Command-Line Interface¶
Installing snputils registers a snputils executable for file-backed workflows. The Python API remains the full surface for object manipulation, custom pipelines, and notebook use; use the CLI when inputs and outputs are primarily paths on disk.
snputils --help
snputils --version
Subcommands¶
Command |
Purpose |
|---|---|
|
Standard PCA with optional coordinate/component export and scatter plot |
|
Missing-data PCA with ancestry masking |
|
Multi-array ancestry-specific MDS |
|
Variant-level association testing |
|
Admixture mapping from local ancestry |
|
Simulate admixed haplotype batches (requires |
|
Manhattan plot from association results |
|
Q–Q plot from association results |
Run snputils <command> --help for the full argument list of any subcommand.
Examples¶
PCA — compute two components, save coordinates and a scatter plot:
snputils pca \
--snp-path cohort.pgen \
--n-components 2 \
--coords pca_coords.tsv \
--plot pca.pdf
mdPCA — ancestry-specific embedding with optional plot:
snputils mdpca \
--snp-path cohort.pgen \
--lai-path local_ancestry.msp \
--labels-file labels.tsv \
--ancestry AFR \
--coords mdpca_coords.tsv \
--plot mdpca.pdf
GWAS — association scan from phenotype and genotype files:
snputils gwas \
--phe-id trait \
--phe-path phenotypes.tsv \
--snp-path cohort.pgen \
--covar-path covariates.txt \
--results-path gwas.tsv.gz
Admixture mapping — window-level association from local ancestry:
snputils admixture-map \
--phe-id trait \
--phe-path phenotypes.tsv \
--lai-path local_ancestry.msp \
--covar-path covariates.txt \
--results-path admixmap.tsv.gz
Simulation — generate admixed haplotype batches (PyTorch required):
snputils simulate \
--snp cohort.pgen \
--metadata metadata.tsv \
--genetic-map gmap.tsv \
--output-dir sim_batches/ \
--batch-size 256 \
--n-batches 10 \
--num-generations 10 \
--device cuda:0
Association plots:
snputils plot-manhattan --results-path gwas.tsv.gz --output-path manhattan.png
snputils plot-qq --results-path gwas.tsv.gz --output-path qq.png