eDNAssay

eDNAssay predicts qPCR cross-amplification in environmental DNA (eDNA) sampling to assess assay specificity when closely related species co-occur.

Key Features:

• Machine learning approach: Uses random forest classifiers to predict qPCR cross-amplification.
• Model types: Provides a primer-only model using SYBR Green intercalating dye results and a full-assay model based on TaqMan hydrolysis probe data.
• Cross-validation performance: Reports 99.6% accuracy for the primer-only model and 100% accuracy for the full-assay model in cross-validation tests.
• Independent validation: Validated on six independent assays not used in training with accuracies of 92.4% for the primer-only model and 96.5% for the full-assay model.

Scientific Applications:

• qPCR assay development: Enables in silico assessment of assay specificity to reduce the need for extensive in vitro testing during qPCR assay design.
• eDNA-based monitoring: Supports wildlife monitoring and biodiversity studies by improving the reliability of species detection and abundance estimates from environmental DNA samples.

Methodology:

Random forest classifiers were trained on data from 82 synthetic gene fragments across 10 qPCR assays, comprising 530 specificity tests conducted using SYBR Green intercalating dye and TaqMan hydrolysis probes.