genomics_QC_pipeline

The genomic QC pipeline is designed to clean and prepare imputed genotype data for pQTL analysis. All the rules were based on Alessia Mapelli’s work: [https://github.com/ht-diva/pqtl-believe-interval/blob/main/Script_QC_INTERVAL_genomics.R]

Requirements

• Singularity

see also environment.yml and Makefile

Getting started

• git clone https://github.com/ht-diva/genomics_QC_pipeline.git
• cd genomics_QC_pipeline
• adapt the submit.sbatch and config/config.yaml files to your environment
• sbatch submit.sbatch

The output is written to the path defined by the workspace_path variable in the config.yaml file. By default, this path is ./results.

Rules description

1. list_rs:
   Purpose: Generate lists of all rsIDs and pseudo biallelic variants from the initial pgen file.
   Output: Two files – one containing all rsIDs and the other containing pseudo biallelic variants.
   

2. recode_pgen:
   Purpose: Replace the IDs in the imputed pgen file with a new format: chr:pos:ref:alt.
   Output: An updated pgen file with the new ID format.
   

3. selected_sample:
   Purpose: Select individuals who are present in both the 2018 data and have corresponding proteomic data.
   Output: A filtered list of individuals.
   

4. filter_var:
   Purpose: Perform several quality control steps: remove additional failed samples, identify and remove heterozygosity outliers, perform minor allele frequency (MAF) filtering, remove related samples based on Hardy-Weinberg equilibrium (HWE).
   Output: A cleaned dataset with high-quality variants and samples.
   

5. create_bgen:
   Purpose: Convert the filtered data from the previous steps into bgen format, a commonly used format for storing large-scale genotype data.
   Output: A bgen file containing the cleaned genotype data.
   

6. qctool:
   Purpose: Compute SNP statistics using qctool, ensuring the quality of the variants.
   Output: SNP statistics file.
   

7. get_hq_variants:
   Purpose: Filter variants to retain only those with an info score greater than 0.7.
   Output: A list of high-quality variants.
   

8. filter_hq_variants:
   Purpose: Extract SNPs with an info score greater than 0.7 from the pgen file and create a new pgen file for each chromosome.
   Output: pgen files for each chromosome containing only high-quality variants.
   

9. merge_filter_hq_variants:
   Purpose: Merge the chromosome-specific pgen files from the previous step into a single pgen file.
   Output: A combined pgen file containing high-quality variants from all chromosomes.
   

10. update_pgen_id:
    Purpose: Update the variant IDs in the pgen file to the format chr:pos:A0:A1, with A0 and A1 in alphabetical order.
    Output: An updated pgen file with harmonised IDs.
    

11. update_pgen_alleles:
    Purpose: Harmonize the alleles in the pgen file to match the new IDs.
    Output: A pgen file with harmonised alleles.
    

12. merge_filter_hq_variants_new_id_alleles_pgen:
    Purpose: Merge all the pgen files from the previous step into a final single pgen file.
    Output: A final combined pgen file with harmonised IDs and alleles, ready for pQTL analysis.
    

13. pgen2bed:
    Purpose: Convert pgen file into bed format. Set hard-call-threshold equal to 0.49999999.
    Output: A bed file with harmonised alleles and minimized missing dosage.
    

14. merge_filter_hq_variants_new_id_alleles_bed:
    Purpose: Merge all the bed files from the previous step into a final single bed file.
    Output: A final combined bed file.
    

Output

1. pgen folder (contains raw pgen files with the new IDs format: chr:pos:ref:alt)
   
   • qc_recoded subfolder (contains pgen files that have been processed through quality control and recoding steps but not yet harmonised.)
   • qc_recoded_harmonised subfolder (contains pgen files that have been both quality controlled, recoded, and harmonised.)
2. bed folder
   • qc_recoded_harmonised subfolder (contains bed files that have been harmonised.)