Biology subjects

McCarroll, M. N.

Publications and source records attributed to McCarroll, M. N..

2 recordsLinked to original sources

bioRxiv2025

High-throughput multi-camera array microscope platform for automated 3D behavioral analysis of freely swimming zebrafish larvae

Understanding the behavioral and morphological dynamics of moving model organisms like the zebrafish larvae requires accurate, high-throughput 3D analysis. However, traditional single-view 2D video tracking fails to capture the full scope of natural 3D movements and postural dynamics. Here, we present a novel high-throughput 24-camera array microscope with a co-designed "mirrored well plate" that allows for snapshot imaging of up to 48 wells over a 118 mm x 82 mm field of view from two orthogonal directions (i.e., a top-view and side-view). Accurate 3D position estimation and tracking is achieved with an efficient machine learning algorithm that scales well to high-throughput measurements. The proposed approach automates parallelized 3D model organism behavioral analysis, providing 3D skeletal tracking, swim bladder morphological dynamics, and kinematics of up to 48 swimming zebrafish larvae at up to several hundred frames per second. The result is an efficient and scalable solution for high-throughput 3D behavioral studies with broad compatibility with standard workflows across laboratories and procedures working with pharmacology, toxicology, and neuroscience.

animal behavior and cognition↗

bioRxiv2024

Deep phenotypic profiling of neuroactive drugs in larval zebrafish

Behavioral larval zebrafish screens leverage a high-throughput small molecule discovery format to find neuroactive molecules relevant to mammalian physiology. We screened a library of 650 central nervous system active compounds in high replicate to train a deep metric learning model on zebrafish behavioral profiles. The machine learning initially exploited subtle artifacts in the phenotypic screen, necessitating a complete experimental re-run with rigorous well-wise randomization. These large matched phenotypic screening datasets (initial and well-randomized) provided a unique opportunity to quantify and understand shortcut learning in a full-scale, real-world drug discovery dataset. The final deep metric learning model substantially outperforms correlation distance-the canonical way of computing distances between profiles-and generalizes to an orthogonal dataset of novel druglike compounds. We validated predictions by prospective in vitro radio-ligand binding assays against human protein targets, achieving a hit rate of 58% despite crossing species and chemical scaffold boundaries. These newly discovered neuroactive compounds exhibited diverse chemical scaffolds, demonstrating that zebrafish phenotypic screens combined with metric learning achieve robust scaffold hopping capabilities.

bioinformatics↗