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_filter_umi_threshold_error_2(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
with pytest.raises(ValueError,match=r"UMI threshold too high, all samples would be removed."):
fdata,fbc,fgenes= qc.filter_count_matrix(remove_cell_cycle=True,
UMI_thresh = 3000,Features_thresh = 39,
log10FeaturesPerUMI_thresh = 0.002,
FeaturesPerUMI_thresh = 0.0001,mtRatio_thresh = 0.5)
def test_filter_without_QC_Obj(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata, fgenes,fbc = qc.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 isinstance(fdata,spsp.csc.csc_matrix)
assert np.shape(fdata) == (32,40)
assert fdata.sum() == 63358
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_filter_allFilters_compare(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
d0,b,a = qc.filter_count_matrix(UMI_thresh=1700)
qc=QualityControl(d0,b,a)
d1,b,a = qc.filter_count_matrix(Features_thresh=39)
qc=QualityControl(d1,b,a)
d2,b,a = qc.filter_count_matrix(FeaturesPerUMI_thresh=.02)
qc=QualityControl(d2,b,a)
d3,b,a = qc.filter_count_matrix(log10FeaturesPerUMI_thresh=.48)
qc=QualityControl(d3,b,a)
d4,fgenes4,fbc4 = qc.filter_count_matrix(mtRatio_thresh=.1)
qc=QualityControl(d4,fgenes4,fbc4)
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
d_all,fgenes_all,fbc_all = qc.filter_count_matrix(UMI_thresh=1700,Features_thresh=39,FeaturesPerUMI_thresh=.02,
log10FeaturesPerUMI_thresh=.48,mtRatio_thresh=.1)
assert (d0.shape[0],d1.shape[0],d2.shape[0],d3.shape[0],d4.shape[0],d_all.shape[0]) == (45, 30, 17, 17, 14, 14)
assert np.all(fgenes4 == fgenes_all)
assert np.all(fbc4 == fbc_all)
assert np.sum(d4) == np.sum(d_all)
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_filter_umi_threshold_error(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata,fgenes,fbc = qc.filter_count_matrix(UMI_thresh=900,remove_cell_cycle=True,verbose=True)
def test_filter_mtRatio_threshold_error(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
with pytest.raises(ValueError,match=r"MT ratio threshold too low, all samples would be removed."):
fdata,fgenes,fbc = qc.filter_count_matrix(mtRatio_thresh=.000001)
def test_filter_log10featuresPerUMI_threshold_error(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
with pytest.raises(ValueError,match=r"log10 Feature per UMI threshold too high, all samples would be removed."):
fdata,fgenes,fbc = qc.filter_count_matrix(log10FeaturesPerUMI_thresh=.9)
def test_filter_mtRatio_threshold_wrong_type(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
with pytest.raises(ValueError,match=r"mtRatio threshold must be a float between 0 and 1."):
fdata,fgenes, fbc = qc.filter_count_matrix(mtRatio_thresh=-1)
def test_filter_log10featuresPerUMI_threshold_wrong_type(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
with pytest.raises(ValueError,match=r"log10featurePerUMI threshold must be an integer or float."):
fdata,fgenes,fbc = qc.filter_count_matrix(log10FeaturesPerUMI_thresh=list([1,2,3]))
def test_filter_allFilters(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata,fgenes,fbc = qc.filter_count_matrix(UMI_thresh=1700,Features_thresh=39,FeaturesPerUMI_thresh=.02,
log10FeaturesPerUMI_thresh=.48,mtRatio_thresh=.1)
assert fdata.shape == (14, 40)
def test_filter_mtRatio(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata,fgenes,fbc = qc.filter_count_matrix(mtRatio_thresh=.1)
assert fdata.shape == (43,40)
def test_filter_log10featuresPerUMI(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata,fgenes,fbc = qc.filter_count_matrix(log10FeaturesPerUMI_thresh=.48)
assert fdata.shape == (43, 40)
def test_filter_features(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata,fgenes,fbc = qc.filter_count_matrix(Features_thresh=39)
assert fdata.shape == (32, 40)
def test_filter_umi(self):
qc=QualityControl(self.mtx_df_50x40,self.genes,self.barcodes)
fdata,fgenes,fbc = qc.filter_count_matrix(UMI_thresh=1700)
assert fdata.shape == (45, 40)