Columbia Technology Ventures

MicronuclAI: pipeline for automated and reliable micronuclei quantification to assess chromosomal instability

This technology is micronuclAI, a pipeline for automated and reliable quantification of micronuclei of varying size, morphology, and location from nuclei-stained images to assess chromosomal instability, a hallmark of cancer.

Unmet Need: Automated quantification of micronuclei applied to multiple cell lines

Chromosomal instability is a hallmark of cancer resulting in accumulation of micronuclei. However, current methods for micronuclei quantification include labor intensive cytogenetics, quantitative imaging, and single cell genomics. Quantitative imaging microscopy is most widely used to quantify micronuclei and is manually scored. During manual quantification, the workflow of quantifying micronuclei is tedious, time-consuming and error-prone, and due to no universal method, quantification varies between observers of a single image. Additionally, some automated methods can quantify micronuclei, but they are not uniform across cell lines. There is a need for a pipeline that automatically quantifies nuclei across a wide range of cell lines.

The Technology: Pipeline for automated and reliable micronuclei quantification

This technology is micronuclAI, a deep learning-based pipeline that automatically quantifies micronuclei from nuclei-stained images for the assessment of chromosomal instability. The pipeline uses segmentation, nuclei isolation, and quantification to reliably quantify micronuclei across multiple cell lines that are ready to use in laboratories. A dataset using micronuclAI was compared against manual single-cell level counts by experts and against routinely used micronuclei ratio, where the classifier was able to achieve a weight F1 score of 0.937. This technology increases speed, accuracy, and robustness of chromosomal instability research.

This technology has been validated with multiple human and murine cancer cell lines.

Applications:

  • Micronuclei quantification
  • Tool for cancer research
  • Tool to quantify chromosomal instability
  • Cell monitoring

Advantages:

  • Automates micronuclei counting
  • Standardizes method of performance
  • Saves time
  • Compatible across different cell lines
  • Reduces errors
  • Nears human-level performance

Lead Inventor:

Benjamin Izar, MD, Ph.D.

Related Publications:

Tech Ventures Reference: