def test_pixel_expansion_ring_plots(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.pixel_expansion_ring_plots(n_expansions=1,
expansions=[1])
test_path = "%s/Ring Plots/%s%s Expansion/Point%s.png" % (temp_dir,
str(round_to_nearest_half(
1 *
config.pixel_interval *
config.pixels_to_distance)),
config.data_resolution_units,
str(1))
assert path.exists(test_path)
def test_removed_vessel_expression_boxplot(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir,
n_points=2,
resolution=(2048, 2048),
n_pseudo_vessel_size=(1, 50))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2,
brain_region_point_ranges=[(1, 2)],
brain_region_names=["Test Region"],
data_resolution_size=(5000, 5000)
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.removed_vessel_expression_boxplot()
test_path = "%s/Kept Vs. Removed Vessel Boxplots/Test/All Points/All_Points.png" % temp_dir
assert path.exists(test_path)
def test_vessel_images_by_categorical_variable(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.vessel_images_by_categorical_variable(primary_categorical_analysis_variable="Vessel Size")
test_path = "%s/Vessel Size Vessel Images/Test/Large/Point_Num_1_Vessel_ID_5.png" % temp_dir
assert path.exists(test_path)
def test_box_plot_brain_expansions(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.box_plot_brain_expansions(n_expansions=1)
test_path = "%s/Expansion Box Plots/Per Marker/0.5μm Expansion/SMA.png" % temp_dir
assert path.exists(test_path)
def test_spatial_probability_maps(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.spatial_probability_maps(mask_size=(2048, 2048))
test_path = "%s/Pixel Expression Spatial Maps/Test/Vessels/Point1.png" % temp_dir
assert path.exists(test_path)
def test_categorical_violin_plot(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.categorical_violin_plot(primary_categorical_analysis_variable="Vessel Size",
mask_type="mask_and_expansion")
test_path = "%s/Categorical Violin Plots/Test/By Vessel Size/Vessels/SMA.png" % temp_dir
assert path.exists(test_path)
def create_test_data(test_dir,
n_points=48,
resolution=(2048, 2048),
n_pseudo_vessel_size=(10, 50)):
config = Config()
mask_dir = os.path.join(test_dir, "masks")
data_dir = os.path.join(test_dir, "data")
mkdir_p(mask_dir)
mkdir_p(data_dir)
for point_idx in range(n_points):
directory_name = "Point" + str(point_idx + 1)
point_data_dir = os.path.join(data_dir, directory_name, "TIFs")
point_mask_dir = os.path.join(mask_dir, directory_name)
mkdir_p(point_data_dir)
mkdir_p(point_mask_dir)
for cluster in config.marker_clusters.keys():
for marker in config.marker_clusters[cluster]:
data = generate_random_data(resolution)
write_tif_data(data, marker, point_data_dir)
mask = generate_random_mask(resolution, n_pseudo_vessel_size=n_pseudo_vessel_size)
write_tif_data(mask, "allvessels", point_mask_dir)
def test_vessel_nonvessel_masks(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2,
brain_region_point_ranges=[(1, 2)],
brain_region_names=["Test Region"]
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.vessel_nonvessel_masks()
test_path = "%s/Associated Area Masks/Test/2.5μm Expansion/Point 1.png" % temp_dir
assert path.exists(test_path)
def test_dimensionality_reduction_clustering(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2)
pipe = MIBIPipeline(config,
temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
analyzer = DimensionalityReductionClusteringAnalyzer(
pipe.config, pipe.all_expansions_features, pipe.marker_names,
pipe.all_feeds_contour_data, pipe.all_feeds_metadata,
pipe.all_feeds_data)
analyzer.analyze(temp_dir)
self.assertIn("UMAP0", pipe.all_expansions_features.columns)
