reGSusie Nextflow Pipeline

Overview

The reGSusie Nextflow pipeline is designed for conducting Genome-Wide Association Studies (GWAS) using a combination of regenie, LDstore2, and susieR. This pipeline offers a comprehensive and customizable solution for processing genetic association data and fine-mapping causal variants.

Key Features

Conduct association testing with regenie.
Compute in-sample LD matrix with LDstore2.
Perform statistical fine-mapping with susieR.

Prerequisites

Before running the pipeline, ensure you have the following dependencies installed:

Running the Pipeline

Clone the reGSusie GitHub repository and navigate to the directory:
```
git clone https://github.com/idarahu/reGSusie.git
cd reGSusie
```
Customize the nextflow.config file based on your analysis requirements (see Pipeline Configuration below).
To execute the pipeline, use the submit_slurm.sh script. This script is configured to submit a Slurm job for pipeline execution. Here's how to run the pipeline:
```
sbatch submit_slurm.sh
```

Pipeline Configuration

Analysis Parameters

For specific parameters, file types, and other details needed for regenie, LDstore2, and susieR, please refer to their respective documentations.

full_pipeline: If set to false, only the LDstore and SuSiE components of the pipeline are executed.
regenie_outputs: A text file (.txt) where each row corresponds to the regenie output file that should be used as input for LDstore and SuSiE when the previous parameter was set to false.

Parameters for regenie step I

step1_input_format: Select the input format for regenie step 1: bed, pgen, or bgen.
step1_bed: When selecting bed as the input format, it assumes that the input folder contains the required files with the extensions .bed, .bim, and .fam.
step1_pgen: When selecting pgen as the input format, it assumes that the input folder contains the required files with the extensions .pgen, .pvar, and .psam.
step1_bgen: When selecting bgen as the input format, it assumes that the input folder contains the required files with the extensions .bgen, .sample, and .bgi.

Parameters for regenie step I and II

phenotype_file: Path to the phenotype file.
phenotype_list: Path to a text file containing a comma-separated list of phenotypes.
covariate_file: Path to the covariate file.
covariate_list: A comma-separated list of covariates.
bsize: Size of the genotype blocks. The default value is 4000.

Parameters for regenie step II

samples_to_keep: Path to a file specifying samples to keep.
minINFO: Minimum imputation info score. The default value is 0.6.

Parameters for regenie step II and LDStore2

All the following files for regenie step 2 must adhere to the format chr*.bgen/chr*.sample/chr*.bgen.bgi, where * represents the chromosome number.

step2_bgen: Path to BGEN files for regenie step 2.
step2_sample: Path to sample files for regenie step 2.
step2_bgi: Path to BGI files for regenie step 2.

Parameters for region filter

p_value: p-value (default is 5e-08) for defining the genome-wide significant locus.
window_size: Half of the size of the fine-mapping region. The default value is 1500000, which corresponds to a 3 Mb total window around a lead variant.
max_region_width: Maximal fine-mapping region width (default is 6 Mb).
window_shrink_ratio: Shrink ratio (default is 0.95) applied when regions exceed the maximal width (original regions are reduced recursively).
bgen_chr_has_zero: Should be set to T (true) if chromosomes 1..9 are given as 01..09 in bgens.
remove_MHC: If set to T (true), then the histocompatibility complex (MHC) region is excluded from the analysis.
GRCh: Select the Right GRCh (choose either 37 or 38 for the GRCh).
MHC_start and MHC_end : MHC Regions in GRCh37 and GRCh38 (specify the MHC region coordinates).

Parameters for LDStore2

samples_ld_incl: Path to the text file with samples that should be included.

Parameters for susieR

n_covariates: Number of covariates used in the analysis.
max_causal_SNPs: Number of maximum causal variants (default is 10).

Configuring output directory names

outdir: Directory where results are stored (default is 'results').
prefix1: Prefix for regenie results (default is 'regenie').
prefix2: Prefix for fine-mapping results (default is 'finemapping').

Execution configuration

excutor: Process Executor
queue: Process Queue

Additional configuration

Include the base.config file.

Singularity Configuration

Singularity is enabled.
Auto-mounts are enabled.
Cache directory: "$baseDir/singularity_img/"

Helper Functions

check_max: A function to check and validate max memory, time, and CPU parameters.

AlasooLab / reGSusie