Predicting Amyloid Proteins Using Attention-based Long Short-term Memory

Alzheimer's disease (AD) is one of the genetically inherited neurodegenerative disorders that mostly occur when people get old. It can be recognized by severe memory impairment in the late stage, affecting cognitive function and general daily living. Reliable evidence confirms that the enhanced symptoms of AD are linked to the accumulation of amyloid proteins. The dense population of amyloid proteins forms insoluble fibrillar structures, causing significant pathological impacts in various tissues. Understanding amyloid protein's mechanisms and identifying them at an early stage plays an essential role in treating AD as well as prevalent amyloid-related diseases. Recently, although several machine learning methods proposed for amyloid protein identification have shown promising results, most of them have not yet fully exploited the sequence information of the amyloid proteins. In this study, we develop a computational model for identification of amyloid proteins using bidirectional long short-term memory in combination with an attention mechanism. In the testing phase, our findings showed that the model developed by our proposed method outperformed those developed by state-of-the-art methods with an area under the receiver operating characteristic curve of 0.9126.