This repository is a copy of https://github.com/pachterlab/BP_2021 with the modifications described in "A like-for-like comparison of lightweight-mapping pipelines for single-cell RNA-seq data pre-processing". The original repository corresponds to the preprint "Benchmarking of lightweight-mapping based single-cell RNA-seq pre-processing " by A. Sina Booeshaghi and Lior Pachter.

How to run the pipeline

Note: All the addresses are relative to the main directory of the repository.

Create the following directories in the main directory of the repository

• navigate to ./analysis/scripts
• Run $bash make_dirs.sh

Download cellranger barcodes

• navigate to ./analysis/scripts
• Run $bash gather_cr_barcodes.sh

Download the samples

• Make sure you have the sratools 2.9 is installed on your system (https://ftp-trace.ncbi.nlm.nih.gov/sra/sdk/2.9.6/sratoolkit.2.9.6-ubuntu64.tar.gz)
• navigate to ./analysis/scripts/
• Run $bash gather_data.sh Downloads the fastq files and moves the ones for each sample to a directory called samples/{species}-{sample-name}

Download the references

• Make sure 'bamtofastq' is downloaded and added to the PATH ((https://github.com/10XGenomics/bamtofastq)
• navigate to ./analysis/scripts/
• Run $bash gather_refs.sh This script, for human, mouse and human_mouse, uses the .mktranscriptome.sh to build the transcriptome from the genome and the annotation files, and use the mkt2g.sh to generate the gene-to-transcript files. For other references, it uses mkt2g_rest.sh to generate the gene-to-transcript files. Then, the fasta file and the corresponding t2g file for each species are moved to a directory called references/{species}/

Add the relavent paths to the config file:

• The configuration file is the JSON file : analysis/scripts/config_all.json The JSON files for the samples you wish to process should be located at a directory and the path of the directory should be provided in config_all.json. The JSON files should be similar to the ones located at ./analysis/scripts/configs/

Prepare the results with both pipelines (kallisto-bustools and alevin-fry)

• navigate to ./analysis/scripts/
• Run $./make_indices.sh PATH/TO/CONFIG
• Run $./run_all_salmon.sh PATH/TO/CONFIG
• Run $./run_all_kallisto.sh PATH/TO/CONFIG

Generate the time and memory comparison plots

• Set the config file's path to the config_all variable
• Execute all the commands in the memtime notebook: ./analysis/notebooks/memtime.ipynb

Generate the gene count comparison plots

Make sure the results for all the samples generated by both tools are located at ./data/kallisto_out and ./data/salmon_out

• Prepare the data for plotting gene set enrichment analysis
  • Make sure the Seurat version 4.0 and othe required R packages are installed
  • navigate to the ./analysis/notebooks/ directory
  • Run $Rscript run_gsea_bar_full.R
• Load the data for making the plots
  • navigate to the ./analysis/scripts/
  • Run $python mkdata.py -d sample_name -o plotting/output/for/the/sample
• Generate all the plots
  • Run $python mkplot.py -d sample_name -i plotting/output/for/the/sample -o output/dir/