gsMap 1.73.0__py3-none-any.whl → 1.73.2__py3-none-any.whl

gsMap/GNN/train.py

@@ -17,7 +17,7 @@ def reconstruction_loss(decoded, x):
 
 def label_loss(pred_label, true_label):
     """Compute the cross-entropy loss."""
-    return F.cross_entropy(pred_label, true_label)
+    return F.cross_entropy(pred_label, true_label.long())
 
 
 class ModelTrainer:

gsMap/__init__.py

@@ -2,4 +2,4 @@
 Genetics-informed pathogenic spatial mapping
 """
 
-__version__ = "1.73.0"
+__version__ = "1.73.2"

gsMap/config.py

@@ -232,6 +232,9 @@ def add_find_latent_representations_args(parser):
         action="store_true",
         help="Enable hierarchical latent representation finding.",
     )
+    parser.add_argument(
+        "--pearson_residuals", action="store_true", help="Using the pearson residuals."
+    )
 
 
 def chrom_choice(value):
@@ -308,7 +311,7 @@ def add_generate_ldscore_args(parser):
         help="Root path for genotype plink bfiles (.bim, .bed, .fam).",
     )
     parser.add_argument(
-        "--keep_snp_root", type=str, required=True, help="Root path for SNP files."
+        "--keep_snp_root", type=str, required=False, help="Root path for SNP files"
     )
     parser.add_argument(
         "--gtf_annotation_file", type=str, required=True, help="Path to GTF annotation file."
@@ -357,7 +360,11 @@ def add_spatial_ldsc_args(parser):
         "--sumstats_file", type=str, required=True, help="Path to GWAS summary statistics file."
     )
     parser.add_argument(
-        "--w_file", type=str, required=True, help="Path to regression weight file."
+        "--w_file",
+        type=str,
+        required=False,
+        default=None,
+        help="Path to regression weight file. If not provided, will use weights generated in the generate_ldscore step.",
     )
     parser.add_argument(
         "--trait_name", type=str, required=True, help="Name of the trait being analyzed."
@@ -678,6 +685,9 @@ def add_run_all_mode_args(parser):
     parser.add_argument(
         "--gM_slices", type=str, default=None, help="Path to the slice mean file (optional)."
     )
+    parser.add_argument(
+        "--pearson_residuals", action="store_true", help="Using the pearson residuals."
+    )
 
 
 def ensure_path_exists(func):
@@ -854,6 +864,7 @@ class FindLatentRepresentationsConfig(ConfigWithAutoPaths):
     var: bool = False
     convergence_threshold: float = 1e-4
     hierarchically: bool = False
+    pearson_residuals: bool = False
 
     def __post_init__(self):
         # self.output_hdf5_path = self.hdf5_with_latent_path
@@ -942,11 +953,11 @@ class GenerateLDScoreConfig(ConfigWithAutoPaths):
     chrom: int | str
 
     bfile_root: str
-    keep_snp_root: str | None
 
     # annotation by gene distance
     gtf_annotation_file: str
     gene_window_size: int = 50000
+    keep_snp_root: str | None = None
 
     # annotation by enhancer
     enhancer_annotation_file: str = None
@@ -1055,7 +1066,7 @@ class GenerateLDScoreConfig(ConfigWithAutoPaths):
 
 @dataclass
 class SpatialLDSCConfig(ConfigWithAutoPaths):
-    w_file: str
+    w_file: str | None = None
     # ldscore_save_dir: str
     use_additional_baseline_annotation: bool = True
     trait_name: str | None = None
@@ -1105,8 +1116,19 @@ class SpatialLDSCConfig(ConfigWithAutoPaths):
         for sumstats_file in self.sumstats_config_dict.values():
             assert Path(sumstats_file).exists(), f"{sumstats_file} does not exist."
 