self.assertIn("UMAP1", pipe.all_expansions_features.columns)
self.assertIn("K-Means", pipe.all_expansions_features.columns)
self.assertIn("Hierarchical", pipe.all_expansions_features.columns)
def hires_example():
conf = Config()
results_dir = "/media/aswin/large_storage/results/experiment_%s/" % datetime.now().strftime("%d_%m_%Y_%H:%M:%S")
mkdir_p(results_dir)
hires_feed = MIBIDataFeed(
feed_data_loc="/media/aswin/large_storage/oliveria_data/data/hires",
feed_mask_loc="/media/aswin/large_storage/oliveria_data/masks/hires",
feed_name="Hires",
n_points=48,
brain_region_point_ranges=[(1, 16), (17, 32), (33, 48)],
brain_region_names=["MFG", "HIP", "CAUD"]
)
pipe = MIBIPipeline(conf, results_dir,
csv_loc="/media/aswin/large_storage/results/5um_impansion_5um_expansion.csv"
)
pipe.add_feed(hires_feed)
pipe.load_preprocess_data()
pipe.add_analyzer(ShapeQuantificationAnalyzer)
# pipe.add_analyzer(PositiveVesselSummaryAnalyzer)
# pipe.add_analyzer(DimensionalityReductionClusteringAnalyzer)
shape_quantification_method = {
"Name": "Solidity",
"Metric": solidity
}
pipe.analyze_data(marker_settings="all_markers",
shape_quantification_method=shape_quantification_method)
def test_brain_region_expansion_heatmap(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2,
brain_region_point_ranges=[(1, 2)],
brain_region_names=["Test Region"]
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.brain_region_expansion_heatmap(n_expansions=1,
primary_categorical_analysis_variable=None)
test_path = "%s/Expansion Heatmaps & Clustermaps/Test/Expansion Clustermaps/0.5μm Expansion/All_Points.png" % temp_dir
assert path.exists(test_path)
def test_load_preprocess_data(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2)
pipe = MIBIPipeline(config,
temp_dir,
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
actual_features = pd.read_csv("data/dummy_test_data.csv",
index_col=[0, 1, 2, 3],
skipinitialspace=True)
assert_frame_equal(actual_features, pipe.all_expansions_features)
self.assertIsNotNone(pipe.visualizer)
def test_positive_vessel_summary(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2)
pipe = MIBIPipeline(config,
temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
analyzer = PositiveVesselSummaryAnalyzer(
pipe.config, pipe.all_expansions_features, pipe.marker_names,
pipe.all_feeds_contour_data, pipe.all_feeds_metadata,
pipe.all_feeds_data)
analyzer.analyze(temp_dir)
self.assertTrue(
os.path.isfile(
os.path.join(temp_dir, "vessel_positive_proportion.csv")))
def test_obtain_embedded_vessel_masks(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.obtain_embedded_vessel_masks(n_points=2)
test_path = "%s/Embedded Vessel Masks/29.5 μm/Test/Original Mask Excluded/Point1.tif" % temp_dir
assert path.exists(test_path)
def test_scatter_plot_umap_marker_projection(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
umap_analyzer = DimensionalityReductionClusteringAnalyzer(
pipe.config,
pipe.all_expansions_features,
pipe.marker_names,
pipe.all_feeds_contour_data,
pipe.all_feeds_metadata,
pipe.all_feeds_data
)
umap_analyzer.analyze(temp_dir)
shape_quant_analyzer = ShapeQuantificationAnalyzer(
pipe.config,
pipe.all_expansions_features,
pipe.marker_names,
pipe.all_feeds_contour_data,
pipe.all_feeds_metadata,
pipe.all_feeds_data
)
shape_quant_analyzer.analyze(temp_dir,
mask_type="expansion_only",
marker_settings="all_markers",
shape_quantification_method={
"Name": "Solidity",
"Metric": solidity
},
img_shape=(2048, 2048))
pipe.visualizer.scatter_plot_umap_marker_projection(mask_type="mask_only",
primary_continuous_analysis_variable="Solidity Score")
test_path = "%s/UMAP Scatter Plot Projection/Test/Marker Clusters/Vessels.png" % temp_dir
