Société Française de Bioinformatique

Mots-Clés

molecular modeling ionic liquids machine learning biopolymers

Description

We are accepting applications for a three-year PhD position on a project that will be performed in the Laboratoire de Glycochimie et des Agroressources d’Amiens (LG2A) in Amiens (France), starting on October 1st, 2025, on the following subject: “Dissolution, generation and cross-linking of biopolymers: Towards machine learning predictions of structure-property relationships via a combined theoretical/experimental approach”

*Context and Research Project
Natural polymers from biomass whether semi-crystalline (cellulose, chitin, silk, … ) or non-crystalline (marine polysaccharides, bacterial, …) represent a virtually inexhaustible green and renewable resource that can be used to design materials with improved mechanical and/or biological properties. The applications of these functionalized and/or cross-linked polymers are very versatile and range from the formation of gels to mimic biological media, the replacement of synthetic polymers (plastics), to the design of specific platforms for the removal of eternal pollutants (drugs, pesticides, etc.).
Natural semi-crystalline polymers are highly crystalline per nature and possess a large and robust hydrogen bond network, making them recalcitrant to solubilization in most of the conventional solvents. Hence, large-scale processing and treatment of these polymers are not environmentally-friendly because of their use of non-green aggressive solvents. To make the dissolution and regeneration processes greener, ionic liquids and deep eutectic solvents might be used. Although the number of such solvents is very large (even infinite, theoretically), few have proved so far to be able to efficiently dissolute and regenerate any type of polymer, which represents a crucial step in obtaining new and original materials.
This PhD project combines both experimental and computational methods to study the dissolution and (re)-generation of biopolymers in a wide range of green solvents, neat and in mixtures, in order to gain some understanding at the molecular level of the mechanisms by which semi-crystalline biopolymer slabs are degraded and then subsequently cross-linked.
This work is divided into several steps:
(i) extraction/dissolution of the semi-crystalline polymers from the biomass using green solvents,
(ii) regeneration of pristine materials in a biofilm/matrix/gel form.
(iii) generation of hybrid cross-linked materials.
Another goal of this PhD project is to produce for each of these work steps, by means of molecular dynamics simulations, datasets which will be fed into machine learning methods to eventually predict compositions of ionic and eutectic solvents able to efficiently dissolute/regenerate a given biopolymer and the composition and size of a hybrid biopolymer featuring ad hoc properties.

*Requirements
Successful candidates should hold a Master’s degree (or equivalent) in chemistry or physics with an excellent academic record. Knowledge or experience in computational chemistry, molecular modeling and computer programming is highly desirable. Familiarity with machine learning would be appreciated. Good knowledge of English (written and oral) is mandatory.

*Amiens and its surroundings
Amiens is the main town of Picardie (now part of region Hauts-de-France), situated just north of Paris (Paris can be easily reached by train in about 1 hour). The University (UPJV) brings into town about 30,000 students.