def test_generic_pandas_mapper():
"""
Test that generic pandas mapping selects and renames columns
"""
df = pd.DataFrame({"a": [1, 2], "b": [3, 4]})
mapper = OrderedDict([('a', 'ardvark')])
r = mappers.generic_pandas_mapper(df, mapper)
expected = pd.DataFrame({"ardvark": [1, 2]})
assert r.equals(expected)
def test_generic_pandas_mapper_raises_assertion_error():
"""
Test that a KeyError is raised if the mapper contains keys not in df.columns
"""
df = pd.DataFrame({"a": [1, 2], "b": [3, 4]})
mapper = OrderedDict([('c', 'canary')])
with pytest.raises(KeyError):
r = mappers.generic_pandas_mapper(df, mapper)
expected = pd.DataFrame({"ardvark": [1, 2]})
def test_data_type_conversion_with_reduce_file_size():
"""
Test that successfull conversion of np attributes from
np.float64 to np.int16
after calling
"""
import pandas as pd
import numpy as np
import tcrdist as td
from tcrdist import mappers
from tcrdist.repertoire import TCRrep
tcrdist_clone_fn = 'tcrdist/test_files/mouse_pairseqs_v1_parsed_seqs_probs_mq20_clones.tsv'
tcrdist_clone_df = pd.read_csv(tcrdist_clone_fn, sep = "\t") #1
ind = (tcrdist_clone_df.epitope == "PA") | (tcrdist_clone_df.epitope == "F2")
tcrdist_clone_df = tcrdist_clone_df[ind].copy()
mapping = mappers.tcrdist_clone_df_to_tcrdist2_mapping #3
tcrdist2_df = mappers.generic_pandas_mapper(df = tcrdist_clone_df, #4
mapping = mapping)
#1
tr = TCRrep(cell_df = tcrdist2_df, organism = "mouse")
#2
tr.infer_cdrs_from_v_gene(chain = 'alpha', imgt_aligned=True)
tr.infer_cdrs_from_v_gene(chain = 'beta', imgt_aligned=True)
#3
tr.index_cols = ['clone_id', 'subject', 'epitope',
'v_a_gene', 'j_a_gene', 'v_b_gene', 'j_b_gene',
'cdr3_a_aa', 'cdr3_b_aa',
'cdr1_a_aa', 'cdr2_a_aa', 'pmhc_a_aa',
'cdr1_b_aa', 'cdr2_b_aa', 'pmhc_b_aa',
'cdr3_b_nucseq', 'cdr3_a_nucseq',
'va_countreps', 'ja_countreps',
'vb_countreps', 'jb_countreps',
'va_gene', 'vb_gene',
'ja_gene', 'jb_gene']
#4
tr.deduplicate()
#5
tr._tcrdist_legacy_method_alpha_beta()
#print(type(tr.cdr3_a_aa_pw[1,1]))
assert isinstance(tr.cdr3_a_aa_pw[1,1], np.int)
assert isinstance(tr.cdr3_b_aa_pw[1,1], np.int)
tr.reduce_file_size()
assert isinstance(tr.cdr3_a_aa_pw[1,1], np.int16)
assert isinstance(tr.cdr3_b_aa_pw[1,1], np.int16)
def generate_tr():
import pandas as pd
import numpy as np
import tcrdist as td
from tcrdist import mappers
from tcrdist.repertoire import TCRrep
tcrdist_clone_fn = 'tcrdist/test_files/mouse_pairseqs_v1_parsed_seqs_probs_mq20_clones.tsv'
tcrdist_clone_df = pd.read_csv(tcrdist_clone_fn, sep = "\t") #1
ind = (tcrdist_clone_df.epitope == "PA") | (tcrdist_clone_df.epitope == "F2")
tcrdist_clone_df = tcrdist_clone_df[ind].copy()
mapping = mappers.tcrdist_clone_df_to_tcrdist2_mapping #3
