@techreport{oai:ipsj.ixsq.nii.ac.jp:00194997,
 author = {黄, 毅聰 and 吉川, 寧 and 和久井, 直樹 and 大上, 雅史 and 秋山, 泰 and Yicong, Huang and Yasushi, Yoshikawa and Naoki, Wakui and Masahito, Ohue and Yutaka, Akiyama},
 issue = {10},
 month = {Mar},
 note = {近年では環状ペプチド医薬品が注目されているが，環状ペプチドは細胞膜透過性が低い傾向があり，高い膜透過性を有するものを計算機で効率的に選別できる予測システムが求められている．本研究では，拡張サンプリングMD法による膜透過性予測手法を開発し、さらにデータ並列とMPI並列で高速化した．102個の環状ペプチドの膜透過シミュレーションを行い，膜透過を再現できた90個の環状ペプチドに対する最良の予測結果では，膜透過性予測値の実験値との有意な相関が示された（相関係数0.55）．また，予測の高速化のために脂質二重膜の対称性を利用して膜透過率の計算に必要なデータを水層から膜に入る段階のみから見積もる手法を提案した．それにより，14個のペプチドに対する計算では約1.4倍の高速化に成功し，予測値と実験値との有意な相関が示された（相関係数0.71）．, Peptide drugs have attracted attention in the drug discovery field over recent years. However, their poor membrane permeability is one of the problems for intracellular drug targets. Thus predicting the membrane permeability is important for the peptide drug discovery projects. This study aims to predict the membrane permeability of cyclic peptides using molecular dynamics (MD) simulation and parallel computing. We performed MD simulations for 102 cyclic hexapeptides using supervised MD enhanced-sampling method. In the best prediction results for 90 cyclic peptides for which we successfully reproduced the membrane permeation, the correlation coefficient was 0.55. For the purpose of speeding up the prediction, we proposed a method of estimating the membrane permeability only from the data archived from half of the permeation process by utilizing the symmetry of the membrane. With using this method for 14 test peptides, we achieved 1.4-times increase of calculation speed, and the correlation coefficient was 0.71.},
 title = {拡張サンプリング法による環状ペプチドの膜透過性予測システムの構築},
 year = {2019}
}