BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:A Bayesian functional principal component analysis framework for l
 ongitudinal genome-wide association studies
DTSTART:20240208T141500
DTEND:20240208T160000
DTSTAMP:20260407T064540Z
UID:66e0fbdf7cf3867977688e9efd6f5c6876a2b1777a6f3037e62b1cec
CATEGORIES:Conferences - Seminars
DESCRIPTION:Daniel Temko\, University of Cambridge        \nIntrod
 uction: When interrogated along with genetic data\, large clinical longitu
 dinal datasets have potential to yield new scientific insights on the gene
 tic contribution to temporal patterns of disease progression and response 
 to treatments. However\, existing approaches for longitudinal genome-wide 
 association studies (GWAS) are not suited to unlock the full power of larg
 e datasets\, as they typically rely on strong distributional assumptions o
 n the longitudinal outcome.\n \nMethods: We propose to reframe the longit
 udinal GWAS problem as a joint latent variable estimation and regression p
 roblem in which genetic variants influence longitudinal trajectories via e
 ffects on functional latent variables. We leverage existing Bayesian infer
 ence frameworks for functional principal component analysis (FPCA) and spa
 rse spike-and-slab regression to develop a two-stage variational inference
  scheme for this model that conveys uncertainty from the principal compone
 nt estimation into the second-stage regression.\n \nResults: Using simula
 tions\, we show that our approach is both scalable and accurate\, and that
  our modelling approach can recover SNPs that influence latent dynamics un
 derlying longitudinal trajectories. We further demonstrate the usefulness 
 and applicability of our framework in a study of genetic effects on longit
 udinal outcomes in the UK biobank.\n \nConclusion: We present a modelling
  scheme that has the computational scalability and flexibility to take adv
 antage of large datasets for longitudinal modelling\, while rigorously han
 dling estimation uncertainty.\n\nDaniel Temko\, Tui Nolan\, Sylvia Richard
 son\, Hélène Ruffieux\n 
LOCATION:MA A1 10 https://plan.epfl.ch/?room==MA%20A1%2010
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR