gsMap 1.60__py3-none-any.whl

gsMap/config.py

@@ -0,0 +1,734 @@
+import argparse
+import logging
+from dataclasses import dataclass, field
+from pprint import pprint
+from typing import Union, Literal
+from pathlib import Path
+
+from collections import OrderedDict, namedtuple
+from typing import Callable
+from gsMap.__init__ import __version__
+import pyfiglet
+
+# Global registry to hold functions
+cli_function_registry = OrderedDict()
+subcommand = namedtuple('subcommand', ['name', 'func', 'add_args_function', 'description'])
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.DEBUG)
+handler = logging.StreamHandler()
+handler.setFormatter(logging.Formatter(
+    '[{asctime}] {levelname:6s} {message}', style='{'))
+logger.addHandler(handler)
+
+
+# Decorator to register functions for cli parsing
+def register_cli(name: str, description: str, add_args_function: Callable) -> Callable:
+    def decorator(func: Callable) -> Callable:
+        def wrapper(*args, **kwargs):
+            name.replace('_', ' ')
+            gsMap_main_logo = pyfiglet.figlet_format("gsMap", font='doom', width=80, justify='center', ).rstrip()
+            print(gsMap_main_logo, )
+            version_number = 'Version: ' + __version__
+            print(version_number.center(80))
+            print('=' * 80)
+            logger.info(f"Running {name}...")
+            func(*args, **kwargs)
+            logger.info(f"Finished running {name}.")
+
+        cli_function_registry[name] = subcommand(name=name, func=wrapper, add_args_function=add_args_function,
+                                                 description=description)
+        return wrapper
+
+    return decorator
+
+
+def add_find_latent_representations_args(parser):
+    parser.add_argument('--input_hdf5_path', required=True, type=str, help='Path to the input hdf5 file.')
+    parser.add_argument('--output_hdf5_path', required=True, type=str, help='Path to the output hdf5 file.')
+    parser.add_argument('--sample_name', required=True, type=str, help='Name of the sample.')
+    parser.add_argument('--annotation', default=None, type=str, help='Name of the annotation layer.')
+    parser.add_argument('--type', default=None, type=str, help="The gene expression layer of the input data (e.g., 'count', 'counts', 'log1p').")
+    parser.add_argument('--epochs', default=300, type=int,
+                        help="Number of training epochs for the GNN-VAE model. Default is 300.")
+
+    parser.add_argument('--feat_hidden1', default=256, type=int,
+                        help="Number of neurons in the first hidden layer of the feature extraction network. Default is 256.")
+    parser.add_argument('--feat_hidden2', default=128, type=int,
+                        help="Number of neurons in the second hidden layer of the feature extraction network. Default is 128.")
+    parser.add_argument('--feat_cell', default=3000, type=int,
+                        help="Number of top variable genes to select. Default is 3000.")
+    parser.add_argument('--gcn_hidden1', default=64, type=int,
+                        help="Number of units in the first hidden layer of the GCN. Default is 64.")
+    parser.add_argument('--gcn_hidden2', default=30, type=int,
+                        help="Number of units in the second hidden layer of the GCN. Default is 30.")
+    parser.add_argument('--p_drop', default=0.1, type=float,
+                        help="Dropout rate used in the GNN-VAE model. Default is 0.1.")
+    parser.add_argument('--gcn_lr', default=0.001, type=float,
+                        help="Learning rate for the GCN network. Default is 0.001.")
+    parser.add_argument('--gcn_decay', default=0.01, type=float,
+                        help="Weight decay (L2 penalty) for the GCN network. Default is 0.01.")
+    parser.add_argument('--n_neighbors', default=11, type=int,
+                        help="Number of neighbors to consider for graph construction in GCN. Default is 11.")
+    parser.add_argument('--label_w', default=1, type=float,
+                        help="Weight of the label loss in the loss function. Default is 1.")
+    parser.add_argument('--rec_w', default=1, type=float,
+                        help="Weight of the reconstruction loss in the loss function. Default is 1.")
+    parser.add_argument('--n_comps', default=300, type=int,
+                        help="Number of principal components to keep if PCA is performed. Default is 300.")
+    parser.add_argument('--weighted_adj', action='store_true',
+                        help="Use a weighted adjacency matrix in GCN. Default is False.")
+    parser.add_argument('--nheads', default=3, type=int,
+                        help="Number of heads in the attention mechanism of the GNN. Default is 3.")
+    parser.add_argument('--var', action='store_true',
+                        help="Enable var. Use --var to enable. Default is False.")
+    parser.add_argument('--convergence_threshold', default=1e-4, type=float,
+                        help="Threshold for convergence during training. Training stops if the loss change is below this threshold. Default is 1e-4.")
