A novel automated protocol for ice crystal segmentation analysis using Cellpose and Fiji

Area of Science:

• Materials Science
• Biophysics

Background:

• Accurate ice crystal size measurement is crucial for quantitative ice recrystallization inhibition (IRI) analysis.
• Current methods like the sucrose sandwiching assay (SSA) and splat assay (SA) are time-consuming and limited in scope.

Purpose of the Study:

• To develop a novel, automated method for measuring ice crystal size and shape.
• To overcome the limitations of traditional methods in IRI analysis.

Main Methods:

• Utilized Fiji and Cellpose, an anatomical segmentation algorithm, for rapid image segmentation (∼70 s per image).
• Implemented automated measurement of all ice crystals within an image field of view.
• Employed Feret's diameter to accurately characterize irregular particle shapes.

Main Results:

• The automated method significantly reduces bias by measuring every ice crystal, unlike subjective largest crystal selection.
• Demonstrated consistency in measuring diverse ice crystal sizes and shapes.
• Generated new outputs including standard deviation, particle size distributions, and circularity for comprehensive analysis.

Conclusions:

• This novel approach provides a more accurate and efficient method for quantitative ice crystal analysis in IRI studies.
• The "shape descriptor" measurement can quantify ice binding, particularly applicable to the SSA.
• Offers enhanced insights into an analyte's IRI ability through detailed ice crystal characterization.