Cumulus Tool on Feature Barcoding

A fast C++ tool to extract feature-count matrix from sequence reads in FASTQ files. We uses isal-l for decompressing and Heng Li's kseq library for read parsing. It is used by Cumulus for feature-count matrix generation of cell hashing, nucleus hashing, CITE-Seq and Perturb-seq protocols, using either 10x Genomics V2, V3 or V4 chemistry.

Installation

The installation has been tested on Debian and Ubuntu Linux.

1. Install dependency packages:

sudo apt install build-essential git libisal2 libisal-dev libdeflate0 libdeflate-dev

Important: Make sure to install libisal2 and libisal-dev version 2.30.0 or later.

2. Check out this repository via Git:

git clone https://github.com/lilab-bcb/cumulus_feature_barcoding.git

3. Enter the directory and compile:

cd cumulus_feature_barcoding
make all

4. Now you'll have an executable named generate_count_matrix_ADTs inside your folder. Type

./generate_count_matrix_ADTs

to see its usage.

5. The 10x barcode inclusion list files are needed. Below is a list of them which you can find in Cell Ranger v9.0.1 source code:

• cellranger-9.0.1/lib/python/cellranger/barcodes/:
  • 3M-3pgex-may-2023_TRU.txt.gz
  • 3M-5pgex-jan-2023.txt.gz
  • 3M-february-2018_TRU.txt.gz
  • 737K-arc-v1.txt.gz
  • 737K-august-2016.txt
• cellranger-9.0.1/lib/python/cellranger/barcodes/translation/:
  • 3M-3pgex-may-2023_NXT.txt.gz
  • 3M-february-2018_NXT.txt.gz

Just put these 7 files in the same folder, and specify it as cell_barcodes_dir argument for generate_count_matrix_ADTs command.

Compile on Mac OS

1. Homebrew installation

brew install isa-l
brew install libdeflate

2. Compile

cd cumulus_feature_barcoding
make all

Usage