<div class="page photo" style="">
<article>
<header style="
background-image:url(/imageLibrary/droplets.jpg);
">
<div class="box">
<div class="intro" style="color: #000;">
<h1 style="color: #000 !important;">Abstract 20150915</h1>
<p class="summary"></p>
</div>
</div>
</header>
<div class="main">
<div class="container">
<p class="byline"> </p>
<p><strong style="background-color: initial;"><의공학연구소 정례세미나></strong></p><p><strong>연자 : 한범 교수 (서울아산병원 융합의학과)</strong></p><p><strong>주제 : 유전체 연관연구를 위한 생물정보학 알고리즘 및 소프트웨어 개발</strong></p><p><strong>일시 : 9월 15일 화요일 17:00~</strong></p><p><strong>장소 : 아산생명과학연구원 교육연구관 4층 회의실 </strong></p><p><strong><br></strong></p><p><strong><abstract></abstract></strong></p><p>Genetic association studies are an
effective means of discovering disease-causing genetic variants, which can lead
to better understanding of disease pathology and development of new drugs.
However, there exist a number of computational challenges in performing
large-scale genome-wide association studies. In this talk, I describe three
challenges and my efforts to address those. First, performing association
studies for HLA genes in the MHC region is extremely difficult, because both
next-generation sequencing and genotyping microarray do not work due to the
highly polymorphic nature of HLA genes. I developed SNP2HLA, a software package
that analytically predicts HLA types based on the SNP data. The method is widely
used for identifying important HLA amino acid positions that affect
susceptibilities to autoimmune diseases. Second, two phenotypes often share
genetic structures, but it is often unclear if the sharing is due to common
genetic basis (pleiotropy) or misdiagnosis of samples (clinical heterogeneity).
I developed BUHMBOX, a method that helps interpret shared genetic structure
between two different diseases by distinguishing pleiotropy from heterogeneity.
Application of the method to autoimmune and psychiatric disorders shows that
their genetic sharing is likely due to pleiotropy. Third, multiple testing
correction is a computationally expensive challenge in eQTL studies. I
developed eGene-MVN, an approach that efficiently corrects for multiple testing
in eQTL studies. The method was chosen as the primary tool for the analysis
pipeline of The Genotype-Tissue Expression (GTEx) Consortium. </p>
</div>
</div>
</article>
</div><!-- /page-->