Improving the Geometric Characterization Methods of Protein Helices by Total Least Squares Analysis

碩士 === 國立成功大學 === 製造工程研究所碩博士班 === 92 ===   Six methods, namely parametric least-squares method, moment matrix, cylindrical fit, cross-product of triad bisector, rotational transformation method, and helix analysis (HELANAL) , have been explored for geometric characterization of a protein helix. Whi...

Full description

Bibliographic Details
Main Authors: Chao-Ching Ho, 何昭慶
Other Authors: Ching-Cheng Wang
Format: Others
Language:zh-TW
Published: 2004
Online Access:http://ndltd.ncl.edu.tw/handle/17108669278054866202
Description
Summary:碩士 === 國立成功大學 === 製造工程研究所碩博士班 === 92 ===   Six methods, namely parametric least-squares method, moment matrix, cylindrical fit, cross-product of triad bisector, rotational transformation method, and helix analysis (HELANAL) , have been explored for geometric characterization of a protein helix. While HELANAL distinguishes among linear, curved, and kinked helix, all the other methods assume only linear ones. Unfortunately, HELANAL employs linear regression analysis making the results dependent of the coordinate system. For the remedy, we explore total least squares analysis (TLS) making results independent from the coordinate system. The thesis presents the TLS method with demonstration examples. TLS results are compared with those obtained by HELANAL.