Modeling the shape of citrus and their oil gland distribution

UCR Citrus Collection       X-Ray CT Imaging       Raw 3D scans Spine          Endocarp         Rind           Exocarp       Oil glands


51 different citrus varieties

  • 166 samples X-ray CT scaned (90 micron resolution)
  • Extracted 5 tissues for all samples

Significant allometry across tissues' volumes

  • Size relationships correspond to hybridization patterns
  • E.g. Sour oranges are between mandarins and pummelos

Ellipsoid approximation of the whole fruit

  • Fruit sphericity with various indices: 0 (line/plane)→1 (sphere)

Size and average distance between closest glands has high R2 correlation coefficients

  • Smaller fruits report higher density of oil glands
  • Similar results when looking at the average distance between each oil gland and its 2nd, 3rd, ..., 25th nearest neighbors

The average distance to the oil gland nearest neighbors follows a square root relationship

  • The average distance from each oil gland to its k-th nearest neighbor follows a square root formula

\[\text{Avg.dist}(k) = \sqrt{Mk + B}\]

  • M is the distance increase rate; B an adjusting intercept
  • This formula suggests that oil glands might be distributing themselves following normal diffusion mechanics
  • Distance increase rate is related to oil gland density
  • The outliers for citrus groups usually correspond to hybrids
  • E.g. lemon outliers are Interdonatos, lemon x pummelo hybrids

Conclusions and takeaways

  • X-ray CT technology offers a new powerful lens to study citrus shape at a much more granular level, even oil glands!
  • Oil gland distribution can shed new light in citrus development
  • Shape patterns can give us insight into citrus hybridization

Acknowledgements

This work is supported in part by Michigan State University and the National Science Foundation Research Traineeship Program (DGE-1828149).