Skip to Main content Skip to Navigation

Sciences de l'information pour l'étude des systèmes biologiques (exemple du vieillissement du système immunitaire)

Abstract : High-throughput experimental approaches for gene expression study involve several processing steps for the quantification, the annotation and interpretation of the results. The i3 lab and the LGIPH, applies these approaches in various experimental setups. However, limitations have been observed when using conventional approaches for annotating gene expression signatures. The main objective of this thesis was to develop an alternative annotation approach to overcome this problem. The approach we have developed is based on the contextualization of genes and their products, and then biological pathways modeling to produce a knowledge base for the study of gene expression. We define a gene expression context as follows: cell population+ anatomical compartment+ pathological condition. For the production of gene contexts, we have opted for the massive screening of literature. We have developed a Python package, which allows annotating the texts according to three ontologies chosen according to our definition of the context. We show here that it ensures better performance for text annotation the reference tool. We used our package to screen an aging immune system text corpus. The results are presented here. To model the biological pathways we have developed, in collaboration with the LIPAH lab a modeling method based on a genetic algorithm that allows combining the results semantics proximity using the Biological Process ontology and the interactions data from db-string. We were able to find networks with an error rate of 0.47.
Document type :
Complete list of metadatas
Contributor : Abes Star :  Contact
Submitted on : Tuesday, November 14, 2017 - 10:16:07 PM
Last modification on : Sunday, October 25, 2020 - 1:57:33 PM
Long-term archiving on: : Thursday, February 15, 2018 - 3:22:26 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01635268, version 1


Walid Bedhiafi. Sciences de l'information pour l'étude des systèmes biologiques (exemple du vieillissement du système immunitaire). Bio-Informatique, Biologie Systémique [q-bio.QM]. Université Pierre et Marie Curie - Paris VI; Université de Tunis El Manar, 2017. Français. ⟨NNT : 2017PA066139⟩. ⟨tel-01635268⟩



Record views


Files downloads