PlantAI-dev/scripts/vcf/07_10_gvcf_to_clean_vcf/README.md

VCF 07-10: gVCF to clean VCF

This directory contains independent extraction scripts for the VCF/resequencing pipeline tail. The original large resequencing script is not modified.

Steps

1. 07_combine_gvcf.sh: collect GVCFs and run GATK CombineGVCFs.
2. 08_genotype_gvcf.sh: run GATK GenotypeGVCFs.
3. 09_select_variants.sh: split genotyped VCF into SNP and INDEL VCFs.
4. 10_hard_filter_clean.sh: apply hard filters and emit GWAS-ready clean.vcf.gz.

Important Boundary

08_genotype_gvcf.sh does not silently fall back to a small interval. If --intervals is supplied, the metrics file marks subset_interval=true.

Test Data

The test runner uses the current single-sample fixture:

/home/xia-lab/wutx_old/wutx/plantAnalysisAgent/vcf_code/vcf/06.gvcf/12.g.vcf.gz

This validates script correctness and chain connectivity only. It is not a population GWAS result.

Run

env PATH=/data/miniconda3/envs/ASAW/bin:/usr/bin:/bin \
  bash scripts/vcf/07_10_gvcf_to_clean_vcf/run_07_10_test.sh

Default output:

/tmp/vcf_07_10_extracted_test