{ "cells": [ { "cell_type": "code", "execution_count": 125, "id": "76129cb7-363c-4efc-a10b-daf4edc7033a", "metadata": {}, "outputs": [], "source": [ "import yaml\n", "import numpy as np\n", "import os\n", "import pandas as pd\n", "from tqdm import tqdm\n", "\n", "\n", "\n", "def read_filter_gff_data(config):\n", " gff_path = os.path.join(config['input_resource']['root'], config['input_resource']['sequence'], config['preprocessing_data_config']['gff_file'])\n", " print(f'Reading gff file from {gff_path}')\n", " gff_df = read_genes(gff_path)\n", "\n", " # Remove genes on chrX, chrY, and chrM\n", " gff_df = gff_df[~gff_df['seqname'].isin(['chrX', 'chrY', 'chrM'])]\n", "\n", " # Keep a few transcript_types\n", " transcript_types = ['protein_coding', 'lincRNA', 'miRNA', 'snRNA', 'snoRNA', 'rRNA']\n", " gff_df = gff_df[gff_df['transcript_type'].isin(transcript_types)]\n", "\n", " na_start = gff_df['start'].isna()\n", " na_end = gff_df['end'].isna()\n", " na_gene = gff_df[na_start | na_end]\n", " if len(na_gene) > 0:\n", " print(f'Warning: {len(na_gene)} genes with NA start or end')\n", " print(na_gene['gene_name'].to_string(index=False))\n", "\n", " gff_df = gff_df[~(na_start | na_end)]\n", "\n", " # Add TSS column\n", " gff_df['tss'] = np.where(gff_df['strand'] == '+', gff_df['start'], gff_df['end']).astype(int)\n", "\n", " tss_radius = config['loci_source_config']['atac_seq_specific_config']['promoter']['tss_radius']\n", " gff_df = add_inference_loci(gff_df, tss_radius, config)\n", " return gff_df\n", "\n", "def save_gene_list(gene_type_dict, config):\n", " save_root = config[\"preprocessing_data_config\"][\"output\"][\"path\"] + '/gene_list'\n", " os.makedirs(save_root, exist_ok=True)\n", " # Save gene list and merged bed file for inference\n", " for gene_list_type, gene_list_df in gene_type_dict.items():\n", " # Save gene list\n", " gene_list_df.to_csv(f'{save_root}/{gene_list_type}_gene_list.csv', index=False)\n", " # Save bed file\n", " bed_df = gene_list_df[['seqname', 'tss_inference_start', 'tss_inference_end', 'transcript_name']].copy()\n", " bed_df['tss_inference_start'] = bed_df['tss_inference_start'].astype(int)\n", " bed_df['tss_inference_end'] = bed_df['tss_inference_end'].astype(int)\n", " bed_df['score'] = 0\n", " bed_df['strand'] = gene_list_df['strand']\n", " bed_df.to_csv(f'{save_root}/{gene_list_type}_gene_list.bed', sep='\\t', index=False, header=False)\n", "\n", " # Merge loci for inference\n", " merge_loci(gene_type_dict, save_root)\n", "\n", "def merge_loci(gene_type_dict, save_root):\n", " # Merge overlapping loci from all gene lists\n", " merged_df = pd.concat(gene_type_dict.values())\n", " \n", " # Sort by chromosome and start position\n", " merged_df = merged_df.sort_values(['seqname', 'tss_inference_start'])\n", " \n", " # Merge overlapping regions\n", " merged_regions = []\n", " current_chrom = None\n", " current_start = None \n", " current_end = None\n", " current_names = []\n", " current_strands = []\n", " \n", " for _, row in merged_df.iterrows():\n", " if current_chrom != row['seqname'] or row['tss_inference_start'] > current_end:\n", " # Save previous merged region\n", " if current_chrom is not None:\n", " merged_regions.append({\n", " 'seqname': current_chrom,\n", " 'start': current_start,\n", " 'end': current_end,\n", " 'transcript_name': '|'.join(current_names),\n", " 'score': 0,\n", " 'strand': '|'.join(current_strands)\n", " })\n", " # Start new region\n", " current_chrom = row['seqname']\n", " current_start = row['tss_inference_start']\n", " current_end = row['tss_inference_end']\n", " current_names = [row['transcript_name']]\n", " current_strands = [row['strand']]\n", " else:\n", " # Extend current region\n", " current_end = max(current_end, row['tss_inference_end'])\n", " current_names.append(row['transcript_name'])\n", " current_strands.append(row['strand'])\n", " \n", " # Add final region\n", " if current_chrom is not None:\n", " merged_regions.append({\n", " 'seqname': current_chrom,\n", " 'start': current_start,\n", " 'end': current_end,\n", " 'transcript_name': '|'.join(current_names),\n", " 'score': 0,\n", " 'strand': '|'.join(current_strands)\n", " })\n", " \n", " # Convert to dataframe and save\n", " merged_df = pd.DataFrame(merged_regions)\n", " merged_df.to_csv(f'{save_root}/merged_loci.csv', index=False)\n", " merged_df.to_csv(f'{save_root}/merged_loci.bed', sep='\\t', index=False, header=False)\n", "\n", "def get_filter_location(diff_df, top_n):\n", " # Make sure not too many loci and from the same gene\n", " g_idx = 0\n", " o_idx = 0\n", " gene_set = set()\n", " if 'gene_name' in diff_df.columns:\n", " gene_list = diff_df['gene_name'].values\n", " for g_idx, gene in enumerate(gene_list):\n", " gene_set.add(gene)\n", " if len(gene_set) == top_n:\n", " break\n", " # Make sure not too many loci and from the same region\n", " overlap_threshold = 0.5\n", " region_df = diff_df[['seqname', 'tss_inference_start', 'tss_inference_end']]\n", " unique_region_length = 0\n", " for o_idx, row in region_df.iterrows():\n", " if o_idx == 0:\n", " continue\n", " prior_df = region_df.iloc[:o_idx]\n", " same_chrom_df = prior_df[prior_df['seqname'] == row['seqname']]\n", " # Calculate if there are overlaps over 50%\n", " starts = same_chrom_df['tss_inference_start'].values\n", " ends = same_chrom_df['tss_inference_end'].values\n", " current_start = row['tss_inference_start']\n", " current_end = row['tss_inference_end']\n", " overlaps = np.maximum(0, np.minimum(current_end, ends) - np.maximum(current_start, starts))\n", " half_length = (current_end - current_start) / 2\n", " if np.sum(overlaps) / half_length > overlap_threshold:\n", " continue\n", " unique_region_length += 1\n", " if unique_region_length > top_n:\n", " break\n", " return max(g_idx, o_idx), max(len(gene_set), unique_region_length)\n", "\n", "def filter_by_rna_promoter_accessibility(gene_list_dict, config):\n", "\n", " for gene_list_type, diff_df in gene_list_dict.items():\n", " # Remove genes with small accessibility \n", " accessibility_threshold = config['loci_source_config']['atac']['active_locus_threshold']\n", " cross_condition_max_accessibility = diff_df[['experimental_max_accessibility', 'control_max_accessibility']].max(axis=1)\n", " select_idx = cross_condition_max_accessibility > accessibility_threshold\n", " gene_list_dict[gene_list_type] = diff_df[select_idx]\n", "\n", " upregulated_df, downregulated_df, conserved_df = gene_list_dict['upregulated'], gene_list_dict['downregulated'], gene_list_dict['conserved']\n", "\n", " acc_log2fc_threshold = config['loci_source_config']['atac']['log2fc_threshold']\n", "\n", " num_diff = config['loci_source_config']['rna']['top_diff_genes']\n", " upregulated_df = add_log2fc(upregulated_df)\n", " upregulated_df = upregulated_df[upregulated_df['accessibility_log2fc'] > acc_log2fc_threshold]\n", " upregulated_df = upregulated_df.sort_values(by='abs_rna_acc_log2fc_sum', ascending=False).reset_index(drop=True)\n", " filter_loc, num_unique = get_filter_location(upregulated_df, num_diff) # Get the location of the top num_diff unique genes\n", " upregulated_df = upregulated_df.head(filter_loc).reset_index(drop=True)\n", " if num_unique < num_diff:\n", " print(f'Warning: Only {num_unique} unique upregulated genes found after filtering by promoter accessibility')\n", "\n", " downregulated_df = add_log2fc(downregulated_df)\n", " downregulated_df = downregulated_df[downregulated_df['accessibility_log2fc'] < -acc_log2fc_threshold]\n", " downregulated_df = downregulated_df.sort_values(by='abs_rna_acc_log2fc_sum', ascending=False).reset_index(drop=True)\n", " filter_loc, num_unique = get_filter_location(downregulated_df, num_diff) # Get the location of the top num_diff unique genes\n", " downregulated_df = downregulated_df.head(filter_loc).reset_index(drop=True)\n", " if num_unique < num_diff:\n", " print(f'Warning: Only {num_unique} unique downregulated genes found after filtering by promoter accessibility')\n", "\n", " num_conserved = config['loci_source_config']['rna']['top_conserved_genes']\n", " conserved_df = add_log2fc(conserved_df)\n", " conserved_df = conserved_df.sort_values(by='abs_rna_acc_log2fc_sum', ascending=True).reset_index(drop=True)\n", " filter_loc, num_unique = get_filter_location(conserved_df, num_conserved) # Get the location of the top num_conserved unique genes\n", " conserved_df = conserved_df.head(filter_loc).reset_index(drop=True)\n", " if num_unique < num_conserved:\n", " print(f'Warning: Only {num_unique} unique conserved genes found after filtering by promoter accessibility')\n", "\n", " gene_list_dict = {'upregulated': upregulated_df, 'downregulated': downregulated_df, 'conserved': conserved_df}\n", "\n", " return gene_list_dict\n", " \n", "def add_log2fc(diff_df):\n", " diff_df['accessibility_log2fc'] = np.log2(diff_df['experimental_max_accessibility'] / diff_df['control_max_accessibility'])\n", " diff_df['rna_acc_log2fc_sum'] = diff_df['log2fc'] + diff_df['accessibility_log2fc']\n", " diff_df['abs_rna_acc_log2fc_sum'] = diff_df['rna_acc_log2fc_sum'].abs()\n", " return diff_df\n", "\n", "def get_promoter_accessibility(diff_df, config, lazy=True):\n", " import zarr\n", " # Load experimental and control zarr files\n", " zarr_path_dict = {condition: [os.path.join(config['input_resource']['root'], config['input_resource']['atac'], f\"{rep}.zarr\") for rep in config['loci_source_config']['atac']['data_path'][condition].split(',')] for condition in ['experimental', 'control']}\n", " if lazy:\n", " zarr_dict = {condition: [zarr.open(zarr_pa, mode='r') for zarr_pa in zarr_path] for condition, zarr_path in zarr_path_dict.items()}\n", " else:\n", " print('Loading zarr files to memory')\n", " zarr_dict = {condition: [load_zarr_to_memory(zarr_pa) for zarr_pa in zarr_path] for condition, zarr_path in zarr_path_dict.items()}\n", "\n", " for zarr_condition, zarr_data in zarr_dict.items():\n", " # Filter by promoter accessibility\n", " max_accessibility_list = []\n", " mean_accessibility_list = []\n", " # Load promoter accessibility data by chromosome\n", " print(f'Loading {zarr_condition} promoter accessibility data')\n", " import tqdm\n", " for row_idx, row in tqdm.tqdm(diff_df.iterrows(), total=len(diff_df)):\n", " chrom = row['seqname']\n", " start = row['tss_inference_start']\n", " end = row['tss_inference_end']\n", " max_accessibility_list.append(np.max(np.stack([zarr_da['chrs'][chrom][start:end] for zarr_da in zarr_data],axis=0)))\n", " mean_accessibility_list.append(np.mean(np.stack([zarr_da['chrs'][chrom][start:end] for zarr_da in zarr_data],axis=0)))\n", " diff_df[f'{zarr_condition}_max_accessibility'] = max_accessibility_list\n", " diff_df[f'{zarr_condition}_mean_accessibility'] = mean_accessibility_list\n", " return diff_df\n", "\n", "def load_zarr_to_memory(zarr_path):\n", " import zarr\n", " zarr_data = zarr.open(zarr_path, mode='r')\n", " zarr_data_dict = {'chrs': {}}\n", " print(f'Loading to memory: {zarr_path}')\n", " for chrom in tqdm(zarr_data['chrs'].keys()):\n", " zarr_data_dict['chrs'][chrom] = zarr_data['chrs'][chrom][:]\n", " return zarr_data_dict\n", "\n", "def read_rna_data(rna_config):\n", " diff_exp_df = pd.read_csv(rna_config['log2fc_csv'])\n", " assert 'gene_name' in diff_exp_df.columns and 'experimental_exp' in diff_exp_df.columns and 'control_exp' in diff_exp_df.columns and 'log2fc' in