The membrane-bound mucin MUC1 is expressed at the apical pole of epithelial cells and bears a long glycosylated extracellular domain (500 nm). MUC1 interacts with ErbB receptor family and plays a role in signal transduction. MUC1 contributes to tumour invasion by simultaneously disrupting existing interactions between opposing cells (anti-adhesion) and establishing new ligands for interaction between the invading cell and the adjoining cells (adhesion). Cancer cells may also use mucins to protect themselves from adverse cellular growth conditions and control the molecular microenvironment during invasion and metastasis. In Renal clear cell carcinoma (RCC), immunohistochemical studies have shown that MUC1 is frequently overexpressed. Its level of expression is correlated with Fuhrman grade and tumour progression and delocalization of its expression is associated with a worse prognosis. Clinical data suggest that MUC1 might be involved in renal tumour progression. In this project, we will test this hypothesis and evaluate the role of MUC1 on RCC by using in vitro and in vivo models. Recently, we have observed an epithelial-mesenchymal transition (EMT) concomitant to the overexpression of MUC1 in two human renal cell lines. EMT is an important mechanism during development, tissue regeneration and tumour progression (metastasis). To better understand the potential role of MUC1 during EMT, we will develop a murine model of renal regeneration. Our goals are (i) to decipher the role of MUC1 in renal tumour and regeneration, (ii) to identify new therapeutic targets and (iii) to propose new biological tools to the clinicians to ameliorate patient care.

Patients with metabolic diseases are affected by a loss of microbiome species and a reduced number of metabolites detected in their blood, which hints that the missed metabolites may be top mechanistic mediator candidates of the human-gut microbiome interaction. For a better understanding of this interaction, improved metabolomics strategies must be developed, capable of exhaustive characterizations and of delving into the previously undetected metabolites (the co-called “dark metabolome”). Our objective is to illuminate the blood “dark metabolome” with the development of an “all-in-all” framework combining state-of-the-art instrumentation with analytical methods using computational approaches (such as chemoinformatics and chemometrics) for addressing key roadblocks in the field of liquid chromatography coupled to high-resolution mass spectrometry metabolomics and, in doing so, bringing altogether next generation metabolomics. This analytical framework will be developed at 3 levels, corresponding to my 3 work tasks: Task 1, for maximizing the number of mass fragmentation spectra obtained; Task 2, for increasing the number of annotations and the quality of these annotations; and Task 3, for the design of data-driven methods to evaluate the role of the new detected compounds. I will set up this framework with plasma samples from the Euroepan cohort study METACARDIS, descriptive of individuals with cardiometabolic diseases, and validate the biological findings with samples of the Lille cohort study Heart&Brain (comprising diabetic patients without initial diabetes-related complications). To sum up, this project will enhance our understanding of the underexplored microbiome-derived metabolites present in the human blood, as well as their roles in metabolism and physiology, that could ultimately lead to new diagnostic tools and novel treatment and prevention strategies.

The TABASCO project (TABlature ASsisted COmposition) aims at elaborating algorithmic methods to assist musicians, of all skills and influence, in the process of composing and writing guitar scores in modern popular musical styles. These methods aim at facilitating the access to musical composition and encouraging a diversification in the composers creative process. They relate to the fields of digital humanities and digital approaches for artistic creation. The first axis of the project is a musicological study on guitarists composition practices based on surveys of guitar music composers. The results will contribute to identify composition contexts for which algorithmic tools can improve the composer's experience, in particular by facilitating and diversifying the expression of its musical style. The second axis of the project focuses on the elaboration of algorithmic methods for the modeling of gesture annotations indicating hand positions and playing techniques specific to the instrument. Predicting these annotations offers to the composer an original approach to gradually vary the stylistic expressivity of its composition. The third axis focus on the elaboration of algorithmic models facilitating the transfer of the style of a reference musical content to an other content being composed. The design of these methods will be inspired by works in text and image generation. Style transfer assists the composer wishing to get inspired by the style of some reference scores and help him to extend stylistic uniformity of the score he is composing. These methods will be experimented and released through user plugins in a software dedicated to music notation and composition.