+    parser.add_argument('--hierarchically', action='store_true',
+                        help="Find latent representations hierarchically. Use --hierarchically to enable. Default is False.")
+
+
+def chrom_choice(value):
+    if value.isdigit():
+        ivalue = int(value)
+        if 1 <= ivalue <= 22:
+            return ivalue
+    elif value.lower() == 'all':
+        return value
+    else:
+        raise argparse.ArgumentTypeError(f"'{value}' is an invalid chromosome choice. Choose from 1-22 or 'all'.")
+
+
+def filter_args_for_dataclass(args_dict, data_class: dataclass):
+    return {k: v for k, v in args_dict.items() if k in data_class.__dataclass_fields__}
+
+
+def get_dataclass_from_parser(args: argparse.Namespace, data_class: dataclass):
+    remain_kwargs = filter_args_for_dataclass(vars(args), data_class)
+    print(f'Using the following arguments for {data_class.__name__}:')
+    pprint(remain_kwargs)
+    return data_class(**remain_kwargs)
+
+
+def add_generate_ldscore_args(parser):
+    parser.add_argument('--sample_name', type=str, required=True, help='Sample name')
+    parser.add_argument('--chrom', type=str, required=True, help='Chromosome number (1-22) or "all"')
+    parser.add_argument('--ldscore_save_dir', type=str, required=True, help='Directory to save ld score files')
+    parser.add_argument('--mkscore_feather_file', type=str, required=True, help='Mkscore feather file path')
+
+    # additional baseline annotation
+    parser.add_argument('--additional_baseline_annotation_dir_path', type=str, default=None,)
+
+    # reference panel
+    parser.add_argument('--bfile_root', type=str, required=True, help='Bfile root path')
+    parser.add_argument('--keep_snp_root', type=str, required=True, help='Keep SNP root path')
+
+    # Annotation by gene distance
+    parser.add_argument('--gtf_annotation_file', type=str, required=True, help='GTF file path')
+    parser.add_argument('--gene_window_size', type=int, default=50000, help='Gene window size')
+
+    # Enhancer annotation
+    parser.add_argument('--enhancer_annotation_file', type=str, default=None,
+                        help='Enhancer annotation bed file path, optional.')
+    parser.add_argument('--snp_multiple_enhancer_strategy', type=str, default='max_mkscore',
+                        choices=['max_mkscore', 'nearest_TSS'], help='Strategy for multiple enhancers per SNP')
+    parser.add_argument('--gene_window_enhancer_priority', type=str, default=None,
+                        choices=['gene_window_first', 'enhancer_first', 'enhancer_only'],
+                        help='Priority between gene window and enhancer')
+
+    # Arguments for calculating ld score
+    parser.add_argument('--spots_per_chunk', type=int, default=5_000, help='Number of spots per chunk')
+    parser.add_argument('--ld_wind', type=int, default=1, help='LD window size')
+    parser.add_argument('--ld_unit', type=str, default='CM', help='LD window unit (SNP/KB/CM)',
+                        choices=['SNP', 'KB', 'CM'])
+
+
+def add_latent_to_gene_args(parser):
+    parser.add_argument('--input_hdf5_with_latent_path', type=str, required=True,
+                        help='Path to the input HDF5 file which contains latent representations.')
+    parser.add_argument('--sample_name', type=str, required=True, help='Name of the sample.')
+    parser.add_argument('--output_feather_path', type=str, required=True,
+                        help='Path to save output gene marker score feather file.')
+    parser.add_argument('--annotation', default=None, type=str, help='Name of the annotation layer.')
+    parser.add_argument('--type', default=None, type=str, help="Type of input data (e.g., 'count', 'counts').")
+
+    # no_expression_fraction
+    parser.add_argument('--no_expression_fraction', action='store_true', default=False,
+                        help='Flag to not use expression fraction as filter when calculate the maker score. Default is False.')
+
+    parser.add_argument('--latent_representation', type=str, default='latent_GVAE',
+                        choices=['latent_GVAE', 'latent_PCA'],
+                        help='Type of latent representation. Default is "latent_GVAE".')
+    parser.add_argument('--num_neighbour', type=int, default=21,
+                        help='Number of neighbours to consider. Default is 21.')
+    parser.add_argument('--num_neighbour_spatial', type=int, default=101,
+                        help='Number of spatial neighbours to consider. Default is 101.')
+    parser.add_argument('--species', type=str, default=None, help='Species name, which is the column name in the homolog gene conversion file.')
+    parser.add_argument('--gs_species', type=str, default=None, help='Homologous gene conversion file path, if applicable.')
+    parser.add_argument('--gM_slices', type=str, default=None, help='The mean Gene marker scores path across multiple slices.')
+
+
+def add_spatial_ldsc_args(parser):
+    # Group for GWAS input data
+    parser.add_argument('--sample_name', required=True, help="Name of the spatial transcriptomic dataset.")
+
+    parser.add_argument('--sumstats_file', default=None, help="Path to GWAS summary statistics file.")
+    parser.add_argument('--sumstats_config_file', default=None, help="Path to GWAS summary statistics config file.")
+    parser.add_argument('--w_file', required=True, help="Path to regression weight file.")
+    parser.add_argument('--ldscore_input_dir', required=True, help="Input directory for LD Score files.")
+    parser.add_argument('--ldsc_save_dir', required=True, help="Directory to save Spatial LDSC results.")
+    parser.add_argument('--trait_name', default=None, help="Name of the trait.")
+    parser.add_argument('--not_M_5_50', action='store_true', help="Flag to not use M 5 50 in calculations.")
+    parser.add_argument('--n_blocks', type=int, default=200, help="Number of blocks for jackknife resampling.")
+    parser.add_argument('--chisq_max', type=int, help="Maximum chi-square value for filtering SNPs.")
+    parser.add_argument('--all_chunk', type=int, help="Number of chunks for processing spatial data.")
+
+    # if use additional baseline annotation
+    parser.add_argument('--disable_additional_baseline_annotation', action='store_true', default=False,)
+
+    parser.add_argument('--num_processes', type=int, default=4, help="Number of processes for parallel computing.")
+
+    return parser
+
+
+def add_Cauchy_combination_args(parser):
+    # Required arguments
+    parser.add_argument('--input_hdf5_path', required=True, type=str, help='Path to the HDF5 file')
+    parser.add_argument('--input_ldsc_dir', required=True, type=str, help='Directory containing LDSC results')
+    parser.add_argument('--output_cauchy_dir', required=True, type=str,
+                        help='Output directory for Cauchy combination results')
+    parser.add_argument('--sample_name', required=True, type=str, help='Name of the sample')
+    parser.add_argument('--trait_name', required=True, type=str, help='Name of the trait')
+    parser.add_argument('--annotation', required=True, type=str, help='Annotation layer name')
+
+    # Optional arguments
+    parser.add_argument('--meta', default=None, type=str, )
+    parser.add_argument('--slide', default=None, type=str, )
+
+
+
+def add_Visualization_args(parser):
+    # Required arguments
+    parser.add_argument('--input_hdf5_path', required=True, type=str, help='Path to the HDF5 file')
+    parser.add_argument('--input_ldsc_dir', required=True, type=str, help='Directory containing LDSC results')
+    parser.add_argument('--output_figure_dir', required=True, type=str, help='Output directory for figures')
+    parser.add_argument('--sample_name', required=True, type=str, help='Name of the sample')
+    parser.add_argument('--trait_name', required=True, type=str, help='Name of the trait')
+    parser.add_argument('--annotation', default=None, type=str, help='Name of the annotation layer.')
+
+
+    # Arguments with defaults
+    parser.add_argument('--fig_title', type=str, default=None, help='Title of figure')
+    parser.add_argument('--fig_height', type=int, default=600, help='Height of figure in pixels')
+    parser.add_argument('--fig_width', type=int, default=800, help='Width of figure in pixels')
+    parser.add_argument('--point_size', type=int, default=None, help='Point size of figure')
+    parser.add_argument('--fig_style', type=str, default='light', choices=['dark', 'light'], help='Plot style of figure')
+
+def add_all_mode_args(parser):
+    parser.add_argument('--input_hdf5_path', required=True, type=str, help='Path to the input hdf5 file.')
+    parser.add_argument('--save_dir', required=True, type=str, help='Path to the running results.')
+    # output
+    # parser.add_argument('--output_hdf5_path', required=True, type=str, help='Path to the output hdf5 file.')
+    parser.add_argument('--sample_name', required=True, type=str, help='Name of the sample.')
+    parser.add_argument('--annotation', default=None, type=str, help='Name of the annotation layer.')
+    parser.add_argument('--type', default=None, type=str, help="The gene expression layer of the input data (e.g., 'count', 'counts', 'log1p').")
+
+    # latent_to_gene
+    # input
+    # parser.add_argument('--input_hdf5_path', type=str, required=True, help='Path to the input HDF5 file.')
+    # parser.add_argument('--sample_name', type=str, required=True, help='Name of the sample.')
+    # output
+    # parser.add_argument('--output_feather_path', type=str, required=True,
+    #                     help='Path to save output gene marker score feather file.')
+    # parser.add_argument('--annotation', default=None, type=str, help='Name of the annotation layer.')
+    # parser.add_argument('--type', default=None, type=str, help="Type of input data (e.g., 'count', 'counts').")
