JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

[A digital droplet PCR detection technique based on filter faster R-CNN]

  • 0School of Biomedical Engineering, Southern Medical University, Guangzhou 510515, China.
Nan Fang Yi Ke Da Xue Xue Bao = Journal of Southern Medical University +

|

March 19, 2024

|

March 19, 2024

View abstract on PubMed

Summary

This summary is machine-generated.

This study introduces a novel Filter Faster R-CNN model to improve digital droplet PCR (ddPCR) accuracy by removing image anomalies. The model ensures stable and precise ddPCR detection even in challenging, dusty environments.

Area Of Science

  • Biotechnology
  • Molecular Biology
  • Bioinformatics

Context

  • Digital droplet PCR (ddPCR) is a sensitive nucleic acid quantification method.
  • Image analysis in ddPCR can be affected by anomalies like dust and scratches.
  • High-throughput, stable, and accurate ddPCR detection is crucial for various applications.

Purpose

  • To develop and validate a method mitigating the impact of image anomalies on ddPCR detection.
  • To enhance the accuracy and robustness of ddPCR results in diverse environmental conditions.

Summary

  • A Filter Faster R-CNN ddPCR detection model was proposed, integrating Faster R-CNN with an outlier filtering module.
  • The model demonstrated superior detection accuracy (98.23% in low-dust, 88.35% in dusty environments) and high F1 scores (99.15%, 99.14%).
  • Absolute quantification experiments showed high consistency with commercial flow cytometry (R²=0.9997).

Impact

  • Provides a robust ddPCR detection method effective under various environmental conditions.
  • Significantly improves positive droplet detection accuracy in environments with image anomalies.
  • Enables high-throughput, stable, and accurate ddPCR analysis, advancing molecular diagnostics.

Keywords: