UPR 5301

Francois Bonnardel’s thesis defense on February, 8th 2021

Such thesis is entitled "Bioinformatic study of lectins: new classification and prediction in genomes" and is co-directed by Dr. Anne Imberty (Cermav) and Dr/Prof. Frédérique Lisacek (Swiss Institute of Bioinformatics). Click on the title for more information.


« Bioinformatics uses mathematical concepts and informatics tools to unravel the knowledge hidden in biological data. When bioinformatics is applied to glycans and glycobiology, it is called glyco-informatics. New technologies allow mass sequencing of new species genomes and of environmental samples metagenomes. But all newly discovered genomes and encoded proteins are only partially annotated with biological function assessed by similarities to reference organisms.
Glycobiology is the research field dedicated to the study of glycan/carbohydrate compounds, composed of one or multiple monosaccharides. Lectins are proteins able to bind reversibly to glycans, and without enzymatic functions. Lectins are powerful tools for the recognition of glycans in samples, and they are also targets for therapeutic compounds due to their involvement in cancer, immunology and infections.
This thesis aims to use bioinformatics for developing new in silico tools for the study of lectins. More specifically, it addresses the need for a new online database covering curated information on lectins for both reference organisms and newly sequenced genomes belonging to other organisms.
To provide a curated classification of lectin 3D structures and their annotation in genomes, a dedicated web portal called UniLectin, was developed and includes several modules. The UniLectin3D module provides manually curated and classified 3D structures together with their interacting glycans. Due to the difficulty of identifying tandem repeated lectins in genomes, a specific method has been developed for the prediction of those particular lectins, now available in the PropLec and TrefLec modules. Finally, the LectomeXplore module includes lectin predictions based on 107 classes defined on the basis of UniLectin3D content, and resulting from screening available sequences stored in the reference protein databases NCBI-nr and UniProt. This made the study of lectomes in different environments possible as collaborative work described in the last part of the thesis. »