diff_exp_df.columns\n", " return diff_exp_df\n", "\n", "def add_loci(diff_exp_df, config):\n", " # Add loci to the gene list\n", " gff_path = os.path.join(config['input_resource']['root'], config['input_resource']['sequence'], config['preprocessing_data_config']['gff_file'])\n", " diff_exp_df = get_gene_loci(diff_exp_df, gff_path)\n", " tss_radius = config['loci_source_config']['rna']['tss_radius']\n", " diff_exp_df = add_inference_loci(diff_exp_df, tss_radius, config)\n", " return diff_exp_df\n", "\n", "def add_inference_loci(diff_exp_df, tss_radius, config):\n", " diff_exp_df['tss_inference_start'] = diff_exp_df['tss'] - tss_radius\n", " diff_exp_df['tss_inference_end'] = diff_exp_df['tss'] + tss_radius\n", "\n", " # Clip to chromosome length\n", " chr_sizes = load_chr_sizes(os.path.join(config['input_resource']['root'], config['input_resource']['sequence'], config['preprocessing_data_config']['chr_sizes_file']))\n", "\n", " diff_exp_df['tss_inference_start'] = np.maximum(diff_exp_df['tss_inference_start'], 0)\n", " diff_exp_df['tss_inference_end'] = np.minimum(diff_exp_df['tss_inference_end'], diff_exp_df['seqname'].map(chr_sizes))\n", "\n", " return diff_exp_df\n", "\n", "def atac_split_filter_gene_type(diff_df, config):\n", " # Add log2fc\n", " diff_df['accessibility_log2fc'] = np.log2(diff_df['experimental_max_accessibility'] / diff_df['control_max_accessibility'])\n", "\n", " # Filter by accessiblity\n", " accessibility_threshold = config['loci_source_config']['atac']['active_locus_threshold']\n", " cross_condition_max_accessibility = diff_df[['experimental_max_accessibility', 'control_max_accessibility']].max(axis=1)\n", " select_idx = cross_condition_max_accessibility > accessibility_threshold\n", " diff_df = diff_df[select_idx]\n", "\n", " # Split by absolute log2fc\n", " diff_df['abs_accessibility_log2fc'] = diff_df['accessibility_log2fc'].abs()\n", " diff_df = diff_df.sort_values(by='abs_accessibility_log2fc', ascending=False)\n", "\n", " # Get conserved genes\n", " atac_config = config['loci_source_config']['atac']\n", " num_conserved = atac_config['top_conserved_genes']\n", " conserved_df = diff_df.copy().sort_values(by='abs_accessibility_log2fc', ascending=True).reset_index(drop=True)\n", " filter_loc, num_unique = get_filter_location(conserved_df, num_conserved) # Get the location of the top num_conserved unique genes\n", " conserved_df = conserved_df.head(filter_loc).reset_index(drop=True)\n", " if num_unique < num_conserved:\n", " print(f'Warning: Only {num_unique} unique conserved genes found after filtering by promoter accessibility')\n", "\n", " # Get upregulated and downregulated genes\n", " atac_threshold = atac_config['log2fc_threshold']\n", " num_diff = atac_config['top_diff_genes']\n", " upregulated_df = diff_df[diff_df['accessibility_log2fc'] > atac_threshold].copy().sort_values(by='accessibility_log2fc', ascending=False).reset_index(drop=True)\n", " filter_loc, num_unique = get_filter_location(upregulated_df, num_diff) # Get the location of the top num_diff unique genes\n", " upregulated_df = upregulated_df.head(filter_loc).reset_index(drop=True)\n", " if num_unique < num_diff:\n", " print(f'Warning: Only {num_unique} unique upregulated genes found after filtering by promoter accessibility')\n", "\n", " downregulated_df = diff_df[diff_df['accessibility_log2fc'] < -atac_threshold].copy().sort_values(by='accessibility_log2fc', ascending=True).reset_index(drop=True)\n", " filter_loc, num_unique = get_filter_location(downregulated_df, num_diff) # Get the location of the top num_diff unique genes\n", " downregulated_df = downregulated_df.head(filter_loc).reset_index(drop=True)\n", " if num_unique < num_diff:\n", " print(f'Warning: Only {num_unique} unique downregulated genes found after filtering by