+
+    # no_expression_fraction
+    # no_expression_fraction
+    parser.add_argument('--no_expression_fraction', action='store_true', default=False,
+                        help='Flag to not use expression fraction as filter when calculate the maker score. Default is False.')
+
+    parser.add_argument('--latent_representation', type=str, default='latent_GVAE',
+                        choices=['latent_GVAE', 'latent_PCA'],
+                        help='Type of latent representation. Default is "latent_GVAE".')
+    parser.add_argument('--num_neighbour', type=int, default=21,
+                        help='Number of neighbours to consider. Default is 21.')
+    parser.add_argument('--num_neighbour_spatial', type=int, default=101,
+                        help='Number of spatial neighbours to consider. Default is 101.')
+    parser.add_argument('--species', type=str, default=None, help='Species name, which is the column name in the homolog gene conversion file.')
+    parser.add_argument('--gs_species', type=str, default=None, help='Homologous gene conversion file path, if applicable.')
+    parser.add_argument('--gM_slices', type=str, default=None, help='The mean Gene marker scores path across multiple slices.')
+
+
+    # generate_ldscore
+    # parser.add_argument('--sample_name', type=str, required=True, help='Sample name')
+    # should be all
+    # parser.add_argument('--chrom', type=chrom_choice, required=True, help='Chromosome number (1-22) or "all"')
+    # output
+    # parser.add_argument('--ldscore_save_dir', type=str, required=True, help='Directory to save ld score files')
+
+    # reference panel
+    parser.add_argument('--bfile_root', type=str, required=True, help='Bfile root path')
+    parser.add_argument('--keep_snp_root', type=str, required=True, help='Keep SNP root path')
+
+    # Annotation by gene distance
+    parser.add_argument('--gtf_annotation_file', type=str, required=True, help='GTF file path')
+    parser.add_argument('--gene_window_size', type=int, default=50000, help='Gene window size')
+
+    # Enhancer annotation
+    parser.add_argument('--enhancer_annotation_file', type=str, default=None,
+                        help='Enhancer annotation bed file path, optional.')
+    parser.add_argument('--snp_multiple_enhancer_strategy', type=str, default='max_mkscore',
+                        choices=['max_mkscore', 'nearest_TSS'], help='Strategy for multiple enhancers per SNP')
+    parser.add_argument('--gene_window_enhancer_priority', type=str, default=None,
+                        choices=['gene_window_first', 'enhancer_first', 'enhancer_only'],
+                        help='Priority between gene window and enhancer')
+
+    # Arguments for calculating ld score
+    parser.add_argument('--spots_per_chunk', type=int, default=5_000, help='Number of spots per chunk')
+    parser.add_argument('--ld_wind', type=int, default=1, help='LD window size')
+    parser.add_argument('--ld_unit', type=str, default='CM', help='LD window unit (SNP/KB/CM)',
+                        choices=['SNP', 'KB', 'CM'])
+
+    # spatial ldsc args:
+    parser.add_argument('--sumstats_file', default=None, help="Path to GWAS summary statistics file.")
+    parser.add_argument('--sumstats_config_file', default=None, help="Path to GWAS summary statistics config file.")
+    parser.add_argument('--w_file', required=True, help="Path to regression weight file.")
+    parser.add_argument('--ldscore_input_dir', required=True, help="Input directory for LD Score files.")
+    parser.add_argument('--ldsc_save_dir', required=True, help="Directory to save Spatial LDSC results.")
+    parser.add_argument('--trait_name', default=None, help="Name of the trait.")
+    parser.add_argument('--not_M_5_50', action='store_true', help="Flag to not use M 5 50 in calculations.")
+    parser.add_argument('--n_blocks', type=int, default=200, help="Number of blocks for jackknife resampling.")
+    parser.add_argument('--chisq_max', type=int, help="Maximum chi-square value for filtering SNPs.")
+    parser.add_argument('--all_chunk', type=int, help="Number of chunks for processing spatial data.")
