hlnicholls/slalom-susie

SLALOM dentist outlier detection + SuSiE fine-mapping pipeline

Tested with python v3.10.8

• Python example version of complete pipeline runs in slalom_susie_pipeline.py
  • slalom_susie_pipeline.py runs with any GWAS catalog study locus and using the UKBiobank LD reference (using PLINK to get LD matrix).
• PySpark work in progress for ETL pipeline in /etl-pipeline/pyspark_finemapping_pipeline.py
  • Functions developed and tested individually in slalom_susie_pipeline.ipynb and pyspark_slalom_susie_pipeline.ipynb
• /old_pipeline scripts 1-4 can also run the same pipeline (steps originally coded in R).
• /susie-r-pipeline scripts for running fine-mapping pipelines in R in parallelised batches.

Installation

Environment:

conda env create -f finemapping_env.yml
conda activate finemaping_env

or:

pip install -r requirements.txt

Additional software:

brew install libomp
export LDFLAGS="-L/opt/homebrew/opt/libomp/lib"
export CPPFLAGS="-I/opt/homebrew/opt/libomp/include"
pip install --no-cache-dir --no-binary :all: bgzip

# clone and install susieinf finemap from: https://github.com/FinucaneLab/fine-mapping-inf/tree/master
cd susieinf
python setup.py bdist_wheel
pip install .

cd finemapinf
python setup.py bdist_wheel
pip install .

Example Requirements

slalom_susie_pipeline.py example runs with any GWAS catalog study locus and using the UKBiobank LD reference (using PLINK to get LD matrix).

Example input required for slalom_susie_pipeline.py to run:

# Example inputs to set at line 255 of slalom_susie_pipeline.py
gwas_file_path = "/33589840-GCST90012877-EFO_0000249.h.tsv.gz"
target="PTK2B",
target_chrom=8,
target_pos=27610986,
lead_snp_ID="8:27610986:C:A",
n_sample=472868,
outlier_method='DENTIST',
r2_threshold = 0.6,
nlog10p_dentist_s_threshold = 1e-4