Recurrent neural network-based modeling of gene regulatory network using elephant swarm water search algorithm

Area of Science:

• Computational Biology
• Bioinformatics
• Systems Biology

Background:

• Inferring gene regulatory networks (GRNs) from time-series data is crucial but challenging in post-genomic research.
• Recurrent Neural Networks (RNNs) are commonly used for modeling gene dynamics and dependencies.

Purpose of the Study:

• To introduce a novel metaheuristic algorithm, the Elephant Swarm Water Search Algorithm (ESWSA), for accurate GRN inference.
• To leverage the social behavior and communication strategies of elephants for biological network modeling.

Main Methods:

• Developed ESWSA based on elephant water-seeking behavior during droughts.
• Tested ESWSA on artificial genetic networks with varying noise levels.
• Validated ESWSA using real gene expression data from the Escherichia Coli SOS Network.

Main Results:

• ESWSA demonstrated high efficiency in terms of parametric error, fitness value, and execution time on artificial networks.
• The algorithm achieved accurate prediction of true gene regulations.
• ESWSA outperformed other state-of-the-art optimization methods on the E. Coli SOS Network dataset.

Conclusions:

• ESWSA is a highly efficient and effective method for gene regulatory network inference.
• The bio-inspired ESWSA offers a promising alternative to existing optimization techniques for biological data analysis.