Prediction of seafloor ecological state using 16S nanopore sequencing

Area of Science:

• Marine ecology
• Environmental DNA (eDNA)
• Bioinformatics

Background:

• Aquaculture and anthropogenic activities impact benthic habitats, necessitating efficient monitoring tools.
• Environmental DNA (eDNA) and advanced sequencing technologies like Oxford Nanopore offer rapid, on-site ecosystem assessment capabilities.
• While Nanopore sequencing shows promise for ecological predictions, its accuracy requires further validation against established methods.

Purpose of the Study:

• To predict the seafloor ecological state using both Illumina and Nanopore 16S rRNA sequencing data.
• To evaluate the performance of different bioinformatic approaches and machine learning for ecological prediction.
• To compare the accuracy and feasibility of Nanopore sequencing against Illumina for marine benthic monitoring.

Main Methods:

• Analysis of 88 seafloor samples along a Norwegian coastal gradient.
• Application of machine learning algorithms combined with feature selection (LASSO regression) for data analysis.
• Comparison of predicted ecological index values (nEQR) derived from both sequencing platforms against macroinvertebrate data.

Main Results:

• Illumina and Nanopore sequencing platforms yielded comparable predictions for seafloor ecological state.
• Stabilized LASSO regression optimized feature sets from thousands to 40-60 Operational Taxonomic Units (OTUs), reducing prediction errors by over 50%.
• High predictive accuracy was achieved, with Pearson correlation coefficients of 0.98 for Illumina and 0.95 for Nanopore data.

Conclusions:

• Nanopore sequencing is a viable and cost-effective alternative to Illumina for marine benthic monitoring.
• Optimized feature selection is crucial for enhancing prediction accuracy and reducing computational demands.
• Continuous improvements in Nanopore sequencing technology and bioinformatic approaches enable precise ecological assessments.