BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260808T052359EDT-60992NRlJt@132.216.98.100
DTSTAMP:20260808T092359Z
DESCRIPTION:Cancer has complex patterns of progression that includes\nconve
 rging as well as diverging progressional pathways. The path\nmodel of colo
 n cancer introduced by Vogelstein was a pioneering\ncontribution to cancer
  research. Since then\, several attempts have\nbeen made to obtain mathema
 tical models of cancer progression\,\ni.e.\, defining classes of models\, 
 devising training algorithms\, and\napplying these to cross-sectional data
 . We study tree-based models\nand mixtures of such\, since they have the d
 esirable features of\nhaving relatively few parameters as well as being re
 latively easy\nto train. In this talk\, we will discuss previous work\, in
 cluding\nthat of Beerenwinkel et al. who provided EM-like algorithms for t
 he\nso-called Oncogenetic Trees and mixtures of such. We will then\nintrod
 uce Hidden Variable Oncogenetic Trees (HOTs) which allow for\nerrors in th
 e data and thereby more realistic modeling\, and we will\npresent proper E
 M-algorithms for the HOTs as well as HOT-mixtures.\nThe EM-algorithm for a
  single HOT is shown to perform very well for\nreasonable-sized data sets\
 , while HOT-mixtures require data sets of\nsizes that can only be obtained
  with tomorrow's more cost efficient\ntechnologies. To facilitate analysis
  of complex cytogenetic data\nsets requiring more than one HOT\, we train 
 parameters on a colon\ncancer (CC) data set and devise a decomposition str
 ategy based on\nPrincipal Component Analysis. The method so obtained is th
 en\nsuccessfully applied to kidney cancer (RCC).\n
DTSTART:20081029T183000Z
DTEND:20081029T200000Z
LOCATION:McIntyre Medical Building\, CA\, QC\, Montreal\, H3G 1Y6\, 3655 pr
 omenade Sir William Osler
SUMMARY:A Structural EM-Algorithm For a Mixture Model Capturing Cancer Prog
 ression
URL:/channels/event/structural-em-algorithm-mixture-mo
 del-capturing-cancer-progression-102244
END:VEVENT
END:VCALENDAR