PhD Position F/M [Numetab project] High-dimensional data analysis in the framework of survival modelling in presence of competing risks.

Context
The PhD candidate will join both the Datavers team based in the Inria center at the University of Lille (France) and the METRICS unit research based in the Faculty of Medicine at the University of Lille.
Inria Datavers team is specialised in statistical learning and her members also belong either to Paul Painlevé laboratory or METRICS.
Research of Paul Painlevé laboratory covers a broad spectrum of mathematical research.
METRICS is a multidisciplinary public health research team that brings together clinicians and methodologists in the fields of ergonomics, biostatistics, medical informatics, data science and health economics.
Assignment
The aim of the PhD thesis is to develop a new statistical method for variable selection in the context of survival analysis, with competing risks, where measurements are repeated over time on thousands of variables simultaneously. This PhD project is funded by the Numetab cross disciplinary project (Numerical transition in Obesity and Metabolic Research, PI: Pr. Pattou and Pr. Marot).
The application will be to identify new omics markers (transcriptomics, metabolomics, proteomics) for predictive purposes in order to adapt patients’ treatment (Precision Health). The originality of the statistical approach consists in simultaneously addressing three scientific challenges: i) high-dimensional analysis (more variables than individuals), ii) analysis of competing risks in the context of survival models and iii) analysis of longitudinal data (repeated measurements on the same individuals).
Results are expected both in terms of biostatistical methodology and software to provide a clinical decision support tool.
Main activities
The main objective is to develop a statistical method for selecting variables in a context of survival prediction in the presence of competing risks.
Secondary objectives are:
to improve the search for biomarkers in analyses with repeated omics data
to develop a decision-support software tool that can be generic and can be used in any context of high-dimensional data analysis with survival models in the presence of competing risks.
PhD candidates who want to understand the concepts necessary for this PhD can read the following references:

• related to statistics

[1] Fine and Gray (1999) A proportional hazards model for the subdistribution of a competing risk. Journal of the American statistical association, 94(446):496–509.
[2] Chi and Ibrahim (2006) Joint models for multivariate longitudinal and multivariate survival data. Biometrics, 62(2):432–445.
[3] Binder, Allignol, Schumacher, and Beyersmann (2009) Boosting for high-dimensional time-to-event data with competing risks. Bioinformatics, 25(7):890–896.

• related to another application which motivated the PhD subject (proteomics data analysis)

[4] Heyse, Vandewalle, Marot, Amouyel, Bauters, and Pinet (2023) Identification of patient subtypes based on protein expression for prediction of heart failure after myocardial infarction. Iscience, 26(3).
Skills
The candidate must hold a MSc in Biostatistics.
Technical skills and level required :
Proficiency in R and at least one scripting language (e.g. python)
In-depth knowledge of statistical learning to handle high-dimensional data
Basic knowledge of biology (molecular and cellular biology, genomics and genetics)
Other valued appreciated : creativity, independence, team working and communication skills