tcrdist2_df = mappers.generic_pandas_mapper(df = tcrdist_clone_df, #4
mapping = mapping)
#1
tr = TCRrep(cell_df = tcrdist2_df, organism = "mouse")
#2
tr.infer_cdrs_from_v_gene(chain = 'alpha', imgt_aligned=True)
tr.infer_cdrs_from_v_gene(chain = 'beta', imgt_aligned=True)
#3
tr.index_cols = ['clone_id', 'subject', 'epitope',
'v_a_gene', 'j_a_gene', 'v_b_gene', 'j_b_gene',
'cdr3_a_aa', 'cdr3_b_aa',
'cdr1_a_aa', 'cdr2_a_aa', 'pmhc_a_aa',
'cdr1_b_aa', 'cdr2_b_aa', 'pmhc_b_aa',
'cdr3_b_nucseq', 'cdr3_a_nucseq',
'va_countreps', 'ja_countreps',
'vb_countreps', 'jb_countreps',
'va_gene', 'vb_gene',
'ja_gene', 'jb_gene']
#4
tr.deduplicate()
#5
tr._tcrdist_legacy_method_alpha_beta()
return tr
def test_CompleteExample_with_TCRMotif_Invoked_From_within_TCRsubset():
import pandas as pd
import numpy as np
import tcrdist as td
#import IPython
from tcrdist import mappers
from tcrdist.repertoire import TCRrep
from tcrdist.cdr3_motif import TCRMotif
from tcrdist.subset import TCRsubset
from tcrdist.storage import StoreIOMotif, StoreIOEntropy
from tcrdist.plotting import plot_pwm
tcrdist_clone_fn = 'tcrdist/test_files/mouse_pairseqs_v1_parsed_seqs_probs_mq20_clones.tsv'
tcrdist_clone_df = pd.read_csv(tcrdist_clone_fn, sep="\t") #1
ind = (tcrdist_clone_df.epitope == "PA") | (tcrdist_clone_df.epitope
== "F2")
tcrdist_clone_df = tcrdist_clone_df[ind].copy()
mapping = mappers.tcrdist_clone_df_to_tcrdist2_mapping #3
tcrdist2_df = mappers.generic_pandas_mapper(
df=tcrdist_clone_df, #4
mapping=mapping)
#1
tr = TCRrep(cell_df=tcrdist2_df, organism="mouse")
#2
tr.infer_cdrs_from_v_gene(chain='alpha', imgt_aligned=True)
tr.infer_cdrs_from_v_gene(chain='beta', imgt_aligned=True)
#3
tr.index_cols = [
'clone_id', 'subject', 'epitope', 'v_a_gene', 'j_a_gene', 'v_b_gene',
'j_b_gene', 'cdr3_a_aa', 'cdr3_b_aa', 'cdr1_a_aa', 'cdr2_a_aa',
'pmhc_a_aa', 'cdr1_b_aa', 'cdr2_b_aa', 'pmhc_b_aa', 'cdr3_b_nucseq',
'cdr3_a_nucseq', 'va_countreps', 'ja_countreps', 'vb_countreps',
'jb_countreps', 'va_gene', 'vb_gene', 'ja_gene', 'jb_gene'
]
#4
tr.deduplicate()
#5
tr._tcrdist_legacy_method_alpha_beta()
#6
distA = tr.dist_a
distB = tr.dist_b
assert np.all(((distA + distB) - tr.paired_tcrdist) == 0)
# 1
criteria = tr.clone_df.epitope == "PA"
clone_df_subset = tr.clone_df[criteria]
# 2
distA_subset = distA.loc[clone_df_subset.clone_id,
clone_df_subset.clone_id].copy()
distB_subset = distB.loc[clone_df_subset.clone_id,
clone_df_subset.clone_id].copy()
# 3
ts = TCRsubset(clone_df_subset,
organism="mouse",
epitopes=["PA"],
epitope="PA",
chains=["A", "B"],
dist_a=distA_subset,
dist_b=distB_subset)
# ts.find_motif()
cnames = [
"file_type", "count", "expect_random", "expect_nextgen", "chi_squared",
"nfixed", "showmotif", "num", "othernum", "overlap", "ep", "ab",
