SETApp

SETApp automates classification and quantification of sea urchin embryo phenotypes to support identification of toxicants in complex environmental samples.

Key Features:

• Machine Learning Integration: Employs partial least squares discriminant analysis (PLS-DA) models to classify larvae based on size increase and malformation levels.
• Hierarchical PLS-DA: Implements Hierarchical PLS-DA as an enhanced classification approach with reported superior performance.
• Prediction Accuracy: Achieves 84% prediction accuracy during validation using the hierarchical PLS-DA approach.
• Automated Quantification: Automates measurement of larval size increase and malformation from image data for high-throughput assessments.
• Training Dataset: Calibrated using a training set of 242 images to define size-increase and malformation levels of sea urchin larvae.

Scientific Applications:

• Automated Toxicant Identification: Supports identification of toxicants by quantifying larval responses (size increase and malformation) in bioassays combined with chemical analysis.
• Effect-Directed Analysis Support: Has been applied in wastewater treatment plant (WWTP) effect-directed analyses for toxicant identification in complex mixtures.
• High-Throughput Ecotoxicology Screening: Facilitates large-scale in vivo/in vitro screening workflows using the sea urchin embryo test as an endpoint.

Methodology:

Trained on 242 images to calibrate size-increase and malformation metrics and constructed two classification models based on PLS-DA, with Hierarchical PLS-DA showing superior classification performance.