+
+
+def get_runall_mode_config(args: argparse.Namespace):
+    # output
+    args.output_hdf5_path = f'{args.save_dir}/{args.sample_name}/find_latent_representations/{args.sample_name}_add_latent.h5ad'
+    args.output_feather_path = f'{args.save_dir}/{args.sample_name}/latent_to_gene/{args.sample_name}_gene_marker_score.feather'
+    args.ldscore_save_dir = f'{args.save_dir}/{args.sample_name}/generate_ldscore'
+    args.ldsc_save_dir = f'{args.save_dir}/{args.sample_name}/spatial_ldsc'
+    args.output_cauchy_dir = f'{args.save_dir}/{args.sample_name}/cauchy_combination/'
+
+    # input
+    args.input_hdf5_with_latent_path = args.output_hdf5_path
+    args.mkscore_feather_file = args.output_feather_path
+    args.ldscore_input_dir = args.ldscore_save_dir
+    args.chrom = 'all'
+    args.input_ldsc_dir = args.ldsc_save_dir
+    args.input_spatial_ldsc = f'{args.save_dir}/{args.sample_name}/spatial_ldsc/{args.sample_name}_{args.trait_name}.gz'
+    # find_latent_representations
+    flr_config = get_dataclass_from_parser(args, FindLatentRepresentationsConfig)
+    # latent_to_gene
+    ltg_config = get_dataclass_from_parser(args, LatentToGeneConfig)
+    # generate_ldscore
+    gls_config = get_dataclass_from_parser(args, GenerateLDScoreConfig)
+    # spatial ldsc
+    ldsc_config = get_dataclass_from_parser(args, SpatialLDSCConfig)
+    # cauchy combination
+    cauchy_config = get_dataclass_from_parser(args, CauchyCombinationConfig)
+    return RunAllModeConfig(flr_config=flr_config, ltg_config=ltg_config, gls_config=gls_config,
+                            ldsc_config=ldsc_config, cauchy_config=cauchy_config)
+
+
+def add_format_sumstats_args(parser):
+    # Required arguments
+    parser.add_argument('--sumstats', required=True, type=str, 
+                        help='Path to gwas summary data')
+    parser.add_argument('--out', required=True, type=str, 
+                        help='Path to save the formatted gwas data')
+    
+    # Arguments for specify column name
+    parser.add_argument('--snp', default=None,type=str, 
+                        help="Name of snp column (if not a name that gsMap understands)")
+    parser.add_argument('--a1', default=None,type=str,
+                        help="Name of effect allele column (if not a name that gsMap understands)")
+    parser.add_argument('--a2', default=None,type=str,
+                        help="Name of none-effect allele column (if not a name that gsMap understands)")
+    parser.add_argument('--info', default=None,type=str,
+                        help="Name of info column (if not a name that gsMap understands)")
+    parser.add_argument('--beta', default=None,type=str,
+                        help="Name of gwas beta column (if not a name that gsMap understands).")
+    parser.add_argument('--se', default=None,type=str,
+                        help="Name of gwas standar error of beta column (if not a name that gsMap understands)")
+    parser.add_argument('--p', default=None,type=str,
+                        help="Name of p-value column (if not a name that gsMap understands)")
+    parser.add_argument('--frq', default=None,type=str,
+                        help="Name of A1 ferquency column (if not a name that gsMap understands)")
+    parser.add_argument('--n', default=None,type=str,
+                        help="Name of sample size column (if not a name that gsMap understands)")
+    parser.add_argument('--z', default=None,type=str,
+                        help="Name of gwas Z-statistics column (if not a name that gsMap understands)")
+    parser.add_argument('--OR', default=None,type=str,
+                        help="Name of gwas OR column (if not a name that gsMap understands)")
+    parser.add_argument('--se_OR', default=None,type=str,
+                        help="Name of standar error of OR column (if not a name that gsMap understands)")
+    
+    # Arguments for convert SNP (chr, pos) to rsid
+    parser.add_argument('--chr', default="Chr",type=str,
+                        help="Name of SNP chromosome column (if not a name that gsMap understands)")
+    parser.add_argument('--pos', default="Pos",type=str,
+                        help="Name of SNP positions column (if not a name that gsMap understands)")
+    parser.add_argument('--dbsnp', default=None,type=str,
+                        help='Path to reference dnsnp file')
+    parser.add_argument('--chunksize', default=1e+6,type=int,
+                        help='Chunk size for loading dbsnp file')
+    
+    # Arguments for output format and quality
+    parser.add_argument('--format',default='gsMap', type=str, 
+                        help='Format of output data',choices=['gsMap', 'COJO'])
+    parser.add_argument('--info_min', default=0.9,type=float,
+                        help='Minimum INFO score.')
+    parser.add_argument('--maf_min', default=0.01,type=float,
+                        help='Minimum MAF.')
+    parser.add_argument('--keep_chr_pos', action='store_true', default=False,
+                    help='Keep SNP chromosome and position columns in the output data')
+
+@dataclass
+class FindLatentRepresentationsConfig:
+    input_hdf5_path: str
+    output_hdf5_path: str
+    sample_name: str
+    annotation: str = None
+    type: str = None
+
+    epochs: int = 300
+    feat_hidden1: int = 256
+    feat_hidden2: int = 128
+    feat_cell: int = 3000
+    gcn_hidden1: int = 64
+    gcn_hidden2: int = 30
+    p_drop: float = 0.1
+    gcn_lr: float = 0.001
+    gcn_decay: float = 0.01
+    n_neighbors: int = 11
+    label_w: float = 1
+    rec_w: float = 1
+    input_pca: bool = True
+    n_comps: int = 300
+    weighted_adj: bool = False
+    nheads: int = 3
+    var: bool = False
+    convergence_threshold: float = 1e-4
+    hierarchically: bool = False
+
+    def __post_init__(self):
+        if self.hierarchically:
+            if self.annotation is None:
+                raise ValueError('annotation must be provided if hierarchically is True.')