promoter accessibility')\n", "\n", " return {'upregulated': upregulated_df, 'downregulated': downregulated_df, 'conserved': conserved_df}\n", "\n", "def split_gene_type(diff_exp_df, rna_config):\n", " # Split by absolute log2fc\n", " diff_exp_df['abs_log2fc'] = diff_exp_df['log2fc'].abs()\n", " diff_exp_df = diff_exp_df.sort_values(by='abs_log2fc', ascending=False)\n", "\n", " # Get conserved genes\n", " num_conserved = rna_config['top_conserved_genes'] * 4 # Leave room for atac-seq filtering\n", " conserved_genes = diff_exp_df.sort_values(by='abs_log2fc', ascending=True).head(num_conserved).reset_index(drop=True)\n", "\n", " # Get upregulated and downregulated genes\n", " rna_threshold = rna_config['log2fc_threshold']\n", " num_diff = rna_config['top_diff_genes'] * 2 # Leave room for atac-seq filtering\n", " upregulated_genes = diff_exp_df[diff_exp_df['log2fc'] > rna_threshold].sort_values(by='log2fc', ascending=False).head(num_diff).reset_index(drop=True)\n", " downregulated_genes = diff_exp_df[diff_exp_df['log2fc'] < -rna_threshold].sort_values(by='log2fc', ascending=True).head(num_diff).reset_index(drop=True)\n", "\n", " return {'upregulated': upregulated_genes, 'downregulated': downregulated_genes, 'conserved': conserved_genes}\n", "\n", "def get_gene_loci(gene_df, gff_path):\n", " # Find location of genes in the genome with a annotation file\n", " print(f'Reading gff file from {gff_path}')\n", " gff_df = read_genes(gff_path)\n", " # Find the gene names in the annotation file\n", " gene_names = gene_df['gene_name'].values\n", " selected_gene_gff_df = []\n", " print(f'Finding gene loci for {len(gene_names)} genes')\n", " for gene_name in gene_names:\n", " gene_gff_df = gff_df[gff_df['gene_name'] == gene_name]\n", " selected_gene_gff_df.append(gene_gff_df)\n", " selected_gene_gff_df = pd.concat(selected_gene_gff_df).reset_index(drop=True)\n", " # Join the split_df with the gff_df\n", " merged_df = gene_df.merge(selected_gene_gff_df, on='gene_name', how='left')\n", "\n", " # Only keep gene on chr1-22 and X\n", " merged_df = merged_df[merged_df['seqname'].isin([f'chr{i}' for i in range(1, 23)] + ['chrX'])]\n", "\n", " # Remove genes with NA start or end and print warning\n", " na_start = merged_df['start'].isna()\n", " na_end = merged_df['end'].isna()\n", " na_gene = merged_df[na_start | na_end]\n", " if len(na_gene) > 0:\n", " print(f'Warning: {len(na_gene)} genes with NA start or end')\n", " print(na_gene['gene_name'].to_string(index=False))\n", "\n", " merged_df = merged_df[~(na_start | na_end)]\n", "\n", " # Add TSS column\n", " merged_df['tss'] = np.where(merged_df['strand'] == '+', merged_df['start'], merged_df['end']).astype(int)\n", "\n", " return merged_df\n", "\n", "def read_genes(gff_path):\n", " names = ['seqname', 'source', 'feature', 'start', 'end', 'score', 'strand', 'frame', 'attribute']\n", " genes = pd.read_csv(gff_path, sep='\\t', comment='#', names=names)\n", " gene_names = []\n", " transcript_types = []\n", " transcript_names = []\n", " for attribute in genes['attribute']:\n", " try:\n", " gene_field, gene_name = attribute.split(';')[5].split('=')\n", " transcript_type_field, transcript_type = attribute.split(';')[6].split('=')\n", " transcript_field, transcript_name = attribute.split(';')[7].split('=')\n", " assert gene_field == 'gene_name'\n", " assert transcript_type_field == 'transcript_type'\n", " assert transcript_field == 'transcript_name'\n", " gene_names.append(gene_name)\n", " transcript_types.append(transcript_type)\n", " transcript_names.append(transcript_name)\n", " except Exception as e:\n", " print(f'Error: {e}')\n", " gene_names.append('')\n", " transcript_names.append('')\n", " genes['gene_name'] = gene_names\n", " genes['transcript_name'] = transcript_names\n", " genes['transcript_type'] = transcript_types\n", " return