This resource provides the R code to reproduce key results described in Greenwald, Galili Darnell, Hoefflin, et al. "Integrative spatial analysis reveals a multi-layered organization of glioblastoma".
The analyses are divided into 6 main modules:
Module 1: Per sample clustering, metaprogram generation, and spot annotation
Module 2: Spatial coherence and organizational zones
Module 3: Measures of spatial associations
Module 4: defining consensus interactions
Module 5: Spatial CNA inference
Module 6: CODEX analysis
- Clone Github repository.
- Download and extract the data provided in Inputs.zip
- Set the working directory to Inputs.
- Run one of the 6 code modules in R.
Please note results of modules 1-3 might slightly differ dependending on the version of R/R packages used.
Each code module can be run independently.
The code uploaded here is the working code being used throughout the work on the project. We are currently working on generating a more user-friendly, readable and easy to run code.
The MP generation approach is based on our earlier work described in Kinker et al. 2020 and Gavish et al. 2023. Further documentation can be found here and here.
The Inputs file contains an additional README regarding alignment of Visium and CODEX samples.
You can find the Visium H&E here.
Requirements R (tested in version 4.1.1).
