CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Joint modeling of floral transition time and leaf appearance process in maize plant Authors:  Sandra Plancade - INRA / Catholic University of Louvain-la-Neuve (France) [presenting]
Anouar El Ghouch - The University catholique de Louvain (Belgium)
Sylvie Huet - INRA (France)
Christine Dillmann - INRA - University of Paris-Sud (France)
Abstract: Plant growth is usually modeled through phases which affect simultaneously various phenotypes, including external traits that can be repeatedly observed, as well as internal traits that require destructive observations, and a question of interest is to get information on the latter based on observations of the former. The focus will be on the estimation of two traits of maize plant development: the floral transition time, a key-step occurring inside the stem, and the process of leaf appearance or phyllochrone, based on repeated measurements of the number of leaves for a set of plants, as well as dissections indicating whether floral transition already occurred; thus, floral transition and leaf appearance times are subject to current status and interval censoring mechanisms respectively. In literature, censoring is usually circumvented by aggregating measurements from a plot of plants selected as homogeneous, in order to create a ``pseudo-plant''. We propose an alternative approach based on an accelerated failure time model estimated via an expectation maximization algorithm, which allows us to account for plant-level variability.