A Spatio-Temporal Model of Embolism Propagation in Leaf Vein Networks

Overview

This dataset contains the image dat, processed data, and code used in the paper "A Spatio-Temporal Model of Embolism Propagation in Leaf Vein Networks", authored by Diane Lu, Chris M. Smith-Martin, Robert Muscarella, María Uriarte, and Tian Zheng, currently accepted by AoB PLANTS, 2025, [DOI link].

Citation

If you use this code or data, please cite our paper:  Diane Lu, Chris M. Smith-Martin, Robert Muscarella, María Uriarte, Tian Zheng, A Spatio-Temporal Model of Embolism Propagation in Leaf Vein Networks, AoB PLANTS, 2025,  [DOI link].

Dataset Structure

Each zip file contains multiple folders, each representing a leaf sample of the same species (e.g., drygla.zip includes drygla2_leaf,  drygla3.2_leaf,  drygla4_leaf,  drygla5_leaf, drygla7_leaf, drygla8_leaf, drygla9_leaf):

speciesIndex_leaf: Folder for each individual leaf sample (e.g., drygla4_leaf)

raw image files (.png or .jpg): Raw image files, with timestamp as the filename

bw_labels.csv: A large CSV file showing each pixel's vein segment index (i.e., ori_region)

bw_labels_img.png: A preview of bw_labels.csv, where the color represents ori_region

real_data.RData: Survival data processed from the raw image files, can be used directly for the code.

The survival data (real_data.RData) contains the following elements:

X is a data frame with 5 cols: ori_region, v, delta,t, new_region. Note nrow(X)= total number of vein segments=max(X$new_region).

ori_region: the vein segment index used for plotting (overlaying vein segments have the same ori_region)

v: vein thickness

delta: event indicator. 1 if the vein segment is embolized during (T1,T2). 0 else

t: if delta==1 (observed emb event), t = the obs emb wp (-0.01MPa). Else (censored), t=NA.

new_region: the region index where overlaying vein segments would receive different indexes.

T1, T2: observation period

E_new: Adjacency matrix with respect to X$new_region (i.e. row index of X). Note that overlaying vein segments are considered not connected here.

main.R: an example code for spatial survival model using real_data.RData.

Code Usage

To get started, download drygla.zip and run the code (main.R) on the survival data (real_data.RData) to fit a spatial survival model for drygla4_leaf.

To fit survival models for all the data, download all the zip files, uncomment the for-loop in main.R, and then run the code (main.R).