CATH非冗余数据集中复合体结构筛选

从CATH网站下载数据（CATH）

wget http://download.cathdb.info/cath/releases/latest-release/non-redundant-data-sets/cath-dataset-nonredundant-S40.list ./

为了从CATH数据库的cath-dataset-nonredundant-S40.list文件中找到含有多条多肽链的复合体结构，整个流程需要使用CATH数据库和PDB文件的结合分析，脚本可以用Python或其他脚本语言编写。以下是大致的分析步骤、可能用到的工具和脚本思路。

数据分析代码

import pandas as pd

### 1. 读取CATH数据
cath_file = '/Users/zhengxueming/Downloads/cath-dataset-nonredundant-S40.list'

pdb_ids = []
chain_ids = []
pdb_chain_ids = []
domain_ids = []

with open(cath_file, 'r') as file:
    for line in file:
        line = line.strip()
        #print(line.strip())  # 使用 strip() 去除每行末尾的换行符
        # 1a15A00
        pdb_id = line[:4]  # 提取PDB ID 
        chain_id = line[4:5] # 提取链ID
        domain_id = line[5:] # 提取结构域ID
        
        pdb_ids.append(pdb_id)
        chain_ids.append(chain_id)
        domain_ids.append(domain_id)      
        
data = {'PDB_ID': pdb_ids, 'Chain_ID': chain_ids, 'Domain_ID':domain_ids}       
cath_data = pd.DataFrame(data)
print(cath_data.shape)

#print(cath_data)

### 2. 提取单一链结构
single_chain_pdbs = cath_data.groupby('PDB_ID').filter(lambda x: x['Chain_ID'].nunique() == 1)
print("单一链的结构有：{}个".format(single_chain_pdbs.shape[0]))

### 3. 提取单一链单一结构域的结构
# 在只包含单条链的结构中，筛选出单一结构域的链
single_chain_single_domain = single_chain_pdbs.groupby(['PDB_ID', 'Chain_ID']).filter(lambda x: len(x) == 1)
print("单一链单一结构域的结构有：{}个".format(single_chain_single_domain.shape[0]))

### 4. 提取单一链多结构域的结构
single_chain_multi_domain = single_chain_pdbs.groupby(['PDB_ID', 'Chain_ID']).filter(lambda x: len(x) > 1)
print("单一链多结构域的结构有：{}个".format(single_chain_multi_domain.shape[0]))

# 5. 提取多条链的结构
multi_chain_pdbs = cath_data.groupby('PDB_ID').filter(lambda x: len(x['Chain_ID'].unique()) > 1)
print("多条链的结构有：{}个".format(multi_chain_pdbs.shape[0]))

进一步，可以根据蛋白质复合体中序列的同源性区分出homo和hetero蛋白质复合体，并根据链长度和结构的分辨率进一步筛选。