A computational approach for nuclear export signals identification using spiking neural P systems

Nuclear export signal (NES) is a nuclear targeting signal within cargo proteins, which is involved in signal transduction and cell cycle regulation. NES is believed to be born to be weak''; hence, it is a challenge in computational biology to identify it from high-throughput data of amino acid sequences. This work endeavors to tackle the challenge by proposing a computational approach to identifying NES using spiking neural P (SN P) systems. Specifically, secondary structure elements of 30 experimentally verified NES are randomly selected for training an SN P system, and then 1224 amino acid sequences (containing 1015 regular amino acid sequences and 209 experimentally verified NES) abstracted from 221 NES-containing protein sequences randomly in NESdb are selected to test our method. Experimental results show that our method achieves a precision rate 75.41 %, better than NES-REBS 47.2 %, Wregex 25.4 %, ELM, and NetNES 37.4 %. The results of this study are promising in terms of the fact that it is the first feasible attempt to use SN P systems in computational biology after many theoretical advancements.