-        # check if additional baseline annotation is exist
-        # self.use_additional_baseline_annotation = False
+        # Handle w_file
+        if self.w_file is None:
+            w_ld_dir = Path(self.ldscore_save_dir) / "w_ld"
+            if w_ld_dir.exists():
+                self.w_file = str(w_ld_dir / "weights.")
+                logger.info(f"Using weights generated in the generate_ldscore step: {self.w_file}")
+            else:
+                raise ValueError(
+                    "No w_file provided and no weights found in generate_ldscore output. "
+                    "Either provide --w_file or run generate_ldscore first."
+                )
+        else:
+            logger.info(f"Using provided weights file: {self.w_file}")
 
         if self.use_additional_baseline_annotation:
             self.process_additional_baseline_annotation()
@@ -1117,16 +1139,6 @@ class SpatialLDSCConfig(ConfigWithAutoPaths):
 
         if not dir_exists:
             self.use_additional_baseline_annotation = False
-            # if self.use_additional_baseline_annotation:
-            #     logger.warning(f"additional_baseline directory is not found in {self.ldscore_save_dir}.")
-            #     print('''\
-            #         if you want to use additional baseline annotation,
-            #         please provide additional baseline annotation when calculating ld score.
-            #         ''')
-            #     raise FileNotFoundError(
-            #         f'additional_baseline directory is not found.')
-            # return
-            # self.use_additional_baseline_annotation = self.use_additional_baseline_annotation or True
         else:
             logger.info(
                 "------Additional baseline annotation is provided. It will be used with the default baseline annotation."
@@ -1227,6 +1239,7 @@ class RunAllModeConfig(ConfigWithAutoPaths):
 
     # == Find Latent Representation PARAMETERS ==
     n_comps: int = 300
+    pearson_residuals: bool = False
 
     # == latent 2 Gene PARAMETERS ==
     gM_slices: str | None = None

gsMap/create_slice_mean.py

@@ -23,6 +23,7 @@ def get_common_genes(h5ad_files, config: CreateSliceMeanConfig):
     common_genes = None
     for file in tqdm(h5ad_files, desc="Finding common genes"):
         adata = sc.read_h5ad(file)
+        sc.pp.filter_genes(adata, min_cells=1)
         adata.var_names_make_unique()
         if common_genes is None:
             common_genes = adata.var_names

gsMap/diagnosis.py

@@ -49,7 +49,10 @@ def compute_gene_diagnostic_info(config: DiagnosisConfig):
 
     # Align marker scores with trait LDSC results
     mk_score = mk_score.loc[trait_ldsc_result.index]
-    mk_score = mk_score.loc[:, mk_score.sum(axis=0) != 0]
+
+    # Filter out genes with no variation
+    has_variation = (~mk_score.eq(mk_score.iloc[0], axis=1)).any()
+    mk_score = mk_score.loc[:, has_variation]
 
     logger.info("Calculating correlation between gene marker scores and trait logp-values...")
     corr = mk_score.corrwith(trait_ldsc_result["logp"])
@@ -66,10 +69,6 @@ def compute_gene_diagnostic_info(config: DiagnosisConfig):
         }
     )
 
-    # Filter based on median GSS score
-    high_GSS_Gene_annotation_pair = high_GSS_Gene_annotation_pair[
-        high_GSS_Gene_annotation_pair["Median_GSS"] >= 1.0
-    ]
     high_GSS_Gene_annotation_pair = high_GSS_Gene_annotation_pair.merge(
         corr, left_on="Gene", right_index=True
     )
@@ -88,19 +87,6 @@ def compute_gene_diagnostic_info(config: DiagnosisConfig):
     gene_diagnostic_info.to_csv(gene_diagnostic_info_save_path, index=False)
     logger.info(f"Gene diagnostic information saved to {gene_diagnostic_info_save_path}.")
 
