The random-model method and a maximum-similarity classification for three-dimensional reconstruction with using simulating phage images

• Main Authors: Jacky, 黃家德
• Other Authors: 吳純純
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/yj56xx

碩士 === 逢甲大學 === 應用數學學系 === 103 === The random-model method and the maximum-similarity classification in the 3D reconstruction with EMAN software were studied. The reliability of the 3D models reconstructed from the 2D images was evaluated by using a simulating model of an icosahedron with an asymmetrical inside as MS2 particles. Primary model were constructed with the random-model method and then refined with the ratio-modified maximum-similarity classification. As examining the reconstructed models with V4 software the maximum-similarity classification not only refined rapidly the primary models, but also reduced the missing angles. The primary models constructed by EMAN with K-means clustering were various even if the same parameters were used in the performance of reconstruction. The random-model method and the maximum-similarity classification used in the 3D reconstruction of asymmetrical particals with EMAN were recommended for getting an acceptable result. With the simulating model the best parameters in the reconstruction performance could be found, but in the study of a real bio-particle as MS2 phage which its 3D microstrcture is vague. The suitable parameters became to be difficultly found. The average model from 3D reconstructed models was suggested to be a reasonable result by the study with the simulating model.