About
A combinatorial in vitro/in silico NAM for automated morphological phenotyping of developmental neurotoxicity.
ConductScreen
ConductScreen is a combinatorial New Approach Methodology (NAM) that integrates two complementary modalities: in vitro human iPSC-derived neural organoid culture and in silico automated computational image analysis. The platform addresses the growing crisis of prenatal substance exposure: Fetal Alcohol Spectrum Disorders alone affect an estimated 1 in 20 U.S. children (~180,000 affected births per year), making it the leading preventable cause of developmental disability.
The morphology module extracts 16 morphological, intensity, and texture features from standard brightfield microscopy images and applies rigorous non-parametric statistics to detect neurodevelopmental phenotypes — entirely automated, observer-independent, and animal-free. Published studies across 9 substance classes (ethanol, VPA, opioids, stimulants, and others) show the same types of morphological changes our pipeline is validated to detect.
Our Approach
A three-phase plan progressing from feasibility demonstration to regulatory-grade validation.
Phase 1 — Feasibility
Validate that the morphology pipeline reliably detects neurodevelopmental morphological differences using published organoid data (1,407 images, 64 organoids, 4 cell lines, 2 labs). Establish concordance between validated features and published substance-exposure effects across 9 substance classes.
Current phase
Phase 2 — Validation
Apply the validated pipeline to substance-exposed organoids (ethanol, VPA, and select opioids at 3 concentrations each). Establish dose-response morphological relationships. 3+ lab reproducibility study with FAIR-compliant NDHCC data submission. Integrate the electrophysiology module.
Phase 3 — Translation
Working platform transfer to VQN with operating instructions. Comprehensive opioid dose-response dataset. Documentation for independent testing, validation, and assessment through the Complement-ARIE VQN framework.
The Team
The ConductScreen team combines academic and industry expertise across clinical medicine, addiction biology, computational engineering, and statistics.
Shuhan He, MD
PI / Clinical Translation
MGH / Harvard Medical School; CEO, ConductScience
Platform architecture, statistical framework, project leadership.
Shivani Pimparkar, MS
Co-PI / Addiction Biology & Disease Modeling
Boston University Bioinformatics; MGH Translational Research
Biological validation, opioid-organoid evidence base, Phase 2 experimental design.
Yijian Henry He, PhD
Technical Lead
PhD Economics; CTO, ConductScience
System architecture, cloud infrastructure, API development.
Louise Corscadden, PhD
Manufacturing, Compliance & Partnerships
PhD Molecular Genetics; Director, ConductScience
Manufacturing oversight, regulatory coordination, FAIR compliance, partnerships.
Bifei Hao, MS
Computer Vision & Supply Chain
Northwestern University EE; ConductScience
Computer vision pipeline, supply chain coordination.
Boyu Peng, MS
Platform Engineering
ConductScience
Deployment pipeline, reproducibility verification.
Santosh Adhikari, MS
ML & Interpretability
Computer vision and robotics; ConductScience
Pipeline interpretability, extension modules.
Allison C. Goff, Ph.D.
Wet-Lab Lead & Substance Exposure Biology
Georgetown Neuroscience; MGH IHP Faculty; Former NIH NIAAA/NIMH Fellow
Wet-lab experimental design, substance exposure protocols, neuroscience domain expertise.
ConductScience, Inc.
Scientific equipment, software, and data services for the research community. ConductScience provides the infrastructure for developing, deploying, and distributing computational tools for biomedical research.
conductscience.comContact
For questions about the ConductScreen platform, the analysis pipeline, or collaboration opportunities:
Explore the Platform
See the analysis results, compare disease models, or read about our computational methods.