+            logger.info(
+                f'------Hierarchical mode is enabled. This will find the latent representations within each annotation.')
+
+        # remind for not providing annotation
+        if self.annotation is None:
+            logger.warning(
+                'annotation is not provided. This will find the latent representations for the whole dataset.')
+        else:
+            logger.info(f'------Find latent representations for {self.annotation}...')
+
+
+@dataclass
+class LatentToGeneConfig:
+    input_hdf5_with_latent_path: str
+    sample_name: str
+    output_feather_path: str
+    no_expression_fraction: bool = False
+    latent_representation: str = 'latent_GVAE'
+    num_neighbour: int = 21
+    num_neighbour_spatial: int = 101
+    species: str = None
+    gs_species: str = None
+    gM_slices: str = None
+    annotation: str = None
+    type: str = None
+
+
+@dataclass
+class GenerateLDScoreConfig:
+    sample_name: str
+    chrom: Union[int, str]
+    ldscore_save_dir: str
+    mkscore_feather_file: str
+    bfile_root: str
+    keep_snp_root: str
+
+    # annotation by gene distance
+    gtf_annotation_file: str
+    gene_window_size: int = 50000
+
+    # annotation by enhancer
+    enhancer_annotation_file: str = None
+    snp_multiple_enhancer_strategy: Literal['max_mkscore', 'nearest_TSS'] = 'max_mkscore'
+    gene_window_enhancer_priority: Literal['gene_window_first', 'enhancer_first', 'enhancer_only',] = None
+
+    # for calculating ld score
+    additional_baseline_annotation_dir_path: str = None
+    spots_per_chunk: int = 5_000
+    ld_wind: int = 1
+    ld_unit: str = 'CM'
+
+    def __post_init__(self):
+        if self.enhancer_annotation_file is not None and self.gene_window_enhancer_priority is None:
+            logger.warning("enhancer_annotation_file is provided but gene_window_enhancer_priority is not provided. "
+                           "by default, gene_window_enhancer_priority is set to 'enhancer_only', when enhancer_annotation_file is provided.")
+            self.gene_window_enhancer_priority = 'enhancer_only'
+        if self.enhancer_annotation_file is None and self.gene_window_enhancer_priority is not None:
+            logger.warning("gene_window_enhancer_priority is provided but enhancer_annotation_file is not provided. "
+                           "by default, gene_window_enhancer_priority is set to None, when enhancer_annotation_file is not provided.")
+            self.gene_window_enhancer_priority = None
+        assert self.gene_window_enhancer_priority in [None, 'gene_window_first', 'enhancer_first', 'enhancer_only', ], \
+            f"gene_window_enhancer_priority must be one of None, 'gene_window_first', 'enhancer_first', 'enhancer_only', but got {self.gene_window_enhancer_priority}."
+        if self.gene_window_enhancer_priority in ['gene_window_first', 'enhancer_first']:
+            logger.info(f'Both gene_window and enhancer annotation will be used to calculate LD score. ')
+            logger.info(
+                f'SNP within +-{self.gene_window_size} bp of gene body will be used and enhancer annotation will be used to calculate LD score. If a snp maps to multiple enhancers, the strategy to choose by your select strategy: {self.snp_multiple_enhancer_strategy}.')
+        elif self.gene_window_enhancer_priority == 'enhancer_only':
+            logger.info(f'Only enhancer annotation will be used to calculate LD score. ')
+        else:
+            logger.info(
+                f'Only gene window annotation will be used to calculate LD score. SNP within +-{self.gene_window_size} bp of gene body will be used. ')
+
+        # remind for baseline annotation
+        if self.additional_baseline_annotation_dir_path is None:
+            logger.info(f'------Baseline annotation is not provided. Default baseline annotation will be used.')
+        else:
+            logger.info(f'------Baseline annotation is provided. Additional baseline annotation will be used with the default baseline annotation.')
+            logger.info(f'------Baseline annotation directory: {self.additional_baseline_annotation_dir_path}')
+            # check the existence of baseline annotation
+            if self.chrom == 'all':
+                for chrom in range(1, 23):
+                    chrom = str(chrom)
+                    baseline_annotation_path = Path(self.additional_baseline_annotation_dir_path) / f'baseline.{chrom}.annot.gz'
+                    if not baseline_annotation_path.exists():
+                        raise FileNotFoundError(f'baseline.{chrom}.annot.gz is not found in {self.additional_baseline_annotation_dir_path}.')
