BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:A Bayesian model-based approach to finding cell-type level associa
 tions in heterogeneous methylation samples - Daniel Kennedy\, Queensland U
 niversity of Technology
DTSTART:20170914T133000Z
DTEND:20170914T143000Z
UID:TALK76732@talks.cam.ac.uk
CONTACT:Alison Quenault
DESCRIPTION:Epigenome-wide association studies are often performed using h
 eterogeneous methylation samples\, especially when there is no prior infor
 mation as to which cell-types are disease associated. While much work has 
 been done on estimating cell-type fractions and removing cell-type heterog
 eneity variation\, relatively little work has been done on identifying cel
 l-type\n\nspecific variation in heterogeneous samples. In this talk I pres
 ent a Bayesian model-based approach for making cell-type specific inferenc
 es in heterogeneous settings\, by utilising a logistic transform to proper
 ly constrain parameters\, and incorporating a prior knowledge of cell-type
  lineage via prior covariance structure. The approach was applied to the d
 etermination of sex-specific cell-type effects in methylation\, where cell
 -type information was present as an independent verification of the result
 s. The approach showed significant improvement in performance over previou
 sly used methods\, particularly for detecting association in several rare 
 cell-types. I outline current and future work on this problem\, which leve
 rages the flexibility of the Bayesian modelling approach by incorporating 
 local methylation correlation and multiple data-types.
LOCATION:Small Seminar Room\, 1st Floor\, Institute of Public Health\, Uni
 versity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR