comparison lrn_risk.py @ 1:f98c92618a6c draft

Uploaded

0:99e04eba4033

def get_species_from_gtdb(f):
    # get GTDB species
    # assumes there is one genome in the GTDB-Tk output file
    with open(f, 'r') as fh:
        for line in fh:
            if line.find('user_genome') < 0:
                items = line.split('\t')
                tax = items[1].strip()
                tax = tax.split(';')[-1].strip()
                # split on GTDB species tag
                tax = tax.split('s__')[1].strip()
                if len(tax) == 0:
                    tax = '(Unknown Species)'
    return tax


def get_blast_genes(f):
    # reads genes detected via BLAST

1:f98c92618a6c

def get_species_from_gtdb(f):
    # get GTDB species
    # assumes there is one genome in the GTDB-Tk output file
    with open(f, 'r') as fh:
        for i, line in enumerate(fh):
            if i == 0:
                # Skip header.
                continue
            items = line.split('\t')
            tax = items[1].strip()
            tax = tax.split(';')[-1].strip()
            # split on GTDB species tag
            tax = tax.split('s__')[1].strip()
            if len(tax) == 0:
                tax = '(Unknown Species)'
    return tax


def get_blast_genes(f):
    # reads genes detected via BLAST