AbstractPost Transactional Modification (PTM) is a vital process which plays an important role in a wide range of biological interactions. One of the most recently identified PTMs is Malonylation. It has been shown that Malonylation has an important impact on different biological pathways including glucose and fatty acid metabolism. Malonylation can be detected experimentally using mass spectrometry. However, this process is both costly and time-consuming which has inspired research to find more efficient and fast computational methods to solve this problem. This paper proposes a novel approach, called SEMal, to identify Malonylation sites in protein sequences. It uses both structural and evolutionary-based features to solve this problem. It also uses Rotation Forest (RoF) as its classification technique to predict Malonylation sites. To the best of our knowledge, our extracted features as well as our employed classifier have never been used for this problem. Compared to the previously proposed methods, SEMal outperforms them in all metrics such as sensitivity (0.94 and 0.89), accuracy (0.94 and 0.91), and Matthews correlation coefficient (0.88 and 0.82), for Homo Sapiens and Mus Musculus species, respectively. SEMal is publicly available as an online predictor at: http://brl.uiu.ac.bd/SEMal/....

SEMal: Accurate protein malonylation site predictor using structural and evolutionary information

10.1016/j.compbiomed.2020.104022