Automated computer vision and dose-response modeling improve throughput and accuracy of an ex vivo functional precision medicine platform

  • 0Eshelman Innovation, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA.

|

|

Summary

This summary is machine-generated.

This study introduces an automated workflow for brain slice assays, significantly reducing analysis time and improving consistency. The approach uses computer vision and dose-response modeling for more reliable functional precision medicine research.

Area Of Science

  • Neuroscience
  • Pharmacology
  • Computational Biology

Background

  • Functional Precision Medicine (FPM) faces challenges in clinical translation due to analytical inconsistencies.
  • Organotypic brain slice cultures are valuable for drug screening but require robust, consistent analysis.

Purpose Of The Study

  • To develop an automated data analysis workflow for organotypic brain slice culture assays.
  • To enhance analytical consistency and throughput for FPM applications.

Main Methods

  • Integration of computer vision and dose-response modeling for automated data analysis.
  • Utilization of cloud-based Machine Learning (ML) tools (e.g., Biodock) with local scripts.
  • Development of strategies for modeling diverse dose-response behaviors (hormesis, plateau effects).

Main Results

  • Analysis time reduced by approximately 99% (from ~20 hours to ~15 minutes for an 11-drug assay).
  • Automation significantly increased consistency within and across experiments, minimizing human subjectivity.
  • Workflow demonstrated effective implementation with limited computational resources and programming expertise.

Conclusions

  • The automated workflow addresses key obstacles in the clinical translation of FPM assays.
  • This approach enhances reliability and efficiency for functional assays, even in the absence of gold-standard measurements.
  • The study provides a practical framework for implementing automated analysis in complex biological systems.