Towards identifying the dynamic cellular patterns
underlying early Arabidopsis floral development
Jonathan Legrand1, Vincent Mirabet1, Frederic Boudon2, Léo Guignard2, Coralie Cellier1, Christophe Godin2, Jan Traas1, Arezki Boudaoud1, Yann Guédon2 & Pradeep Das1
1. Laboratoire de Reproduction et Développement des Plantes & LJC, ENS Lyon , Lyon, France 2. VirtualPlant, CIRAD / INRIA, Montpellier, France
1. Segmented tissues of early Arabidopsis flower development
2. Data management: Graph based approach and available variables
Black outlines are simplified cells represented by the colored vertex.
Green and dashed red edges are cell spatial neighborhood and temporal relations.
Number Size Shape
Spatial ●●CellsNeighbors ●●VolumeSurface (contact, epidermis)
●Axes of inertia (shape anisotropy) ●Gaussian and mean curvature
(epidermis, contact)
●Walls orientation
Temporal ●●Time intervalNumber of daugthers
Spatio-temporal ●Division rate
●Surface growth
●Volumetric growth
●Strain (rate, direction, anisotropy)
●Division plane orientation
Number of cells Relative Growth Rate t1 24 -t2 60 10 %.h-1 t3 151 8 %.h-1 t4 332 8 %.h-1 t5 468 4 %.h-1
Only L1 cells with a lineage from t1 to t5 have been counted
3. Spatial projection of quantified growth parameters
RGR(t)= 1 Δt .
(
∑ Vdaughters−∑ Vmothers
∑V mothers
)
Epidermal strain anisotropy
Relative volumetric growth rate ( h-1) Epidermal strain orientation
4. Exploratory analysis with manually annotated regions
Left sepal Adaxial sepal Central zone Peripheric zone Abaxial sepal Boundaries Right sepal
Expert definition of regions
5. Unsupervised clustering: towards global spatio-temporal cellular patterns
Aim: To identify and characterize cellular patterns based on measured cellular properties, linking their spatio-temporal behaviour to the
observed changes induced by organogenesis.
t3, N=7 t4, N=9 t2, N=4
t5, N=10
Relative volumetric growth (h-1)
Central zone region
t3, N=41 t4, N=70 t2, N=22
t5, N=98
Relative volumetric growth (h-1)
Boundaries region
t1 t3
t2 t4
Backward temporal projection of regions
Growth between t4 & t5
Conclusions and perspectives:
We propose a new methodology allowing the analysis of growing tissues in 3D using an integrative pipeline to access spatio-temporal parameters at a cellular resolution. We also provide tools for the definition of
cellular patterns based on the unsupervised clustering of spatio-temporal data. One important goal is to
determine how these cellular patterns diverge in mutants displaying extra sepals (e.g. perianthia, ettin), and to identify shifts or changes in properties within these groups or the presence or absence of certain groups.
Relative volumetric growth (h-1)
N=34 N=30 N=10 N=248 N=126 N=98 N=64 26h 44h 56h 69h t1 t2 t1 t3 t4 t5 0h 2 7 5 4 6 1 3 2, N=149 7, N=0 5, N=293 4, N=372 6, N=112 1, N=231 3, N=309 2, N=149 7, N=86 5, N=293 4, N=372 6, N=112 1, N=231 3, N=309
Growth between tn and tn+1 by clusters
Relative volumetric growth (h-1)
Cell volume (µm3)
Ward clustering on the whole time-course
Ward clustering method for seven clusters. Based on a pairwise distance matrix constructed with the following information: 60% volumetric growth, 20% volume and 20% topology.
* * * * * perianthia Volume by clusters
