BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//University of Iceland//AI Centre//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
UID:c888f6ec-bf63-47c7-8860-5f7d34756e9c@ai.hi.is
DTSTAMP:20261011T043458Z
DTSTART:20260203T090000Z
DTEND:20260203T100000Z
SUMMARY:Improved AI-assisted detection of deletions and duplications from S
 NP array data
DESCRIPTION:Copy number variants (CNVs) are an important source of genetic 
 variation in the human genome implicated in evolution and disease suscepti
 bility. The presence of larger types of copy number variants can be inferr
 ed from SNP array data by examining probe intensity (log-R-ratio\, LRR) an
 d allelic ratios (B-allele frequency\, BAF). Existing CNV-calling methods 
 such as PennCNV are proficient in detecting true CNVs but suffer a high fa
 lse positive (FP) call rate as well as inaccurate estimation of CNV bounda
 ries\, which limits their use for genome-wide analyses in large datasets\,
  as CNV calls need to be validated through sequencing and/or visual inspec
 tion of LRR and BAF patterns.\n\nWe visually inspected 60\,000 CNV calls f
 rom 22\,500 samples genotyped on different SNP arrays and found the majori
 ty to be false positive or unclear. Using a subset of this dataset\, we tr
 ained a convolutional neural network to automate the validation of CNVs th
 rough machine vision. Out-of-sample accuracy of the model exceeded 90%\, a
 pproximating that of a human analyst. Orthogonal validation with genome se
 quencing data found our visual validation to be highly accurate\, with onl
 y 1.7% of calls supported by the sequencing dataset deemed as false by the
  human analyst.\n\nhttps://ai.hi.is/is/events/colloquium-in-statistics-and
 -ai-9/
LOCATION:Íslensk erfðagreining\, Tjarnarsalur
URL:https://ai.hi.is/is/events/colloquium-in-statistics-and-ai-9/
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR