PupStruct: Prediction of Pupylated Lysine Residues Using Structural Properties of Amino Acids

Abstract

Post-translational modification (PTM) is a critical biological reaction which adds to the diversification of the proteome. With numerous known modifications being studied, pupylation has gained focus in the scientific community due to its significant role in regulating biological processes. The traditional experimental practice to detect pupylation sites proved to be expensive and requires a lot of time and resources. Thus, there have been many computational predictors developed to challenge this issue. However, performance is still limited. In this study, we propose another computational method, named PupStruct, which uses the structural information of amino acids with a radial basis kernel function Support Vector Machine (SVM) to predict pupylated lysine residues. We compared PupStruct with three state-of-the-art predictors from the literature where PupStruct has validated a significant improvement in performance over them with statistical metrics such as sensitivity (0.9234), specificity (0.9359), accuracy (0.9296), precision (0.9349), and Mathew’s correlation coefficient (0.8616) on a benchmark dataset.

Electrostatic interactions guide substrate recognition of the prokaryotic ubiquitin-like protein ligase PafA

“…Attempts have been made to identify a substrate recognition pattern for PafA through bioinformatic analysis of the pupylome data sets. However, the results have been inconclusive 25 – 27 . In proteomics experiments for the identification of modified proteins, a common approach is to include an enrichment step during purification to separate the modified from unmodified proteins.…”

“…These studies identified hundreds of putative pupylation substrates, spanning proteins from diverse functional classes without a clear preference. Bioinformatic analysis of these pupylomes has revealed statistical differences in amino acid frequencies in the sequence neighborhood of target lysine residues compared to non-target lysines 25 – 27 . However, these patterns are insufficient to predict experimentally validated target lysines.…”

PUP-Fuse: Prediction of Protein Pupylation Sites by Integrating Multiple Sequence Representations

“…To know the molecular mechanisms of pupylation, it is necessary to define the substrates of pupylation and its sites precisely. Typically, this task is labor-intensive and time-consuming because of the large-scale analysis of proteomics; thus, a few computational methods of predicting pupylation sites have been proposed [ 12 , 13 , 14 , 15 , 16 ]. Liu et al first developed the GPS-PUP predictor for the prediction of pupylation sites by the group-based prediction system (GPS) method [ 17 ].…”