Planted (l, d) motif search using Bat algorithm with inertia weight and opposition based learning

Understanding the molecular mechanism of transcriptional regulation is the fundamental task of identifying and characterizing gene regulatory binding motifs using computational techniques. In computational biology, finding the planted ( ℓ , d ) motif search is a challenging task, even though it was addressed by various computational intelligence techniques, still it has the limitations of less accuracy and more computational time. Bat algorithm (BA) and together with two more techniques, namely inertia weight (IW) and opposition based learning (OBL) have been proposed to search planted ( ℓ , d ) motif. IW is used to control BA's exploration and exploitation capabilities, and OBL is in charge of reducing execution time by modifying the random directions. Particularly the constant, random inertia weight and quasi OBL improve the accuracy and execution time during the simulation. The proposed model was tested with the simulated and real dataset; it is 25 times faster than the hybrid method and 11 times faster than the DNAMotif_PSO for larger ( ℓ , d ) values. In the real dataset [ Escherichia coli cyclic AMP receptor protein (CRP)], the proposed model gives better accuracy when compared to the state-of-the-art methods and can find the transcription factor binding sites accurately in the simulated and real datasets with minimal time.