-    # TODO: A new script is needed to save the gene diagnostic info to adata.var and trait_ldsc_result to adata.obs when running multiple traits
-    # # Save to adata.var with the trait_name prefix
-    # logger.info('Saving gene diagnostic info to adata.var...')
-    # gene_diagnostic_info.set_index('Gene', inplace=True)  # Use 'Gene' as the index to align with adata.var
-    # adata.var[f'{config.trait_name}_Annotation'] = gene_diagnostic_info['Annotation']
-    # adata.var[f'{config.trait_name}_Median_GSS'] = gene_diagnostic_info['Median_GSS']
-    # adata.var[f'{config.trait_name}_PCC'] = gene_diagnostic_info['PCC']
-    #
-    # # Save trait_ldsc_result to adata.obs
-    # logger.info(f'Saving trait LDSC results to adata.obs as gsMap_{config.trait_name}_p_value...')
-    # adata.obs[f'gsMap_{config.trait_name}_p_value'] = trait_ldsc_result['p']
-    # adata.write(config.hdf5_with_latent_path, )
-
     return gene_diagnostic_info.reset_index()
 
 
@@ -171,6 +157,20 @@ def generate_manhattan_plot(config: DiagnosisConfig):
         + gwas_data_to_plot["Annotation"].astype(str)
     )
 
+    # Verify data integrity
+    if gwas_data_with_gene_annotation_sort.empty:
+        logger.error("Filtered GWAS data is empty, cannot create Manhattan plot")
+        return
+
+    if len(gwas_data_to_plot) == 0:
+        logger.error("No SNPs passed filtering criteria for Manhattan plot")
+        return
+
+    # Log some diagnostic information
+    logger.info(f"Creating Manhattan plot with {len(gwas_data_to_plot)} SNPs")
+    logger.info(f"Columns available: {list(gwas_data_to_plot.columns)}")
+    logger.info(f"Chromosome column values: {gwas_data_to_plot['CHR'].unique()}")
+
     fig = ManhattanPlot(
         dataframe=gwas_data_to_plot,
         title="gsMap Diagnosis Manhattan Plot",

gsMap/find_latent_representation.py

@@ -50,6 +50,15 @@ def preprocess_data(adata, params):
         # HVGs based on count
         logger.info("Dealing with count data...")
         sc.pp.highly_variable_genes(adata, flavor="seurat_v3", n_top_genes=params.feat_cell)
+
+        # Get the pearson residuals
+        if params.pearson_residuals:
+            sc.experimental.pp.normalize_pearson_residuals(adata, inplace=False)
+            pearson_residuals = sc.experimental.pp.normalize_pearson_residuals(
+                adata, inplace=False, clip=10
+            )
+            adata.layers["pearson_residuals"] = pearson_residuals["X"]
+
         # Normalize the data
         sc.pp.normalize_total(adata, target_sum=1e4)
         sc.pp.log1p(adata)
@@ -64,8 +73,13 @@ class LatentRepresentationFinder:
     def __init__(self, adata, args: FindLatentRepresentationsConfig):
         self.params = args
 
-        self.expression_array = adata[:, adata.var.highly_variable].X.copy()
-        self.expression_array = sc.pp.scale(self.expression_array, max_value=10)
+        if "pearson_residuals" in adata.layers:
+            self.expression_array = (
+                adata[:, adata.var.highly_variable].layers["pearson_residuals"].copy()
+            )
+        else:
+            self.expression_array = adata[:, adata.var.highly_variable].X.copy()
+            self.expression_array = sc.pp.scale(self.expression_array, max_value=10)
 
         # Construct the neighboring graph
         self.graph_dict = construct_adjacency_matrix(adata, self.params)
@@ -103,6 +117,8 @@ def run_find_latent_representation(args: FindLatentRepresentationsConfig):
     # Load the ST data
     logger.info(f"Loading ST data of {args.sample_name}...")
     adata = sc.read_h5ad(args.input_hdf5_path)
+    sc.pp.filter_genes(adata, min_cells=1)
+
     logger.info(f"The ST data contains {adata.shape[0]} cells, {adata.shape[1]} genes.")
 
     # Load the cell type annotation