Analysis of Medicago-AMF single-nuclei and spatial transcriptomics datasets

This is a repository for the publication "Spatial Co-transcriptomics Reveals Discrete Stages of Arbuscular Mycorrhizal Symbiosis" in which authors applied single-nuclei RNA-seq and spatial RNA-seq to the symbiosis between Medicago truncatula and Rhizophagus irregularis.

---

All figures can be produced using code found in the following Juptyer Notebooks:

1. sNucRNAseq_med_AMF.ipynb: Figure 1a, Figure 3a, Figure 3b, Figure 4b, Figure 4c, Supp. Sheet 4
2. sNucRNAseq_med_all.ipynb: Supp Figure 3
3. spatialRNAseq_pre_integration.ipynb: spatial data conversion into seurat object + filtering/normalization
4. spatialRNAseq_integration.ipynb: spatial object integration, Supp Figure 4
5. spatialRNAseq_post_integration.ipynb: Figure 2b, Figure 2c, Figure 4a (data), Figure 4b, Figure 5c
6. colonizationanalysis.ipynb: Supp Figure 2

Fastq files, Cellranger and Spaceranger matrix files, and fully processed .RDS Seurat Object files can be downloaded here: https://www.ncbi.nlm.nih.gov/geo/info/linking.html](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE240107.

The following packages are required:
"Seurat" "ggplot2" "patchwork" "dplyr" "here" "tidyverse" "viridis" "lattice" "reshape2" "cowplot" "Matrix" "Matrix.utils" "edgeR" "S4Vectors" "SingleCellExperiment" "pheatmap" "apeglm" "png" "DESeq2" "RColorBrewer" "data.table"

---

For optimal reproducibility, it is best to use the docker image listed below rather than rely on packages downloaded locally. You may also find it necessary to use NERSC or a cloud computing platform to analyze larger datasets; below are instructions for pulling the docker image during an interactive R session on NERSC or in a NERSC-hosted Jupyter notebook (preferred but a few more steps to set up).

---

Very basic instructions for pulling a docker image that supports these Jupyter Notebooks:

1. docker is installed on your computer and running (might as well be the latest version)
2. adjust docker preferences>resources to at least 8GB memory and at least 16GB disk image size
3. in terminal:

docker pull bcoli/renv_single_cell:2.1.1

docker run --rm --mount type=bind,source="your-project-directory",target="/home/rstudio/" -p 8787:8787 bcoli/renv_single_cell:2.1.1

4. in web browser, type: http://localhost:8787

Username: rstudio

PW: provided in terminal

To use this docker image for R on NERSC (for larger objects)

1. be in an interactive node
2. be in directory where files are

run:

shifterimg pull bcoli/renv_single_cell:2.0.1 shifter --image bcoli/renv_single_cell:2.0.1 R

To use R in a Jupyter notebook on NERSC

To create a kernel to use shifter to connect to the bcoli/renv_single_cell:2.1.1 docker image:

Log in to Perlmutter and navigate to: ~/.local/share/jupyter/kernels

mkdir seurat-shifter cd seurat-shifter

vim kernel.json

this kernel.json should contain the following:

{ "argv": ["shifter","--image=bcoli/renv_single_cell:2.1.1","R","--slave", "-e", "IRkernel::main()", "--args", "{connection_file}"],
"display_name": "seurat_shifter", "language": "R" }

Go to 'jupyter.nersc.gov' and select a CPU (shared or exclusive) on Perlmutter– this kernel should now be available

R objects identities:

see doc

reference genomes:

Medicago truncatula genome assembly and annotation: MedtrA17_4.0

Rhizophagus irregularis genome assembly and annotation: Rir_HGAP_ii_V2 (DAOM 181602, DAOM 197198) .

Please reach out to Karen Serrano at karenserrano@lbl.gov or Margot Bezrutcyzk at mbezrutczyk@lbl.gov for any questions.