Find cgMLST alleles unique to a group/cluster

dummy_data.csv

unique_alleles.py

'''
Find unique alleles to a group
'''

import click
import pandas as pd

def load_data(filename):
    '''
    Load CSV in to a pandas.DataFrame
    '''
    return pd.read_csv(filename, engine="python", sep=None)

def find_unique(tab):
    '''
    1. One-hot encode the data.
    2. Groub By the "GROUP" variable --- this assumes that you have only two groups an IN and an OUT, or an A and B groups
    3. Find the proportion of individuals in each group that have a particular allele in each of the two groups
    4. If group of interest is A, identify those columns that have proportion 1.0 for the allele in A, and 0.0 in group B.
    5. Return a Boolean vector 
    '''
    tab_one = pd.get_dummies(tab, columns=list(tab.columns[tab.columns.str.startswith("LOCUS")]))
    return tab_one.groupby("GROUP").apply(lambda x: x.sum()/x.count()).apply(lambda x: True if x.A == 1.0 and x.B == 0.0 else False)
 

@click.command()
@click.argument("input_file")
def run_finder(input_file):
    tab = load_data(input_file)
    unq = find_unique(tab)
    print(unq)
    
if __name__ == "__main__":
    run_finder()

When running this script with the dummy data, you get:

LOCUS1 _1    False
LOCUS1 _2    False
LOCUS2_1      True
LOCUS2_3     False
LOCUS2_4     False
LOCUS3_1     False
LOCUS3_2     False
dtype: bool