BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Using topic models to help cure cancer - Prof Quaid Morris (Toront
 o)
DTSTART:20100611T130000Z
DTEND:20100611T140000Z
UID:TALK25052@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:Better diagnosis of cancer using topic models may hold the\nke
 y to curing cancer.  This is because cancer is not a single disease\nbut a
  large number of related diseases. Different types of cancer\nrespond bett
 er to different types of therapy.  Targeted therapies\nwould improve patie
 nt survival\, but before we can do that\, we need to\nbe able to better su
 b-categorize cancer tumours.  Right now\,\ninternational consortia are spe
 nding hundreds of millions of dollars\nto sequence the genomes of 25\,000 
 cancers in an effort to identify\ncancer-associated mutations and try to u
 nderstand their role.\nHowever\, many of these mutations do not impact the
  coding sequence of\ngenes but instead they change the instructions that d
 escribe when the\ngenes should be activated and how the gene’s messenger
  RNA (mRNA)\nshould be processed. So\, in order to understand how these mu
 tations\nare associated with cancer\, we need to be able to understand the
 \ngenome regulatory code and we are far from being able to do that.  On\nt
 he other hand\, we can (at least in theory) directly measure the mRNA\nlev
 els and determine how they change in cancer.  But cancer RNA levels\nare d
 ifficult to measure accurate because tumours are “contaminated”\nby wi
 dely varying amounts of normal\, healthy tissue.  Because RNA\nlevels chan
 ge rapidly once a tumour biopsy is taken\, there’s simply\nnot enough ti
 me to physically purify the sample.   I will describe how\nto perform this
  purification statistically using topic models.  Our\nmethods deconvolve m
 easured tumour RNA levels into those coming from\nhealthy and cancerous ce
 ll populations.  Our modeling framework is\nnaturally extended to identify
 ing the site of origin of metastatic\ncancers and to finding descriptive g
 roupings of cancers into\nsub-types.
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR