Literature_component.py

import pandas as pd
import numpy as np

########################################################
### PART 1: LOAD Metadata                            ###
########################################################

id_column      = 'pfcode'
species_column = 'species_desc'
genus_column   = 'genus_desc'
family_column  = 'family_desc'

metadata = pd.read_csv('../PF_metadata.csv',
            usecols=[id_column, species_column, genus_column, family_column],
            sep=',')#.drop(['Unnamed: 0'],axis=1)


########################################################
### PART 2: Paramters                                ###
########################################################

max_comp_reported_sp = 20          # max number of compounds reported at species level, more than this value, the plant is considered less interesting
max_comp_reported_g = 100          # max number of compounds reported at genus level,more than this value, the plant is considered less interesting
max_comp_reported_f = 500           # max number of compounds reported at genus level,more than this value, the plant is considered less interesting

calculate_LC = True

########################################################
### PART 3: Compute the Literature report            ###
########################################################
 
def literature_component(df, max_comp_reported_sp, max_comp_reported_g, max_comp_reported_f):
    """ function to compute the literature component based on the metadata and combinend information of the Dictionary of natural products and the Lotus DB, 
    Args:
        df2 = metadata_df

    Returns:
        None
    """
    LotusDB = pd.read_csv('../data_loc/LotusDB_inhouse_metadata.csv', 
                       sep=',').dropna()

    #create a set of species from the metadata table
    set_sp = set(df[species_column].dropna()) #dropna is used to erase all the QC, blanks, etc not having a species associated
    #create a set of genera from the metadata table 
    set_g = set(df[genus_column].dropna()) #dropna is used to erase all the QC, blanks, etc not having a species associated
    #create a set of families from the metadata table 
    set_f = set(df[family_column].dropna()) #dropna is used to erase all the QC, blanks, etc not having a species associated

    #reduce LotusDB to sp in set 
    LotusDBs= LotusDB[LotusDB['organism_taxonomy_09species'].isin(set_sp)]
    LotusDBs = LotusDBs[['organism_name','organism_taxonomy_09species','Reported_comp_Species']].drop_duplicates()
    LotusDBs['Reported_comp_Species'] =LotusDBs['Reported_comp_Species'].astype(int)

    #reduce LotusDB to genera in set 
    LotusDBg= LotusDB[LotusDB['organism_taxonomy_08genus'].isin(set_g)]
    LotusDBg = LotusDBg[['organism_taxonomy_08genus','Reported_comp_Genus']].drop_duplicates()
    LotusDBg['Reported_comp_Genus'] =LotusDBg['Reported_comp_Genus'].astype(int)

    #reduce LotusDB to families in set 
    LotusDBf= LotusDB[LotusDB['organism_taxonomy_06family'].isin(set_f)]
    LotusDBf = LotusDBf[['organism_taxonomy_06family','Reported_comp_Family']].drop_duplicates()
    LotusDBf['Reported_comp_Family'] =LotusDBf['Reported_comp_Family'].astype(int)
    #LotusDB.head()
        
    df = pd.merge(df[[id_column, family_column, genus_column, species_column]],
                LotusDBs[['organism_taxonomy_09species', 'Reported_comp_Species']],
                how= 'left', left_on=species_column, right_on='organism_taxonomy_09species').drop_duplicates(subset=[id_column])
    df.drop('organism_taxonomy_09species', axis=1, inplace=True)
    df = pd.merge(df,
                LotusDBg,
                how= 'left', left_on=genus_column, right_on='organism_taxonomy_08genus').drop_duplicates(subset=[id_column])
    df.drop('organism_taxonomy_08genus', axis=1, inplace=True)
    df = pd.merge(df,
                LotusDBf,
                how= 'left', left_on=family_column, right_on='organism_taxonomy_06family').drop_duplicates(subset=[id_column])
    df.drop('organism_taxonomy_06family', axis=1, inplace=True)
    df = df.fillna(0) #assumign species not present in LotusDB the number of reported compounds is set to 0

    if calculate_LC == True: 
        df['LC'] = 1-df['Reported_comp_Species'].div(max_comp_reported_sp*100)-df['Reported_comp_Genus'].div(max_comp_reported_g*100)-df['Reported_comp_Family'].div(max_comp_reported_f*100)
        df.to_csv('../LC_results.tsv', sep='\t')
    else:
        print('Literature component not calculated')
    
    df.to_csv('../LC_results.tsv', sep='\t')
    return df
        
df_LC = literature_component(metadata,max_comp_reported_sp, max_comp_reported_g, max_comp_reported_f)