of structure-based methods increases at low levels of sequence identity, or for residues in regular secondary structures or buried ones. Concerning gap management, sequence-based programs set less gaps than structure-based programs. Concerning the databases, the alignments of the manually built databases are more challenging for the programs.
Assemblage de génome et topologie algébriqueJean-François Gibrat (INRA)Thursday, January 16, 2020 – 10:30Room AurignyTalk abstract:
Le séminaire présente un travail, encore préliminaire, sur l’assemblage de génome à partir de données de séquenceurs de 3e génération (PacBio et Oxford Nanopore Technology). Dans un premier temps, je présenterai un algorithme efficace pour assembler ces lectures, basé sur l’analyse d’un graphe OLC (Overlap-Layout-Consensus). Dans un second temps, je montrerai comment on peut utiliser certains concepts de topologie algébrique (les nombres de Betti) pour analyser les caractéristiques du graphe OLC et déterminer à l’avance si l’assemblage sera simple ou s’il faudra rechercher les “noeuds” et identifier les “boucles” dans le graphe occasionnées par les régions répétées non résolues du génome.
Qualitative modeling of the immune response: from mechanisms to populationsAurelien Naldi (ENS)Thursday, January 9, 2020 – 10:30 to 12:00Room AurignyTalk abstract:
Qualitative dynamical models, based on generalized Boolean functions, have been used over the last decades to study complex biological systems. As a reasoning tool for researchers in biology, they facilitate the identification of missing or inconsistent knowledge, and can ultimately guide experimental design. We are interested in particular in the regulation of the immune response. In this context, we assembled and analyzed a comprehensive model of the differentiation and plasticity of regulatory T cells, which are heavily define the nature and amplitude of this response. However, these cells are not isolated actors, they interact with each other and with several other types of cells through chemical signals. We are currently extending existing modelling tools to account for large-scale populations of mechanical models. In the long term, we aim to reduce the gap between mechanistic models at the single cell scale and phenomenological ones at the population scale.