CMStatistics 2022: Start Registration
View Submission - CMStatistics
B0655
Title: Functional trait locus mapping by functional data clustering Authors:  Marie-Helene Descary - University of Quebec in Montreal (Canada) [presenting]
Abstract: The main goal of trait locus mapping (gene mapping) is to identify the locus (gene), i.e. a region of the DNA, that affects a trait of interest. Traditionally, traits of interest were binary (e.g. case vs control) or quantitative (e.g. blood pressure), but it is more and more common to work with functional traits (e.g. growth curve). We will present a new methodology of gene mapping for a functional trait that uses tools developed to analyze functional data, i.e. data that can be seen as realizations of a random function. The idea behind the new method is to translate the problem of identifying a gene associated with a functional trait into the problem of finding the ``best'' clustering of a set of functions. We consider different measures of functional dissimilarity leading to a flexible gene mapping method, in the sense that it can detect a wide range of effects of a gene on a functional trait. The performance of the new method is assessed with a simulation study and an application to real data.