baseline

Run an L1-regularized logistic regression baseline (with QC filtering and optional covariates) across several regularization strengths, and print train/validation metrics for each. Useful as a fast sanity check before investing in MLP training via attribute or pal.

posthoc baseline

Usage

posthoc baseline [OPTIONS]

Options

--pfile <pfile>: Required Prefix of PLINK2 .pgen/.pvar/.psam fileset.

--pheno <pheno>: Required Phenotype file.

--pheno-name <pheno_name>: Phenotype column name (default: first value column).

--covar <covar>: Covariate file.

--maf <min_maf>

--geno <max_missing>

--indep-pairwise <indep_pairwise>

--val-fraction <val_fraction>

--seed <seed>

Behavior notes

The phenotype is always treated as binary/logistic — there is no --linear mode for baseline.
Missing genotype calls are mean-imputed per-variant using training-split means only, then applied to both train and validation folds.
If --covar is given, covariates are concatenated to the genotype matrix as additional predictor columns.
For each C in (0.001, 0.01, 0.1, 1.0), a sklearn.linear_model.LogisticRegression with l1_ratio=1, solver="liblinear" is fit, and the script prints train loss, validation loss, validation AUC, and the number of non-zero coefficients.
This command does not write an output file — results are printed to stdout only.

Example output

n_train=403 n_val=101 n_features=6112
       C   train_loss   val_loss  val_auc  n_nonzero
   0.001       0.6512     0.6603    0.612          3
    0.01       0.5981     0.6120    0.671         41
     0.1       0.4872     0.5734    0.703        387
     1.0       0.3105     0.6488    0.681       1822