Bioengineering for multiple PAHs degradation for contaminated sediments: Response surface methodology (RSM) and artificial neural network (ANN)

Scientific community around the globe have major focus on designing bioremediation strategies for persistent, recalcitrant, highly toxic and carcinogen/mutagen polycyclic aromatic hydrocarbons (PAHs) present in marine environment. For the bioremediation strategy, components of growth medium are a key factor, which enhance degradation of the PAHs through simulating the microbial growth. Thus, present study involves bioengineering of growth medium (ONR7a) using response surface methodology (RSM) and artificial neural network (ANN) for enhanced multiple PAHs biodegradation. Microbes were isolated from contaminated sediments of Alang Sosiya Ship Breaking Yard (ASSBRY), Gulf of Khambhat, Gujarat, India. RSM - a process centric approach has resulted in an increase in PAHs degradation from 69% (Unoptimized) to 90.03% with 1.29 folds increase on 5th day with R-2 value of 0.98. Moreover, use of Artificial Neural Network (ANN) - a data centric approach resulted in better prediction of PAHs degradation of 93.36% compared to the CCD-RSM predicted PAHs degradation of 90.03% with R-2 value of 0.98. Based on various error functions such as mean absolute deviation (MAD), mean squared error (MSE), root mean squared error (RMSE) and mean absolute percentage error (MAPE), the predictive ability of the constructed ANN models was found to be higher compared to RSM. As this is the first ever report on PAHs degradation by bacterial mixed culture using data centric approach, this study bridges the gap between fundamental research and its application for policymakers and stakeholders which would be helpful in designing appropriate bioremediation technologies.