The repository is currently being upgraded to DSpace 7. Temporarily, only admins can login. Submission of items and changes to existing items is prohibited until the completion of this upgrade process.
Deep learning-based approaches for prediction of post-translational modification sites in proteins
Abstract
Protein post-translational modification plays an important role in a myriad of biological processes. Computational prediction approaches serve as complementary methods for the characterization of post-translational modification sites in proteins. Computational prediction of N-linked glycosylation sites confined to N-X-[S/T] sequons is an important problem. This dissertation reports on DeepNGlyPred, a deep neural network-based approach for N-linked glycosylation sites PTM prediction and it encodes the positive and negative sequences in the human proteome dataset using sequence-based features (gapped-dipeptide), predicted structural features, and evolutionary information. Similarly, this dissertation presents LMNglyPred, a deep learning-based approach to predict N-linked glycosylated sites in human proteins using embeddings from a pre-trained protein language model.
To efficiently explore more undiscovered ubiquitylation sites, a novel multimodal deep learning architecture tool that identifies ubiquitination sites in proteins is studied. This study proposes a novel integrated deep learning-based approach named UbiIDN, for general ubiquitination site prediction, extracts and combines sequence and physicochemical properties information.
Moreover, a novel integrated deep learning-based approach named LMPhosSite, for general phosphorylation site prediction is developed. LMPhosSite extracts and combines sequence and protein language model information. Using an independent test set of experimentally identified N-linked glycosylation, ubiquitination, and phosphorylation sites the respectively developed predictors were able to outperform state-of-the-art predictors. These results demonstrate that developed predictors are a robust computational technique to predict PTM sites in proteins.
Description
Thesis (Ph.D.)-- Wichita State University, College of Engineering, School of Computing