Deep neural networks for 3D protein structural analysis
Abstract/Contents
- Abstract
- Central to protein biology is the understanding of how structural elements give rise to observed function. The surfeit of protein structural data enables development of computational methods to systematically derive rules governing structural-functional relationships. However, performance of these methods depends critically on the choice of protein structural representation. Most current methods rely on features that are manually selected based on knowledge about protein structures. These are often general-purpose but not optimized for the specific application of interest. In this dissertation, I present a general framework that applies deep neural networks to structure-based protein analysis. The framework automatically extracts task-specific features from raw atom distributions or simple descriptors, driven by supervised labels. As a pilot study, I applied 3D convolutional neural networks (3DCNNs) to analyze the 20 amino acid microenvironments and predict the amino acids most compatible with a given location in a protein. Our 3DCNN achieves a two-fold increase in prediction accuracy compared to models that employ conventional engineered features and successfully recapitulates known information about similar and different microenvironments. I further applied the 3DCNN framework to detecting protein functional sites and demonstrated that 3DCNNs performed better on site detection tasks compared to recently published methods. Finally, I propose a graph-convolutional (Graph-CNN) framework for predicting protein-ligand interactions and show that Graph-CNNs achieved better or comparable performance to recent methods on virtual screening tasks, without relying on protein-ligand co-complexes. End-to-end trained deep learning networks consistently outperform methods using hand-engineered features on various important biological tasks, suggesting that the 3DCNN and Graph-CNN frameworks are well suited for protein structural analysis and hold great promise for advancing our understanding in protein science and drug discovery.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2019; ©2019 |
| Publication date | 2019; 2019 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Torng, Wen |
|---|---|
| Degree supervisor | Altman, Russ |
| Thesis advisor | Altman, Russ |
| Thesis advisor | Fordyce, Polly |
| Thesis advisor | Huang, Possu |
| Thesis advisor | Kundaje, Anshul, 1980- |
| Degree committee member | Fordyce, Polly |
| Degree committee member | Huang, Possu |
| Degree committee member | Kundaje, Anshul, 1980- |
| Associated with | Stanford University, Department of Bioengineering. |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Wen Torng. |
|---|---|
| Note | Submitted to the Department of Bioengineering. |
| Thesis | Thesis Ph.D. Stanford University 2019. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2019 by Wen Torng
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...