def test_metrics_features_threshold_error(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
with pytest.raises(ValueError,match=r"Feature threshold too high, all samples would be removed."):
fdata,fgenes,fbc, QC_metaobj_50x40 = qc.metrics(filter_count_matrix=True,remove_cell_cycle=True,
UMI_thresh = 1500,Features_thresh = 50,
log10FeaturesPerUMI_thresh = 0.002,
FeaturesPerUMI_thresh = 0.0001,mtRatio_thresh = 0.5)
def test_metrics_with_filter_DATAFRAME(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata,fgenes,fbc, QC_metaobj_50x40 = qc.metrics(UMI_thresh = 1500,Features_thresh = 39,log10FeaturesPerUMI_thresh = 0.002,
FeaturesPerUMI_thresh = 0.0001,mtRatio_thresh = 0.5,
filter_count_matrix=True, remove_cell_cycle=False)
assert QC_metaobj_50x40.shape == (len(self.mtx_df_50x40),6)
assert np.all(QC_metaobj_50x40.columns == ['nUMI', 'nFeatures', 'FeaturesPerUMI','log10FeaturesPerUMI', 'mtUMI','mitoRatio'])
assert np.any(QC_metaobj_50x40.isna()) == False
assert isinstance(fdata,spsp.csc.csc_matrix)
assert isinstance(fbc,list)
assert isinstance(fgenes,list)
assert len(fbc) == 32
assert len(fgenes) == 40
assert fdata.shape == (32,40)
assert fdata.sum() == 63358
assert len(set(fbc).difference(set(['AAACATTGAGCTAC-1','AAACATACAACCAC-1','AAACATTGATCAGC-1','AAACCGTGTATGCG-1','AAACGCACTGGTAC-1',
'AAACGCTGACCAGT-1','AAACTTGATCCAGA-1', 'AAAGAGACGCGAGA-1', 'AAAGAGACGGCATT-1','AAAGCAGAAGCCAT-1', 'AAAGCAGATATCGG-1','AAAGCCTGTATGCG-1',
'AAAGTTTGATCACG-1', 'AAAGTTTGGGGTGA-1', 'AAAGTTTGTAGAGA-1','AAAGTTTGTAGCGT-1', 'AAATCAACCCTATT-1', 'AAATCAACGGAAGC-1','AAATCAACTCGCAA-1',
'AAATCCCTCCACAA-1', 'AAATCCCTGCTATG-1','AAATGTTGAACGAA-1', 'AAATGTTGCCACAA-1','AAATGTTGTGGCAT-1','AAATTCGAAGGTTC-1','AAATTCGAGCTGAT-1','AAACATTGATCTAC-1','AAACATTGATCCGC-1',
'ATACGCACTGGTAC-1', 'AAACGCTGATCAGT-1','AAACGCGGGTTCTT-1','AAAAGCTGTAGCCA-1']))) == 0
assert len(set(fgenes).difference(set(['ENSG00000243485',
'ENSG00000237613','ENSG00000186092','ENSG00000238009','ENSG00000239945','ENSG00000237683', 'ENSG00000239906',
'ENSG00000241599','ENSG00000228463','ENSG00000237094','ENSG00000235249', 'ENSG00000236601','ENSG00000236743', 'ENSG00000231709',
'ENSG00000239664', 'ENSG00000230021', 'ENSG00000223659','ENSG00000185097','ENSG00000235373', 'ENSG00000240618','ENSG00000229905',
'ENSG00000010292','ENSG00000011426','ENSG00000129055', 'ENSG00000177757', 'ENSG00000225880', 'ENSG00000230368','ENSG00000269308',
'ENSG00000272438','ENSG00000230699','ENSG00000210049','ENSG00000211459','ENSG00000097007', 'ENSG00000210082',
'ENSG00000241180', 'ENSG00000223764', 'ENSG00000187634','ENSG00000268179',
'ENSG00000188976','ENSG00000187961']))) == 0
def test_log10FeaturesPerUMI_thresh(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata1,fgenes1,fbc1 = qc.filter_count_matrix(log10FeaturesPerUMI_thresh=.48)
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata2,fgenes2,fbc2, QC_metaobj_50x40 = qc.metrics(filter_count_matrix=True,log10FeaturesPerUMI_thresh=.48)
assert fdata1.shape == fdata2.shape
def test_metrics_no_filter_CSC(self):
qc=QualityControl(self.csc_50x40,self.genes,self.barcodes)
QC_metaobj_50x40_csc = qc.metrics(filter_count_matrix=False) # UMI_thresh = 1500,Features_thresh = 39,log10FeaturesPerUMI_thresh = 0.002,FeaturesPerUMI_thresh = 0.0001,mtRatio_thresh = 0.5,
assert QC_metaobj_50x40_csc.shape == (self.csc_50x40.shape[0],6)
assert np.all(QC_metaobj_50x40_csc.columns == ['nUMI', 'nFeatures', 'FeaturesPerUMI','log10FeaturesPerUMI', 'mtUMI','mitoRatio'])
