predict

Generate genomic predictions from fitted SNP effects.

Use this command after fit to score target samples.

Basic Syntax

Minimum Working Command
gelex predict -b target_data -e model.snp.eff -o target.pred.tsv
Full Syntax Template
gelex predict --bfile <genotype_prefix> --snp-eff <snp_effect_file> --out <prediction_file> [OPTIONS]

Required inputs are genotype prefix (--bfile), SNP effects (--snp-eff), and output path (--out).

Options

Quick Start Options

-b, --bfile required

PLINK binary prefix (.bed/.bim/.fam) for target samples.

-e, --snp-eff required

SNP effects file from gelex fit (usually <out>.snp.eff).

-o, --out required

Output file path for prediction results.

Input and Covariate Options

--covar-eff

Optional covariate effect file (usually <out>.param from fit).

--qcovar

Quantitative covariate TSV in format FID IID covar1 ....

--dcovar

Categorical covariate TSV in format FID IID factor1 ....

Processing Options

--iid-only false

Match samples by IID only and ignore FID.

-c, --chunk-size 10000

Number of SNPs per processing chunk. Lower values reduce peak memory.

Output Files

predict writes one prediction table at the exact path passed to --out.

File

Contents

Reference

<out>

Per-sample predictions (FID/IID and predicted values)

Prediction Results (.pred.tsv)

Warnings and Notes

Warning

Use --iid-only only when IID uniquely identifies individuals in every file. Otherwise, samples may be mismatched.

Note

If you use --covar-eff, keep covariate files consistent with the fit stage (same variables, coding, and compatible sample IDs).

Examples

Basic Genomic Prediction
gelex predict \
   -b target_data \
   -e trained_model.snp.eff \
   -o predictions.pred.tsv
Add Covariate Effects
gelex predict \
   -b target_data \
   -e trained_model.snp.eff \
   --covar-eff trained_model.param \
   --qcovar target_age.tsv \
   --dcovar target_sex.tsv \
   -o predictions_with_covar.pred.tsv
IID-Only Matching
gelex predict \
   -b target_data \
   -e trained_model.snp.eff \
   --iid-only \
   -o predictions_iid_only.pred.tsv
Low-Memory Chunking
gelex predict \
   -b target_data \
   -e trained_model.snp.eff \
   --chunk-size 2000 \
   -o predictions_low_mem.pred.tsv

See Also