免费com域名网站中山建设网站官网

AlphaFold3 generate_chain_data_cache 脚本在源代码的scripts文件夹下。该脚本从指定目录中批量解析 mmCIF/PDB 文件的工具，并将每个链的基本信息（序列、分辨率、是否属于聚类等）提取并写入 JSON 文件，主要用于后续蛋白质建模、过滤或训练数据准备。

源代码：

import argparse
from functools import partial
import json
import logging
from multiprocessing import Pool
import os
import string
from collections import defaultdict
from tqdm import tqdm
from src.data.mmcif_parsing import parse
from src.common import protein, residue_constants
import sys
sys.path.append("../../../Downloads")  # an innocent hack to get this to run from the top leveldef parse_file(f, args,chain_cluster_size_dict
):file_id, ext = os.path.splitext(f)if ext == ".cif":with open(os.path.join(args.data_dir, f), "r") as fp:mmcif_string = fp.read()mmcif = parse(file_id=file_id, mmcif_string=mmcif_string)if mmcif.mmcif_object is None:logging.info(f"Could not parse {f}. Skipping...")return {}else:mmcif = mmcif.mmcif_objectout = {}for chain_id, seq in mmcif.chain_to_seqres.items():full_name = "_".join([file_id, chain_id])out[full_name] = {}local_data = out[full_name]local_data["release_date"] = mmcif.header["release_date"]local_data["seq"] = seqlocal_data["resolution"] = mmcif.header["resolution"]if chain_cluster_size_dict is not None:cluster_size = chain_cluster_size_dict.get(full_name.upper(), -1)local_data["cluster_size"] = cluster_sizeelif ext == ".pdb":with open(os.path.join(args.data_dir, f), "r") as fp:pdb_string = fp.read()protein_object = protein.from_pdb_string(pdb_string, None)aatype = pr