Caenorhabditis elegans Multi-Tracker Based on a Modified Skeleton Algorithm.

Area of Science:

• Computational Biology
• Bioimage Analysis
• Neuroscience

Background:

• Automatic tracking of Caenorhabditis elegans (C. elegans) is crucial for behavioral analysis.
• Existing tracking methods fail when worms aggregate, overlap, or make body contact, leading to data loss.
• Automating the analysis of worm contact behaviors requires robust solutions for maintaining individual worm identity.

Purpose of the Study:

• To develop a computer vision-based solution for accurate Caenorhabditis elegans tracking during aggregation and contact.
• To address the challenge of maintaining individual worm identity in crowded experimental conditions.

Main Methods:

• Application of a skeletonization method to extract worm skeletons in situations of overlap and contact.
• Development of novel optimization methods to resolve worm identity issues during aggregation.
• Evaluation of various cost functions and criteria using experimental data from 70 tracks and 3779 poses.

Main Results:

• The modified skeleton algorithm achieved 99.42% accuracy in identifying worms during overlaps and in the presence of noise.
• The classical skeleton algorithm demonstrated 98.73% precision.
• The proposed methods effectively solve the problem of lost worm identity during aggregation.

Conclusions:

• The developed computer vision techniques significantly improve the accuracy and reliability of Caenorhabditis elegans tracking.
• This solution enables more comprehensive automated analysis of worm behaviors, including social interactions.
• The findings pave the way for advanced research in C. elegans behavior and neurobiology.