celescope 1.1.8

GEXSCOPE Single cell analysis

CeleScope

GEXSCOPE Single Cell Analysis Pipelines
Chinese Docs(中文文档): https://gitee.com/singleron-rd/celescope/wikis/

Requirements

• conda
• git
• minimum 32GB RAM(to run STAR aligner)

Installation

1. Clone repo

git clone https://github.com/singleron-RD/CeleScope.git
# If github is blocked
git clone https://gitee.com/singleron-rd/celescope.git

2. Install conda packages

cd CeleScope
conda create -n celescope
conda activate celescope
conda install --file conda_pkgs.txt --channel conda-forge --channel bioconda --channel r --channel imperial-college-research-computing

3. Install celescope

pip install celescope
# Use pypi mirror to accelerate downloading if you are in china
pip install -i https://pypi.tuna.tsinghua.edu.cn/simple celescope

4. Install Beta version(optional)

# If you want to use Beta version of celescope
python setup.py install

Reference genome

Homo sapiens

mkdir -p hs/ensembl_99
cd hs/ensembl_99

wget ftp://ftp.ensembl.org/pub/release-99/fasta/homo_sapiens/dna/Homo_sapiens.GRCh38.dna.primary_assembly.fa.gz
wget ftp://ftp.ensembl.org/pub/release-99/gtf/homo_sapiens/Homo_sapiens.GRCh38.99.gtf.gz

gunzip Homo_sapiens.GRCh38.dna.primary_assembly.fa.gz
gunzip Homo_sapiens.GRCh38.99.gtf.gz

conda activate celescope
gtfToGenePred -genePredExt -geneNameAsName2 Homo_sapiens.GRCh38.99.gtf /dev/stdout | \
    awk '{print $12"\t"$1"\t"$2"\t"$3"\t"$4"\t"$5"\t"$6"\t"$7"\t"$8"\t"$9"\t"$10}' > Homo_sapiens.GRCh38.99.refFlat

STAR \
    --runMode genomeGenerate \
    --runThreadN 6 \
    --genomeDir ./ \
    --genomeFastaFiles Homo_sapiens.GRCh38.dna.primary_assembly.fa \
    --sjdbGTFfile Homo_sapiens.GRCh38.99.gtf \
    --sjdbOverhang 100

Mus musculus

mkdir -p mmu/ensembl_99
cd mmu/ensembl_99

wget ftp://ftp.ensembl.org/pub/release-99/fasta/mus_musculus/dna/Mus_musculus.GRCm38.dna.primary_assembly.fa.gz
wget ftp://ftp.ensembl.org/pub/release-99/gtf/mus_musculus/Mus_musculus.GRCm38.99.gtf.gz

gunzip Mus_musculus.GRCm38.dna.primary_assembly.fa.gz 
gunzip Mus_musculus.GRCm38.99.gtf.gz

conda activate celescope
gtfToGenePred -genePredExt -geneNameAsName2 Mus_musculus.GRCm38.99.gtf /dev/stdout | \
    awk '{print $12"\t"$1"\t"$2"\t"$3"\t"$4"\t"$5"\t"$6"\t"$7"\t"$8"\t"$9"\t"$10}' > Mus_musculus.GRCm38.99.refFlat

STAR \
    --runMode genomeGenerate \
    --runThreadN 6 \
    --genomeDir ./ \
    --genomeFastaFiles Mus_musculus.GRCm38.dna.primary_assembly.fa \
    --sjdbGTFfile Mus_musculus.GRCm38.99.gtf \
    --sjdbOverhang 100

Usage

Single cell RNA-Seq

1. Prepare mapfile

mapfile is a tab-delimited text file(.tsv) containing at least three columns. Each line of mapfile represents a pair of fastq files(Read 1 and Read 2).

First column: Fastq file prefix. Fastq files must be gzipped.

Second column: Fastq directory.

Third column: Sample name, which is the prefix of all generated files. One sample can have multiple fastq files.

Fourth column: Optional, force cell number (scRNA-Seq) or match_dir (scVDJ).

Sample mapfile:

$cat ./my.mapfile
R2007197    /SGRNJ/DATA_PROJ/dir1	sample1
R2007199    /SGRNJ/DATA_PROJ/dir2	sample1
R2007198    /SGRNJ/DATA_PROJ/dir1   sample2

$ls /SGRNJ/DATA_PROJ/dir1
R2007198_L2_2.fq.gz
R2007198_L2_1.fq.gz
R2007197_L2_2.fq.gz
R2007197_L2_1.fq.gz

$ls /SGRNJ/DATA_PROJ/dir2
R2007199_L2_2.fq.gz
R2007199_L2_1.fq.gz

2. Run multi_rna to create shell scripts

conda activate celescope
multi_rna \
 --mapfile ./my.mapfile \
 --genomeDir {some path}/hs/ensembl_99 \
 --thread 8 \
 --mod shell

--mapfile Required, mapfile path.

--genomeDir Required, genomeDir directory.

--thread Maximum number of threads to use, default=4.

--mod Create "sjm"(simple job manager https://github.com/StanfordBioinformatics/SJM) or "shell" scripts.

Shell scripts will be created in ./shell directory, one script per sample. The shell scripts contains all the steps that need to be run.

3. Run shell scripts under current directory

sh ./shell/{sample}.sh

Single Cell VDJ

Running single Cell VDJ is almost the same as running single Cell RNA-Seq, except that the arguments of multi_vdj are somewhat different.

1. Prepare mapfile

If you have paired single cell RNA-seq and VDJ samples, the single cell RNA-Seq directory after running CeleScope is called matched_dir. You can write matched_dir's path as the fourth column of mapfile(optional).

R2007197    /SGRNJ/DATA_PROJ/dir    sample1 /SGRNJ/Projects/sample1

2. Run multi_vdj to create shell scripts

conda activate celescope
multi_vdj \
 --mapfile ./my.mapfile \
 --type TCR \
 --thread 8 \
 --mod shell \

--type Required. TCR or BCR.

3. Run shell scripts under current directory