Biological Sequence Representation Methods and Recent Advances: A Review

Abstract

Simple Summary This paper investigates how to convert biological sequences (protein and nucleotide sequences) into a format that computers can understand to better advance computational biology research. Our goal is to explain the principles and technical details of these methods, mainly to illustrate their application areas and advantages and disadvantages. We reviewed early techniques that relied on counting patterns, later methods that imitated language-processing techniques to capture context, and state-of-the-art methods based on large models in recent years. Our results show that new methods are more accurate but require advanced computing power. The development and improvement of these methods will help scientists design effective drugs, predict diseases, and reveal the connection between genetic material and proteins. We conduct a fundamental research work, this review can provide useful guidance and help for researchers in computational biology, especially those new to the field.