|
Descriptif
|
Context: All the cells of a multicellular organism contain the same genetic information but differ by their shapes, their physiologies and their functions. These differences result from specific patterns of gene expression, which largely rely on biochemical tags, the so-called epigenetic marks, that are deposited on top of the genetic information. Failure of preserving the proper epigenetic mark profiles might result in inappropriate gene activity and diseases like cancer. Biological processes involved in the maintenance of such epigenetic information are intrinsically stochastic. Understanding the paradigm of maintaining a robust information from stochastic processes is therefore crucial to better characterize epigenetic (de)regulation in normal and pathological tissues.
Objectives and expected results: We propose an internship to develop statistical – data-driven- models to characterize the stochasticity of epigenetic regulation in normal and cancerous tissues. The project is part of a collaboration between the group of Daniel Jost in the TIMC-IMAG laboratory for the modeling part, the group of Saadi Khochbin at the Institute of Advanced Biosciences (Grenoble) for the experimental part, and the team of Elisabeth Brambilla at the University Hospital of Grenoble, specialized in the pathology of lung tumors. The student will develop computational methods, based on maximum entropy principles, to analyze single-cell epigenomic data. Such approaches, closed to the well known Ising model in Physics, will allow to describe quantitatively the experimental data. A challenging part of the internship will be to develop efficient inference scheme to estimate the model parameters (the inverse Ising problem). After validation of the method on in silico data, the student will apply it to data issued from publicly available studies on healthy and cancerous lung tissues. Results of the models will be compared to predictions from mechanistic models of epigenetic regulation allowing to get new insights into the underlying regulatory mechanisms. We are looking for a highly motivated master student, who likes team work and interdisciplinary environments. A strong background in physics and/or statistics and/or bioinformatics, coding and familiarity with Linux environment are required.
References: Decelle et al, Phys Rev E 94: 012112 (2016). Lezon et al, PNAS 103: 19033 (2006). Guetierrez-Arcelu et al, ELife 2:e00523 (2013). Jenkinson et al, Nat. Genet. 2017.
|
