feat: enhance logging for sample intersection to assist debugging "No gene exist." errors

src/jaxqtl/cli.py

@@ -321,12 +321,19 @@ def main(args):
     geno, bim, sample_info = geno_reader(args.geno)
 
     pheno_reader = PheBedReader()
+    # 1. Load raw phenotype data
     pheno = pheno_reader(args.pheno)
+    n_genes_raw = pheno.shape[1] if hasattr(pheno, 'shape') else len(pheno)
+    logging.info(f"Raw phenotype data loaded. Total genes detected: {n_genes_raw}")
 
     covar = covar_reader(args.covar, args.add_covar, args.covar_test, args.rm_covar)
 
     if args.genelist is not None:
+        # Load provided gene list
         genelist = pd.read_csv(args.genelist, header=None, sep="\t").iloc[:, 0].to_list()
+        logging.info(f"User-provided gene list loaded. Total target genes: {len(genelist)}")
+        # Use DEBUG level for verbose data dumping (requires --verbose True)
+        logging.debug(f"Gene list preview (first 5): {genelist[:5]}")
     else:
         genelist = None
 
@@ -336,8 +343,18 @@ def main(args):
     else:
         indList = None
 
+    # 2. Create ReadyData (Intersection of Genotype, Phenotype, and Covariates)
     dat = create_readydata(geno, bim, pheno, covar, autosomal_only=args.autosomal_only, ind_list=indList)
 
+    # Check dimensions after sample intersection
+    n_genes_intersect = dat.pheno.count.shape[1] if hasattr(dat.pheno, 'count') else 0
+    n_samples_intersect = dat.pheno.count.shape[0] if hasattr(dat.pheno, 'count') else 0
+
+    logging.info(f"Data intersection complete. Retained Genes: {n_genes_intersect}, Retained Samples: {n_samples_intersect}")
+
+    if n_genes_intersect == 0 or n_samples_intersect == 0:
+        logging.warning("Data intersection resulted in empty dataset. Check sample ID consistency between phenotype, genotype, and covariates.")
+
     # before filter gene list, calculate library size and set offset, or read in pre-computed log(offset)
     if args.offset is None:
         total_libsize = jnp.array(dat.pheno.count.sum(axis=1))[:, jnp.newaxis]
@@ -347,8 +364,9 @@ def main(args):
         offset_eta = offset_eta.loc[offset_eta.index.isin(dat.pheno.count.index)].sort_index()
         offset_eta = jnp.array(offset_eta)
 
-    # filter out genes with no expressions at all
+    # 3. Filter out non-expressed genes
     dat.filter_gene(geneexpr_percent_cutoff=0.0)
+    logging.info(f"Filtered non-expressed genes (all zeros). Remaining genes: {dat.pheno.count.shape[1]}")
 
     # add expression PCs to covar, genotype PC should appended to covar outside jaxqtl
     if args.addpc > 0:
@@ -359,8 +377,14 @@ def main(args):
         # note: use pre-processed file as in tensorqtl
         offset_eta = jnp.zeros_like(offset_eta)
 
-    # filter gene list
+    # 4. Apply target gene list and expression percentage filter
     dat.filter_gene(gene_list=genelist, geneexpr_percent_cutoff=args.express_percent)
+    logging.info(f"Applied target gene list and expression threshold. Final gene count for mapping: {dat.pheno.count.shape[1]}")
+
+    if dat.pheno.count.shape[1] == 0:
+        logging.error("No genes remaining after filtering. Please check if gene IDs in the provided list match the phenotype file.")
+        # Optionally raise an error here to stop execution cleanly
+        # raise ValueError("No genes available for analysis.")
 
     # permute gene expression for type I error calibration
     if args.perm_pheno: