Date of Award

Summer 2018

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Science

Committee Director

Jing He

Committee Member

Mohammad Zubair

Committee Member

Desh Ranjan

Committee Member

Jiangwen Sun

Abstract

Cryo-electron microscopy (cryo-EM) is an emerging biophysical technique for structural determination of protein complexes. However, accurate detection of secondary structures is still challenging when cryo-EM density maps are at medium resolutions (5-10 Å). Most existing methods are image processing methods that do not fully utilize available images in the cryo-EM database. In this paper, we present a deep learning approach to segment secondary structure elements as helices and β-sheets from medium- resolution density maps. The proposed 3D convolutional neural network is shown to detect secondary structure locations with an F1 score between 0.79 and 0.88 for six simulated test cases. The architecture was also applied to experimentally-derived cryo- EM density regions of 571 protein chains. . The average F1 score for helix detection is 0.747 and 0.674 for β-sheets in a test involving seven cryo-EM density regions. Additionally, we extend an arc-length association method to β -strands and show that this method for measuring error is superior to many popular methods. An interactive tool is also presented that can visualize the results of this arc-length association method.

DOI

10.25777/kjtq-x893

ISBN

9780438538405

Share

COinS