Kernel extreme learning with harmonized bat algorithm for prediction of pyrene toxicity in rats

Area of Science:

• Environmental Toxicology
• Bioinformatics
• Computational Biology

Background:

• Polycyclic aromatic hydrocarbons (PAHs) are toxic organic pollutants with significant human health implications.
• Pyrene, a specific PAH, was investigated for its harmful effects on rat models.
• Understanding PAH toxicity mechanisms requires effective feature selection from complex biological datasets.

Purpose of the Study:

• To investigate the toxicological impact of pyrene on rat gene expression.
• To develop and validate a high-performance feature selection strategy for PAH toxicity assessment.
• To identify specific rat genes most affected by PAH exposure.

Main Methods:

• Development of a novel optimization method, SCBA (Sequence Crossover Bat Algorithm).
• Integration of SCBA with the Kernel Extreme Learning Machine (KELM) model (SCBA-KELM).
• Validation of SCBA-KELM performance using benchmark functions and comparative experiments on rat gene expression data.

Main Results:

• The SCBA-KELM model demonstrated high accuracy and stability in feature selection.
• Key genes significantly impacted by pyrene exposure were identified: PXR, CAR, CYP2B1/2, and CYP1A1/2.
• The SCBA-KELM model achieved 100% classification performance on the gene dataset and ~96% precision on a public dataset.

Conclusions:

• The SCBA-KELM model is a reliable and effective tool for toxicological classification and assessment.
• The identified genes provide insights into the molecular mechanisms of PAH toxicity in rats.
• This approach offers a valuable method for analyzing animal datasets in toxicological studies.