École Nationale Supérieure de
Techniques Avancées

Résumé

Understanding the composition of a cancer tumor is the first step towards its elimination. Here, we will describe the development of a method designed to determine the number and proportions of subpopulations in a cell population. This method has a statistical approach: it consists in finding the best model to fit the data, by maximizing the log-likelihood and optimizing the model selection criterion – here, the most consistent results were found with the Bayesian Information Criterion (BIC). This method has been developed in the form of an algorithm on Matlab, and tested with both test data and real patient data.