def test_get_bets_j_gene():
from tcrdist.pairwise import hm_metric
tc = TCRcodon(organism="human", db_file="gammadelta_db.tsv")
df = pd.read_csv("tcrdist/test_files_compact/sant.csv")
someseq = df['cdr3_d_aa'][1]
x = tc.get_best_j_gene(aa_seq=someseq, verbose=True)
assert x == 'TRDJ1*01'
# test for all
xx = [tc.get_best_j_gene(aa_seq=x, verbose=False) for x in df['cdr3_d_aa']]
vc = pd.Series(xx).value_counts().to_dict()
assert vc == {
'TRDJ1*01': 271,
'TRDJ3*01': 66,
'TRDJ2*01': 20,
'TRDJ4*01': 9
}
def test_get_bets_j_gene_room_for_improvement():
tc = TCRcodon(organism="mouse", db_file="alphabeta_db.tsv")
df = clone_df_subset[[
'v_b_gene', 'j_b_gene', 'cdr3_b_aa', 'cdr3_b_nucseq'
]]
xx = [tc.get_best_j_gene(aa_seq=x, verbose=False) for x in df['cdr3_b_aa']]
# THIS CAN'T RESOLVE TIES, OR WHO KNOW WHAT ELSE
assert np.all(df['j_b_gene'][0:2] == xx[0:2])
assert np.sum(df['j_b_gene'] == xx) > 20
def test_TCRcodon_smal_dataframe_gama_lots():
tc = TCRcodon(organism="human", db_file="gammadelta_db.tsv")
df = pd.read_csv("tcrdist/test_files_compact/sant.csv")
df['j_d_gene'] = [
tc.get_best_j_gene(aa_seq=x, verbose=False) for x in df['cdr3_d_aa']
]
df = df[df['v_d_gene'].notna()].copy()
syn_nucs = df.apply(lambda r: \
tc.guess_reverse_translation(\
r['v_d_gene'], r['j_d_gene'], r['cdr3_d_aa'],\
verbose = False), axis = 1)
# Check that synthestic and real seqs are same length
len_syn = [len(x) for x in syn_nucs]
len_real = [3 * len(x) for x in df['cdr3_d_aa']]
assert np.all(len_syn == len_real)
assert np.all(len_syn == len_real)