Analysis of Complex Traits in Families and Populations
Chair: Jinko Graham
Organizer: Jinko Graham
Sponsor: Biostatistics Section 
 

[Monday, June 12, 2017 10:20-11:50]

10:20-10:50 
J. Concepcion Loredo-Osti (Memorial University of Newfoundland) 
The Analysis of Longitudinal Multivariate (Discrete or Continuous) Traits under Irregular Time Measurements 
 

There are well known methods to analyse longitudinal continuous variables with regularly spaced measurements and some of them (in particular, mixed models) have been adapted to deal with longitudinal quantitative traits. Nonetheless, the problem is far from being solved. The modelling and analysis of longitudinal multivariate discrete and or continuous traits poses various challenges to accommodate the family structure and irregular time measurements. In this presentation we discuss these challenges and propose some ways of addressing the problems.

10:50-11:20
Kelly M. Burkett (University of Ottawa)
An Ancestral Tree-Based Approach to Detect Rare and Common Variants 

For detecting genetic variants associated with a disease or trait, it is useful to consider the ancestral trees that gave rise to the sample's genetic variability. For both rare and common disease or trait influencing genetic variants, we expect to see haplotypes from individuals with similar values of the disease or trait clustered together in the ancestral tree corresponding to the genomic location of the variant. In this presentation, we describe how tree-based statistics can be used for detecting both rare and common genetic variants associated with either continuous or dichotomous outcomes. We summarize the performance of these statistics on simulated data having known and missing tree structures and we compare results to those obtained using conventional approaches to detect genetic association. Finally, application of the tree-based method to real data is also discussed.

11:20-11:50
Fabrice Larribe (Université du Québec à Montréal)
Mapping Complex Traits, Rare Variants and Interaction via the Coalescent Process with Recombination 

Genetic population data is the result of the evolution of chromosomes on this population; thus, it has been found natural to try to analyze such data by modeling the unknown history of this population. By combining an approach based on the coalescent process with recombination and importance sampling techniques, we can show how this stochastic process can be used to map genes influencing a disease. We address issues pertaining to complex diseases, whether this complexity is due to incomplete penetrance or phenocopy, to multiple rare variants, or to the interaction between genes. We introduce the progress made in building such histories and outline the remaining challenges.