genes\n", "\n", "def load_chr_sizes(chr_size_path):\n", " chr_sizes = {}\n", " with open(chr_size_path) as f:\n", " for line in f:\n", " chrom, size = line.strip().split()\n", " chr_sizes[chrom] = int(size)\n", " return chr_sizes\n", "\n", "def copy_merged_loci_to_inputs(config):\n", " source_path = os.path.join(config[\"preprocessing_data_config\"][\"output\"][\"path\"], 'gene_list', 'merged_loci.bed')\n", " dest_path = os.path.join(config[\"inference_config\"][\"input\"][\"root\"], config[\"inference_config\"][\"input\"][\"locus_list_path\"])\n", " os.makedirs(os.path.dirname(dest_path), exist_ok=True)\n", " import shutil\n", " shutil.copy(source_path, dest_path)\n", "\n", "\n" ] }, { "cell_type": "code", "execution_count": 98, "id": "feae8583-5dfc-4275-acbd-3987215232ba", "metadata": {}, "outputs": [ { "ename": "IndentationError", "evalue": "unexpected indent (1107164402.py, line 43)", "output_type": "error", "traceback": [ " \u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[98]\u001b[39m\u001b[32m, line 43\u001b[39m\n\u001b[31m \u001b[39m\u001b[31melif atac_seq_specific_config['mode'] == 'All_peaks':\u001b[39m\n ^\n\u001b[31mIndentationError\u001b[39m\u001b[31m:\u001b[39m unexpected indent\n" ] } ], "source": [ "\n", "# Load config\n", "with open('config.yaml', 'r') as f:\n", " config = yaml.safe_load(f)\n", "\n", "# Get differential vs conserved gene lists\n", "# RNA-seq mode\n", "\n", "# ATAC-seq mode\n", "atac_seq_specific_config = config['loci_source_config']['atac_seq_specific_config']\n", "\n", "atac_config = config['loci_source_config']['atac']\n", "diff_acc_df = read_filter_gff_data(config)\n", "###read diff_acc_df\n", "gff_path = os.path.join(config['input_resource']['root'], config['input_resource']['sequence'], config['preprocessing_data_config']['gff_file'])\n", "print(f'Reading gff file from {gff_path}')\n", "gff_df = read_genes(gff_path)\n", "\n", "# Remove genes on chrX, chrY, and chrM\n", "gff_df = gff_df[~gff_df['seqname'].isin(['chrX', 'chrY', 'chrM'])]\n", "\n", "# Keep a few transcript_types\n", "transcript_types = ['protein_coding', 'lincRNA', 'miRNA', 'snRNA', 'snoRNA', 'rRNA']\n", "gff_df = gff_df[gff_df['transcript_type'].isin(transcript_types)]\n", "\n", "na_start = gff_df['start'].isna()\n", "na_end = gff_df['end'].isna()\n", "na_gene = gff_df[na_start | na_end]\n", "if len(na_gene) > 0:\n", " print(f'Warning: {len(na_gene)} genes with NA start or end')\n", " print(na_gene['gene_name'].to_string(index=False))\n", "\n", "gff_df = gff_df[~(na_start | na_end)]\n", "\n", "# Add TSS column\n", "gff_df['tss'] = np.where(gff_df['strand'] == '+', gff_df['start'], gff_df['end']).astype(int)\n", "\n", "tss_radius = config['loci_source_config']['atac_seq_specific_config']['promoter']['tss_radius']\n", "gff_df = add_inference_loci(gff_df, tss_radius, config)\n", "\n", "\n", "diff_acc_df = get_promoter_accessibility(diff_acc_df, config, lazy=False)\n", "gene_type_dict = atac_split_filter_gene_type(diff_acc_df, config)\n", " elif atac_seq_specific_config['mode'] == 'All_peaks':\n", " raise NotImplementedError('All_peaks mode not implemented')\n", " else:\n", " raise ValueError(f'Invalid mode: {config[\"loci_source_config\"][\"mode\"]}')\n", "else:\n", " raise ValueError(f'Invalid mode: {config[\"loci_source_config\"][\"mode\"]}')\n", "\n", "# Save gene list and merged bed file for inference\n", "save_gene_list(gene_type_dict, config,config_name)\n", "\n", "# Copy merged_loci.bed to inputs folder\n", "copy_merged_loci_to_inputs(config,config_name)" ] } ], "metadata": { "kernelspec": { "display_name": "Python (transformer)", "language": "python", "name": "transformer" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.13" } }, "nbformat": 4, "nbformat_minor": 5 }