+            else:
+                baseline_annotation_path = Path(self.additional_baseline_annotation_dir_path) / f'baseline.{self.chrom}.annot.gz'
+                if not baseline_annotation_path.exists():
+                    raise FileNotFoundError(f'baseline.{self.chrom}.annot.gz is not found in {self.additional_baseline_annotation_dir_path}.')
+
+
+
+@dataclass
+class SpatialLDSCConfig:
+    sample_name: str
+    w_file: str
+    ldscore_input_dir: str
+    ldsc_save_dir: str
+    disable_additional_baseline_annotation: bool = False
+    trait_name: str = None
+    sumstats_file: str = None
+    sumstats_config_file: str = None
+    num_processes: int = 4
+    not_M_5_50: bool = False
+    n_blocks: int = 200
+    chisq_max: int = None
+    all_chunk: int = None
+
+    def __post_init__(self):
+        if self.sumstats_file is None and self.sumstats_config_file is None:
+            raise ValueError('One of sumstats_file and sumstats_config_file must be provided.')
+        if self.sumstats_file is not None and self.sumstats_config_file is not None:
+            raise ValueError('Only one of sumstats_file and sumstats_config_file must be provided.')
+        if self.sumstats_file is not None and self.trait_name is None:
+            raise ValueError('trait_name must be provided if sumstats_file is provided.')
+        if self.sumstats_config_file is not None and self.trait_name is not None:
+            raise ValueError('trait_name must not be provided if sumstats_config_file is provided.')
+        self.sumstats_config_dict = {}
+        # load the sumstats config file
+        if self.sumstats_config_file is not None:
+            import yaml
+            with open(self.sumstats_config_file) as f:
+                config = yaml.load(f, Loader=yaml.FullLoader)
+            for trait_name, sumstats_file in config.items():
+                self.sumstats_config_dict[trait_name] = sumstats_file
+        # load the sumstats file
+        elif self.sumstats_file is not None:
+            self.sumstats_config_dict[self.trait_name] = self.sumstats_file
+        else:
+            raise ValueError('One of sumstats_file and sumstats_config_file must be provided.')
+
+        # check if additional baseline annotation is exist
+        self.use_additional_baseline_annotation = False
+        self.process_additional_baseline_annotation()
+
+    def process_additional_baseline_annotation(self):
+        additional_baseline_annotation_dir_path = Path(self.ldscore_input_dir) / 'additional_baseline'
+        dir_exists = additional_baseline_annotation_dir_path.exists()
+
+        if not dir_exists:
+            if self.use_additional_baseline_annotation:
+                logger.warning(f"additional_baseline directory is not found in {self.ldscore_input_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. You should disable use_additional_baseline_annotation')
+            return
+
+        self.use_additional_baseline_annotation = self.use_additional_baseline_annotation or True
+
+        if self.disable_additional_baseline_annotation:
+            logger.warning(
+                f"additional_baseline directory is found in {self.ldscore_input_dir}, but use_additional_baseline_annotation is disabled.")
+            print('''\
+                if you want to use additional baseline annotation,
+                please enable by not adding --disable_additional_baseline_annotation.
+                ''')
+            self.use_additional_baseline_annotation = False
+        else:
+            logger.info(
+                f'------Additional baseline annotation is provided. It will be used with the default baseline annotation.')
+            logger.info(f'------Additional baseline annotation directory: {additional_baseline_annotation_dir_path}')
+
+            chrom_list = range(1, 23)
+            for chrom in chrom_list:
+                baseline_annotation_path = additional_baseline_annotation_dir_path / f'baseline.{chrom}.l2.ldscore.feather'
+                if not baseline_annotation_path.exists():
+                    raise FileNotFoundError(
+                        f'baseline.{chrom}.annot.gz is not found in {additional_baseline_annotation_dir_path}.')
