tavareshugo / publication_otero2022_phloempoleatlas

analysis code for Otero et al 2020

Code for analysis of scRNA-seq data of phloem ring

Assuming full pipeline is run, the file structure is as follows:

• data: contains all the data needed for the project
  • raw: the raw data files. This should be shared in public repositories upon publication.
  • intermediate: intermediate files used in the analysis. These can be re-created by running the workflow and can be removed to save space.
  • processed: data files that can also be re-created by running the workflow, but we keep them as they are used for making key figures/analysis. These should also be shared in a public repository upon publication.
  • external: contains external data downloaded from web (e.g. reference genome).
• logs: contains log files resulting from running different steps of the workflow.
• workflow: contains the actual workflow scripts. We're using snakemake for our pipelines. The file pipeline.pdf contains a diagram of the workflow.
  • Snakefile: is the script with the workflow specification
  • scripts: contains scripts used in the pipeline
  • envs: contain YAML files with the conda environment for each step of the pipeline.

Data

• 10x data generated in this project is deposited on SRA: accession GSE181999.
• The list of public datasets used for the integrated analysis is available here.
• For a quick look at our data, two files with processed results are attached to this repository's release. These are detailed below.

cell_metadata.csv

A CSV file containing information about our cell annotation. This can be opened in any spreadsheet or data analysis software. The columns in the table are:

• cell_id is a unique identifier of each cell (a combination of the sorted sample identifier and the 10x barcode).
• sample is the sorted sample identifier (refers to the marker used for cell-sorting).
• barcode is the 10x barcode of the cell.
• total_counts is the total UMI count detected in the cell.
• detected_genes is the number of genes with detected expression (at least 1 UMI).
• cluster is the cluster number used throughout the paper.
• annotation is the cell type annotation inferred in our paper.
• UMAP1 is the first axis of the UMAP projection used in the paper.
• UMAP2 is the second axis of the UMAP projection used in the paper.

SingleCellExperiment_filtered.rds

This is the SingleCellExperiment object (for use with R/Bioconductor packages), containing the full analysis of the data. The object contains:

• Matrices of raw and normalised counts (in the assays slots).
• Dimensionality reduction matrices including UMAP, t-SNE and diffusion maps (in the reducedDims slots).
• The trajectory analysis results from slingshot, whose pseudotime can be found in the colData slot of the object.

Here is some R code to read this object:

library(SingleCellExperiment)

# read the object
sce <- readRDS("SingleCellExperiment_filtered.rds")

# cluster assignment used in paper
colData(sce)$cluster_mnn_logvst

# assays
assay(sce, "counts")    # raw counts
assay(sce, "logvst")    # counts normalised using sctransform
assay(sce, "logcounts") # counts normalised using scran's deconvolution method

(re-)Running the pipeline

⚠️ This pipeline is not being maintained and some of the code may not run as expected.

To re-run the analysis (or if new samples are generated), this command alone should re-run everything (make sure you cd to the project's directory):

snakemake --use-conda

Note for SLCU HPC users - to submit the jobs on the cluster, do the following instead:

• run tmux (this will launch a terminal that will persist even when you logout)
• run: snakemake --use-conda --cluster "sbatch -c {threads} --mem-per-cpu=5G -J {rulename} -o logs/slurm/job#%j-{rulename}-{wildcards.sample}.log" --jobs 10
• you can exit tmux by using the keyboard shortcut 'Ctrl + b' followed by 'd'
• when you log back in, to get back your tmux session back run tmux attach -t 0

publication_otero2022_phloempoleatlas's Issues

where is the public_datasets.csv?

hi.

I do not find the dataset info. Would you please upload it? Also, the sample_info.csv is missing too.

Sorry, I do not know how much data I need to re reproduce the result in the paper.

Best
Mu