assert np.any(QC_metaobj_50x40_csc.isna()) == False
np.testing.assert_approx_equal(sum(sum(QC_metaobj_50x40_csc.values)), 107641.13463,significant=4, err_msg='metrics sum incorrect')
def test_metrics_no_filter_DATAFRAME(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
QC_metaobj_50x40 = qc.metrics(UMI_thresh = 1500,Features_thresh = 39,log10FeaturesPerUMI_thresh = 0.002,
FeaturesPerUMI_thresh = 0.0001,mtRatio_thresh = 0.5,filter_count_matrix=False)
assert QC_metaobj_50x40.shape == (len(self.mtx_df_50x40),6)
assert np.all(QC_metaobj_50x40.columns == ['nUMI', 'nFeatures', 'FeaturesPerUMI','log10FeaturesPerUMI', 'mtUMI','mitoRatio'])
assert np.any(QC_metaobj_50x40.isna()) == False
assert sum(sum(QC_metaobj_50x40.values)) == 107641.1346368048
np.testing.assert_approx_equal(sum(sum(QC_metaobj_50x40.values)), 107641.13463,significant=4, err_msg='metrics sum incorrect')
def test_filter_both(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
QC_metaobj_50x40 = qc.metrics(filter_count_matrix=False)
fdata1,genes1,bc1 = qc.filter_count_matrix(QC_metaobj_50x40,remove_cell_cycle=False,UMI_thresh = 1500,Features_thresh = 39,log10FeaturesPerUMI_thresh = 0.002,FeaturesPerUMI_thresh= 0.0001,mtRatio_thresh = 0.5)
qc2=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata2, genes2, bc2 = qc2.filter_count_matrix(QC_metaobj=None,remove_cell_cycle=False,UMI_thresh = 1500,Features_thresh = 39,log10FeaturesPerUMI_thresh = 0.002,FeaturesPerUMI_thresh= 0.0001,mtRatio_thresh = 0.5)
assert fdata1.shape == fdata2.shape
assert np.all(bc1 == bc2)
assert np.all(genes1 == genes2)
def test_metrics_with_filter_df_verbose(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata,fgenes,fbc, QC_metaobj_50x40 = qc.metrics(UMI_thresh = 1500,
Features_thresh = 39,
log10FeaturesPerUMI_thresh = 0.002,
FeaturesPerUMI_thresh = 0.0001,
mtRatio_thresh = 0.5,
filter_count_matrix=True,
remove_cell_cycle=False,
verbose=True)
def test_filter_with_QC_Obj(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
QC_metaobj_50x40 = qc.metrics(filter_count_matrix=False)
fdata,fgenes,fbc = qc.filter_count_matrix(QC_metaobj_50x40,remove_cell_cycle=False,UMI_thresh = 1500,Features_thresh = 39,log10FeaturesPerUMI_thresh = 0.002,
FeaturesPerUMI_thresh = 0.0001,mtRatio_thresh = 0.5)
assert isinstance(fdata,spsp.csc.csc_matrix)
assert isinstance(fbc,list)
assert isinstance(fgenes,list)
assert np.shape(fdata) == (32,40)
assert fdata.sum() == 63358
def test_filter_with_QC_Obj_removeCC(self):
'''test QC filter function with QC_metaobj (call metrics to get QC_metaobj
and pass it to filter_count_matrix with mtx_df)'''
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
QC_metaobj_50x40 = qc.metrics(UMI_thresh = 1500,Features_thresh = 39,log10FeaturesPerUMI_thresh = 0.002,
FeaturesPerUMI_thresh = 0.0001,mtRatio_thresh = 0.5,filter_count_matrix=False)
fdata, fgenes, fbc = qc.filter_count_matrix(QC_metaobj_50x40,remove_cell_cycle=True,UMI_thresh = 1500,Features_thresh = 39,log10FeaturesPerUMI_thresh = 0.002,
FeaturesPerUMI_thresh = 0.0001,mtRatio_thresh = 0.5)#,nUMI=500,nFeatures=500,FeaturesPerUMI=0.3,mtRatio=0.05)
assert isinstance(fdata,spsp.csc.csc_matrix)
assert isinstance(fbc,list)
assert isinstance(fgenes,list)
assert np.shape(fdata) == (32,38)
assert fdata.sum() == 60257
def test_metrics_filter_no_args(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
with pytest.raises(ValueError):
fdata,fgenes,fbc,QC_metaobj_50x40 = qc.metrics(filter_count_matrix=True)