%A Rahu, Sikandar %A Ghulam, Ali %A Farman, Ali %A Talpur, Dhani Bux %A Talpur, Mir Sajjad Hussain %A Saba, Erum %A Maher, Zulfikar Ahmed %A Tunio, Saima %D 2022 %T UBI-XGB: IDENTIFICATION OF UBIQUITIN PROTEINS USING MACHINE LEARNING MODEL %K %X A recent line of research has focused on Ubiquitination, a pervasive and proteasome-mediated protein degradation that controls apoptosis and is crucial in the breakdown of proteins and the development of cell disorders, is a major factor.  The turnover of proteins and ubiquitination are two related processes. We predict ubiquitination sites; these attributes are lastly fed into the extreme gradient boosting (XGBoost) classifier. We develop reliable predictors computational tool using experimental identification of protein ubiquitination sites is typically labor- and time-intensive. First, we encoded protein sequence features into matrix data using Dipeptide Deviation from Expected Mean (DDE) features encoding techniques. We also proposed 2nd features extraction model named dipeptide composition (DPC) model. It is vital to develop reliable predictors since experimental identification of protein ubiquitination sites is typically labor- and time-intensive. In this paper, we proposed computational method as named Ubipro-XGBoost, a multi-view feature-based technique for predicting ubiquitination sites. Recent developments in proteomic technology have sparked renewed interest in the identification of ubiquitination sites in a number of human disorders, which have been studied experimentally and clinically.  When more experimentally verified ubiquitination sites appear, we developed a predictive algorithm that can locate lysine ubiquitination sites in large-scale proteome data. This paper introduces Ubipro-XGBoost, a machine learning method. Ubipro-XGBoost had an AUC (area under the Receiver Operating Characteristic curve) of 0.914% accuracy, 0.836% Sensitivity, 0.992% Specificity, and 0.839% MCC on a 5-fold cross validation based on DPC model, and 2nd 0.909% accuracy, 0.839% Sensitivity, 0.979% Specificity, and 0. 0.829% MCC on a 5-fold cross validation based on DDE model. The findings demonstrate that the suggested technique, Ubipro-XGBoost, outperforms conventional ubiquitination prediction methods and offers fresh advice for ubiquitination site identification. %U https://journal.kiu.edu.pk/index.php/JMAR/article/view/167 %J Journal of Mountain Area Research %0 Journal Article %R 10.53874/jmar.v8i0.167 %P 14-26%V 8 %@ 2518-850X %8 2022-12-02