Quantifying bristle cell organization in Drosophila melanogaster using spatial clustering features

Area of Science:

• Developmental biology
• Biophysics
• Computational biology

Background:

• Developing tissues exhibit reproducible patterns despite cellular stochasticity.
• Quantifying complex cell and tissue spot patterns is a significant research challenge.
• The fruit fly thorax's sensory bristles serve as a model for studying pattern organization.

Purpose of the Study:

• To evaluate the utility of spatial clustering for quantifying and classifying cell and tissue spot patterns.
• To develop quantitative features for distinguishing between bristle patterns.
• To assess the effectiveness of these features in identifying known and novel biological perturbations.

Main Methods:

• Utilized the density-based spatial clustering of applications with noise (DBSCAN) algorithm.
• Proposed four novel clustering features derived from DBSCAN.
• Combined clustering features with pattern density for analysis.

Main Results:

• A combination of pattern density and cluster size variance distinguished 70% of patterns.
• The proposed features successfully differentiated between wild-type and mutant fly bristle patterns.
• The method demonstrated potential for confirming known perturbations and discovering new ones.

Conclusions:

• Spatial clustering offers a robust method for quantifying and classifying complex biological patterns.
• Pattern density and cluster size variance are key quantitative features for distinguishing tissue organization.
• This approach advances the study of developmental patterning and genetic perturbations.