"nseqs", "v_rep_counts", "j_rep_counts"
]
motif_fn = 'tcrdist/test_files/mouse_pairseqs_v1_parsed_seqs_probs_mq20_clones_cdr3_motifs_PA.log'
x = open(motif_fn, "r").readlines()
ts.motif_df = pd.DataFrame([l.split() for l in x], columns=cnames)
i = 0
row = ts.motif_df.iloc[i, :].to_dict()
motif_list = list()
motif_logo = list()
for i, row in ts.motif_df.iterrows():
StoreIOMotif_instance = ts.eval_motif(row)
motif_list.append(StoreIOMotif_instance)
motif_logo.append(
plot_pwm(StoreIOMotif_instance,
create_file=False,
my_height=200,
my_width=600))
if i > 1:
break
def test_Complete_Performance_Example():
import pandas as pd
import numpy as np
import tcrdist as td
from collections import namedtuple
from tcrdist import mappers
from tcrdist.repertoire import TCRrep
tcrdist_clone_fn = 'tcrdist/test_files/mouse_pairseqs_v1_parsed_seqs_probs_mq20_clones.tsv'
tcrdist_clone_df = pd.read_csv(tcrdist_clone_fn, sep="\t")
mapping = mappers.tcrdist_clone_df_to_tcrdist2_mapping
tcrdist2_df = mappers.generic_pandas_mapper(df=tcrdist_clone_df,
mapping=mapping)
tr = TCRrep(cell_df=tcrdist2_df, organism="mouse")
tr.infer_cdrs_from_v_gene(chain='alpha', imgt_aligned=True)
tr.infer_cdrs_from_v_gene(chain='beta', imgt_aligned=True)
tr.index_cols = [
'clone_id', 'subject', 'epitope', 'v_a_gene', 'j_a_gene', 'v_b_gene',
'j_b_gene', 'cdr3_a_aa', 'cdr3_b_aa', 'cdr1_a_aa', 'cdr2_a_aa',
'pmhc_a_aa', 'cdr1_b_aa', 'cdr2_b_aa', 'pmhc_b_aa', 'cdr3_b_nucseq',
'cdr3_a_nucseq', 'va_countreps', 'ja_countreps', 'vb_countreps',
'jb_countreps', 'va_gene', 'vb_gene', 'ja_gene', 'jb_gene'
]
tr.deduplicate()
tr._tcrdist_legacy_method_alpha_beta(processes=1)
distA = tr.dist_a
distB = tr.dist_b
assert np.all(((distA + distB) - tr.paired_tcrdist) == 0)
# K NEAREST NEIGHBORS
pr = namedtuple("perf", ["observed", "predicted", "dist"])
obsereved = tr.clone_df.epitope.to_list()
performance = list()
k = 5
for i, row in tr.clone_df.iterrows():
ind = (tr.clone_df.subject != row.subject
) # Index hold out all data from that subject
distances = tr.paired_tcrdist[
i, ind] # Get Distances from the ith row, holding out subject
sorted_indices = np.argsort(
distances) # Get index of storted distances small to large
sorted_epitopes = tr.clone_df.epitope.iloc[sorted_indices].to_list(
) # Get epitopes associated wtih those indices
sorted_distances = distances[
sorted_indices] # Get distances associated with those neighbors
predicted = sorted_epitopes[
0:k] # Get Predicted epitopes for K nearest neighbors
predicted_distance = sorted_distances[
0:k] # Get distances for K nearest neighbots
performance.append(
pr(obsereved[i], predicted,
predicted_distance)) # Save Performance Information
performance[1:10]
