Prediction Of O-Glycosylation Site Using Pre-Trained Language Model And Machine Learning

Abstract

O-glycosylation is a typical type of protein post-translational modifications (PTMs), which is linked to several diseases and has significant roles in many biological processes. Identification of Oglycosylation sites is important to know the mechanism of the O-glycosylation process. However, the identification of PTM sites by laboratory experimental tools is time and money-consuming. Thus, the utilization of computational and artificial intelligence is becoming essential to predict O-glycosylation sites. In this paper, we proposed a new model to improve O-glycosylation site prediction using a transformer-based protein language model and machine learning. The dataset was collected and prepared from a recent data source called OGP (O-glycoprotein repository). The TAPE (Tasks Assessing Protein Embeddings) protein language model was used to feature extraction from the peptide sequences using the embedding strategy. Then, feature selection was implemented using the linear support vector machine (SVM) to select informative features. The XGBoost ensemble-based machine learning method was utilized for classification and prediction. The proposed model achieved high-performance results with 0.7761 accuracy, 0.7391 sensitivity, 0.8130 specificity, 0.8295 AUC, and 0.5537 MCC when compared with the traditional machine learning methods. On an independent dataset, the proposed method performed better than the latest available methods for predicting O-glycosylation sites.

HumanO-linked Glycosylation Site Prediction Using Pretrained Protein Language Model

“…These features were merged and utilized for training the SVM classifier. Recently, Alkuhlani et al 24 developed TAPE PLM-based 25 O- linked glycosylation sites predictor and used XGBoost to classify the O- glycosylation site in the proteins.…”

“…However, they have not extensively explored other complementary and recently developed PLMs. The O- linked glycosylation prediction method of Alkuhlani et al 24 still uses a window size of 31 around site of interest to extract the TAPE-based PLM features. Additionally, this method uses SVM based feature selection approaches which are indeed meticulous, yet also onerous.…”

“…The majority of the methods mentioned above [26][27][28][29][30][31] have their input features manually curated for prediction. Furthermore, the O-linked glycosylation prediction method from Alkuhlani et al 24 is the only one that leveraged the benefits of embeddings from large protein language models (TAPE 25 ). However, they have not extensively explored other complementary and recently developed PLMs.…”

Human O-linked Glycosylation Site Prediction Using Pretrained Protein Language Model

“…These features were merged and utilized for training the SVM classi er. Recently, Alkuhlani et al 24 developed TAPE PLM-based 25 O-linked glycosylation sites predictor and used XGBoost to classify the O-glycosylation site in the proteins.…”

“…The majority of the methods mentioned above [26][27][28][29][30][31] have their input features manually curated for prediction. Furthermore, the O-linked glycosylation prediction method from Alkuhlani et al 24 is the only one that leveraged the bene ts of embeddings from large protein language models (TAPE 25 ). However, they have not extensively explored other complementary and recently developed PLMs.…”