yyoshiaki / sclinker-skg

sclinker-skg

Setup

0. Prerequisite

• gsutil
• conda
• scanpy (in python3)
• bedtools
• R (calling /usr/bin/Rscript in the script.)
• tidyverse (in R)

1. Download data

git clone https://github.com/yyoshiaki/sclinker-skg.git
cd sclinker-skg
bash scripts/prep_data.sh ${PWD}

2. create env for ldsc

cd ldsc
conda env create --file environment.yml
conda create --name scanpy1.8.1 python=3.8 scanpy=1.8.1

sclinker-celltype

This includes scanpy's scanpy.tl.rank_genes_group.

Usage

bash ./scripts/sclinker-workflow.celltype.skg.sh ./configs/pbmc_all.sh

Example (PBMC)

mkdir -p example
cd example
# download scanpy object
wget https://storage.googleapis.com/broad-alkesgroup-public/LDSCORE/Jagadeesh_Dey_sclinker/scdata/pbmc-processed-annotated.h5ad
cd ..

bash ./scripts/sclinker-workflow.celltype.skg.sh ./configs/pbmc_all.sh

input

BASEDIR=~/yyoshiaki-git/sclinker-skg

NAME=exmaple.pbmc.all
CELLTYPE_COL=celltype
WORKDIR=$BASEDIR/$NAME
ADATA=$BASEDIR/example/pbmc-processed-annotated.h5ad
# 0 for all genes, >=1 to select top n gene for each celltype
N_TOP_GENE=0
TISSUE=BLD # BLD, BRN
CSV_SUMSTATS=${BASEDIR}/data/ldsc/sumstats.csv

sclinker-genescores

This pipline starts from a genescore's CSV file.

bash ./scripts/sclinker-workflow.genescores.skg.sh ./configs/pbmc_all.genescores.sh

Example (PBMC, after execution of celltype mode)

bash ./scripts/sclinker-workflow.genescores.skg.sh ./configs/pbmc_all.genescores.sh

input

BASEDIR=~/yyoshiaki-git/sclinker-skg

NAME=exmaple.pbmc.genescores.all
WORKDIR=$BASEDIR/$NAME
GENE_SCORES=$BASEDIR/example.pbmc.all/celltype_genescores.csv
# 0 for all genes, >=1 to select top n gene for each celltype
N_TOP_GENE=0
TISSUE=BLD # BLD, BRN
CSV_SUMSTATS=${BASEDIR}/data/ldsc/sumstats.csv