+
+
+logger
+@dataclass
+class CauchyCombinationConfig:
+    input_hdf5_path: str
+    input_ldsc_dir: str
+    output_cauchy_dir: str
+    sample_name: str
+    trait_name: str
+    annotation: str
+    meta: str = None
+    slide: str = None
+
+
+@dataclass
+class VisualizeConfig:
+    input_hdf5_path: str
+    input_ldsc_dir: str
+    output_figure_dir: str
+    sample_name: str
+    trait_name: str
+
+    annotation: str = None
+    fig_title: str = None
+    fig_height: int = 600
+    fig_width: int = 800
+    point_size: int = None
+    fig_style: Literal['dark', 'light'] = 'light'
+
+
+@dataclass
+class RunAllModeConfig:
+    flr_config: FindLatentRepresentationsConfig
+    ltg_config: LatentToGeneConfig
+    gls_config: GenerateLDScoreConfig
+    ldsc_config: SpatialLDSCConfig
+    cauchy_config: CauchyCombinationConfig
+
+
+@dataclass
+class FormatSumstatsConfig:
+    sumstats: str
+    out: str
+    dbsnp: str
+    snp: str = None
+    a1: str = None
+    a2: str = None
+    info: str = None
+    beta: str = None
+    se: str = None
+    p: str = None
+    frq: str = None
+    n: str = None
+    z: str = None
+    OR: str = None
+    se_OR: str = None
+    format: str = None
+    chr: str = None
+    pos: str = None
+    chunksize: int = 1e+7
+    info_min: float = 0.9
+    maf_min: float = 0.01
+    keep_chr_pos:bool = False
+
+
+@register_cli(name='run_find_latent_representations',
+              description='Run Find_latent_representations \nFind the latent representations of each spot by running GNN-VAE',
+              add_args_function=add_find_latent_representations_args)
+def run_find_latent_representation_from_cli(args: argparse.Namespace):
+    from gsMap.find_latent_representation import run_find_latent_representation
+    config = get_dataclass_from_parser(args, FindLatentRepresentationsConfig)
+    run_find_latent_representation(config)
+
+
+@register_cli(name='run_latent_to_gene',
+              description='Run Latent_to_gene \nEstimate gene marker gene scores for each spot by using latent representations from nearby spots',
+              add_args_function=add_latent_to_gene_args)
+def run_latent_to_gene_from_cli(args: argparse.Namespace):
+    from gsMap.latent_to_gene import run_latent_to_gene
+    config = get_dataclass_from_parser(args, LatentToGeneConfig)
+    run_latent_to_gene(config)
+
+
+@register_cli(name='run_generate_ldscore',
+              description='Run Generate_ldscore \nGenerate LD scores for each spot',
+              add_args_function=add_generate_ldscore_args)
+def run_generate_ldscore_from_cli(args: argparse.Namespace):
+    from gsMap.generate_ldscore import run_generate_ldscore
+    config = get_dataclass_from_parser(args, GenerateLDScoreConfig)
+    run_generate_ldscore(config)
+
+
+@register_cli(name='run_spatial_ldsc',
+              description='Run Spatial_ldsc \nRun spatial LDSC for each spot',
+              add_args_function=add_spatial_ldsc_args)
+def run_spatial_ldsc_from_cli(args: argparse.Namespace):
+    from gsMap.spatial_ldsc_multiple_sumstats import run_spatial_ldsc
+    config = get_dataclass_from_parser(args, SpatialLDSCConfig)
+    run_spatial_ldsc(config)
+
+
+@register_cli(name='run_cauchy_combination',
+              description='Run Cauchy_combination for each annotation',
+              add_args_function=add_Cauchy_combination_args)
+def run_Cauchy_combination_from_cli(args: argparse.Namespace):
+    from gsMap.cauchy_combination_test import run_Cauchy_combination
+    config = get_dataclass_from_parser(args, CauchyCombinationConfig)
+    run_Cauchy_combination(config)
+
+
+@register_cli(name='run_visualize',
+              description='Visualize the gsMap results',
+              add_args_function=add_Visualization_args)
+def run_Visualize_from_cli(args: argparse.Namespace):
+    from gsMap.visualize import run_Visualize
+    config = get_dataclass_from_parser(args, VisualizeConfig)
+    run_Visualize(config)
+
+
+@register_cli(name='run_all_mode',
+              description='Run gsMap method (the full process)',
+              add_args_function=add_all_mode_args)
+def run_all_mode_from_cli(args: argparse.Namespace):
+    from gsMap.find_latent_representation import run_find_latent_representation
+    from gsMap.latent_to_gene import run_latent_to_gene
+    from gsMap.generate_ldscore import run_generate_ldscore
+    from gsMap.spatial_ldsc_multiple_sumstats import run_spatial_ldsc
+    from gsMap.cauchy_combination_test import run_Cauchy_combination
+    config = get_runall_mode_config(args)
+    run_find_latent_representation(config.flr_config)
+    run_latent_to_gene(config.ltg_config)
+    run_generate_ldscore(config.gls_config)
+    run_spatial_ldsc(config.ldsc_config)
+    if args.annotation is not None:
+        config.cauchy_config.annotation = args.annotation
+        run_Cauchy_combination(config.cauchy_config)
+
+@register_cli(name='format_sumstats',
+              description='Format gwas summary statistics',
+              add_args_function=add_format_sumstats_args)
+def gwas_format_from_cli(args: argparse.Namespace):
+    from gsMap.format_sumstats import gwas_format
+    config = get_dataclass_from_parser(args, FormatSumstatsConfig)
+    gwas_format(config)