Drugbank的医药数据XML解析

pyhon解析drugbank数据库。代码[原文地址]

#!/usr/bin/python3

import untangle
import pandas as pd
import numpy as np
import os

 
filename="drugbank_all_full_database.xml"
obj=untangle.parse(filename)

#Building dataframe of chemical descriptors
#Data Frame of DrugBank Small Molecule Type Drugs
df_drugbank_sm=pd.DataFrame(columns=["drugbank_id","name","cas","smiles","logP ALOGPS", "logP ChemAxon", "solubility ALOGPS", "pKa (strongest acidic)", "pKa (strongest basic)"])
df_drugbank_sm 
 
i=-1
#iterate over drug entries to extract information
for drug in obj.drugbank.drug:
    drug_type= str(drug["type"])

# select for small molecule drugs
if drug_type in ["small molecule", "Small Molecule", "Small molecule"]:
        i=i+1

#Get drugbank_id
for id in drug.drugbank_id:
if str(id["primary"])=="true":
                df_drugbank_sm.loc[i, "drugbank_id"]=id.cdata
#Drug name
        df_drugbank_sm.loc[i,"name"]=drug.name.cdata

#Drug CAS
        df_drugbank_sm.loc[i, "cas"]=drug.cas_number.cdata

#Get SMILES, logP, Solubility
#Skip drugs with no structure. ("DB00386","DB00407","DB00702","DB00785","DB00840",
#                                            "DB00893","DB00930","DB00965", "DB01109","DB01266",
#                                           "DB01323", "DB01341"...)
if len(drug.calculated_properties.cdata)==0: #If there is no calculated properties
continue
else:
for property in drug.calculated_properties.property:
if property.kind.cdata == "SMILES":
                    df_drugbank_sm.loc[i, "smiles"]=property.value.cdata

if property.kind.cdata == "logP":
if property.source.cdata == "ALOGPS":
                        df_drugbank_sm.loc[i, "logP ALOGPS"]=property.value.cdata
if property.source.cdata == "ChemAxon":
                        df_drugbank_sm.loc[i, "logP ChemAxon"]=property.value.cdata

if property.kind.cdata == "Water Solubility":
                    df_drugbank_sm.loc[i, "solubility ALOGPS"]=property.value.cdata

if property.kind.cdata == "pKa (strongest acidic)":
                    df_drugbank_sm.loc[i, "pKa (strongest acidic)"]=property.value.cdata

if property.kind.cdata == "pKa (strongest basic)":
                    df_drugbank_sm.loc[i, "pKa (strongest basic)"]=property.value.cdata
          
df_drugbank_sm.head(10)
print (df_drugbank_sm.shape) 
 
#Drop drugs without SMILES from the dataframe
df_drugbank_smiles = df_drugbank_sm.dropna()
df_drugbank_smiles= df_drugbank_smiles.reset_index(drop=True)
print(df_drugbank_smiles.shape) 
 
df_drugbank_smiles.head() 
 
#write to csv
df_drugbank_smiles.to_csv("drugbank_smiles.csv", encoding='utf-8',index=False)