第26回インシリコ・メガバンク研究会を下記のとおり行いますのでご案内いたします。今回は東京大学医科学研究所 George Chalkidisさんを講師としてお迎えし、「遺伝子発現制御機序としての長鎖非コードRNAの情報理論分析」について講演していただきます。

・日時：平成25年4月19日(金)　17：00‐18：30
・場所：東北メディカル・メガバンク機構2階会議室1 http://www.megabank.tohoku.ac.jp/info/access.html

・演題：Information Theoretic Analysis of Long non-coding RNA as a gene expression control mechanism
・講師：George Chalkidis ( Institute of Medical Science, The University of Tokyo)

・概要：An emerging theme from multiple model systems is that lncRNAs form
extensive networks of regulatory mechanisms that control the higher-order
organization of the transcriptome. The importance of these modes of
regulation is underscored by the newly recognized roles of long non-coding
RNAs for proper gene control across all kingdoms of life.
The means by which such lncRNAs regulate transcription are speculated to
encompass a diversity of mechanisms and the identification and
characterization of these mechanisms is an active area of research.
There will be 30-50 RNA-seq and TSS-seq data samples from human cultured
cells from various cell lines and various conditions. The objective will be
to discover potential control mechanisms and model them from small, noisy
samples. The Minimum Description Length Principle (MDL) together with the
maximal information coefficient (MIC) are particularly suited for that
task, because MDL shows excellent non- asymptotical performance in terms of
learning the statistics of finite size context sources from finite length
training data and thus enables the de-noising and correct model
identification even from small sample sizes. Furthermore, MIC is a powerful

novel dependency measure for two-variable relationships of the class of
maximal information-based nonparametric exploration statistics that
captures a wide range of associations both functional and non-functional
and provides a score close to 1 for functional and non-functional
relationships and score close to 0 for statistically independent random
variables.

・世話人：長﨑正朗