Segmentation of occluded hematopoietic stem cells from tracking

Area of Science:

• * Stem cell biology
• * Live cell imaging and analysis
• * Computational biology

Background:

• * Analyzing live cell image sequences of proliferating hematopoietic stem cells (HSCs) presents challenges due to visual ambiguities.
• * HSCs require co-culture with OP9 stromal cells for long-term imaging, but frequently become occluded beneath the stromal layer.
• * This occlusion makes accurate cell segmentation and tracking difficult, even for expert analysis.

Purpose of the Study:

• * To develop an advanced computational approach for segmenting occluded hematopoietic stem cells (HSCs) in live cell time-lapse microscopy.
• * To improve the accuracy of cell tracking and lineage reconstruction in complex cellular environments.
• * To enable more robust analysis of HSC proliferation and development from image data.

Main Methods:

• * Implementation of an a priori segmentation using a Monte Carlo technique to estimate cell numbers in clumps.
• * Development of a tracking and lineage analysis pipeline for the image data.
• * An a posteriori resegmentation step, informed by tracking data, to delineate occluded HSCs.

Main Results:

• * The new segmentation approach successfully identified HSCs that were previously hidden or unresolvable.
• * Applied to 3031 occluded segmentations across 77 cell tracks, the method recovered over 84% of these challenging cases.
• * This significantly enhances the ability to track and determine the lineage of individual HSCs.

Conclusions:

• * The developed resegmentation technique effectively overcomes the limitations of occluded cell segmentation in HSC live imaging.
• * This computational advancement provides a more accurate and comprehensive understanding of hematopoietic stem cell behavior and development.
• * The method offers a valuable tool for researchers in stem cell biology and regenerative medicine.