General and Species-Specific Lysine Acetylation Site Prediction Using a Bi-Modal Deep Architecture

Abstract

Acetylation, as one of the most important post-translation modifications, plays a key role in a variety of biological functions, such as transcriptional regulation, cytokine signaling, and apoptosis. To understand the mechanism of acetylation profoundly, it is necessary to identify acetylation sites in proteins accurately. The existing methods for identifying protein acetylation sites can be divided into two major categories, i.e., mass spectrometry and computational methods. Mass spectrometry-based experimental methods are capable of discovering acetylation sites from eukaryotes, but can be time-consuming and expensive. Therefore, it is necessary to develop computational approaches that can effectively and accurately identify protein acetylation sites. The existing computational methods usually involve feature engineering, which may lead to redundancy and biased representations. While deep learning is capable of excavating the underlying characteristics from large-scale training data set via multiple-layer networks and non-linear mapping operations. In this paper, we propose a new method (named DeepAce) for predicting general and species-specific lysine acetylation sites based on deep neural network. We critically evaluate the performance of DeepAce and compare it with other existing predictors. The comparative results show the effectiveness of our Bi-modal deep architecture and also indicate that our method is very promising for predicting acetylation sites. The source code of DeepAce can be freely accessed at https://github.com/jiagenlee/DeepAce .