assert path.exists(test_path)
def test_init(self):
conf = Config()
mask = generate_random_mask((2048, 2048, 1))
object_extractor = ObjectExtractor(conf, None)
mibi_contours = MIBIPointContours(mask, 1, conf, object_extractor)
self.assertIsNotNone(mibi_contours.contours)
self.assertIsNotNone(mibi_contours.removed_contours)
self.assertIsNotNone(mibi_contours.removed_areas)
def test_pseudo_time_heatmap(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2,
brain_region_point_ranges=[(1, 2)],
brain_region_names=["Test Region"]
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
analyzer = ShapeQuantificationAnalyzer(
pipe.config,
pipe.all_expansions_features,
pipe.marker_names,
pipe.all_feeds_contour_data,
pipe.all_feeds_metadata,
pipe.all_feeds_data
)
analyzer.analyze(temp_dir,
mask_type="expansion_only",
marker_settings="all_markers",
shape_quantification_method={
"Name": "Solidity",
"Metric": solidity
},
img_shape=(2048, 2048))
pipe.visualizer.pseudo_time_heatmap()
test_path1 = "%s/Pseudo-Time Heatmaps/Test/Test Region/" \
"Solidity Score_pseudo_time_heatmap_Test Region.png" % temp_dir
test_path2 = "%s/Pseudo-Time Heatmaps/Test/Test Region/" \
"Solidity Score_pseudo_time_heatmap_binned_Test Region.png" % temp_dir
test_path3 = "%s/Pseudo-Time Heatmaps/Test/Solidity Score_pseudo_time_heatmap_all_points.png" % temp_dir
assert path.exists(test_path1)
assert path.exists(test_path2)
assert path.exists(test_path3)
def test_add_feed(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=1, resolution=(2048, 2048))
config = Config()
loader = MIBILoader(config)
example_feed = MIBIDataFeed(feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=1)
loader.add_feed(example_feed)
self.assertEqual(loader.feeds[example_feed.name], example_feed)
def medres_example():
conf = Config()
medres_feed = MIBIDataFeed(
feed_data_loc="/media/large_storage/oliveria_data/data/medres",
feed_mask_loc="/media/large_storage/oliveria_data/masks/medres",
feed_name="Medres",
n_points=300,
points_per_dir=100,
brain_region_point_ranges=[(1, 100), (101, 200), (201, 300)]
)
loader = MIBILoader(conf)
loader.add_feed(medres_feed)
loader.read()
def test_categorical_violin_plot_with_images(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
analyzer = ShapeQuantificationAnalyzer(
pipe.config,
pipe.all_expansions_features,
pipe.marker_names,
pipe.all_feeds_contour_data,
pipe.all_feeds_metadata,
pipe.all_feeds_data
)
analyzer.analyze(temp_dir,
mask_type="expansion_only",
marker_settings="all_markers",
shape_quantification_method={
"Name": "Solidity",
"Metric": solidity
},
img_shape=(2048, 2048))
pipe.visualizer.categorical_violin_plot_with_images(primary_categorical_analysis_variable="Solidity",
mask_size=(2048, 2048),
order=["25%", "50%", "75%", "100%"])
test_path = "%s/Categorical Violin Plots with Images/Test/By Solidity/Vessels/SMA.png" % temp_dir
assert path.exists(test_path)
def test_line_plots(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(
feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2
)
pipe = MIBIPipeline(config, temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False
)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
pipe.visualizer.expansion_line_plots_per_vessel(n_expansions=1,
vessel_line_plots_points=[1])
assert path.exists("%s/%s%s Expansion" % ("%s/Line Plots Per Vessel" % temp_dir,
str(round_to_nearest_half(config.pixel_interval *
config.pixels_to_distance)),
config.data_resolution_units))
pipe.visualizer.expansion_line_plots_per_point(n_expansions=1,
n_points=2)
assert path.exists("%s/%s%s Expansion" % ("%s/Line Plots Per Point" % temp_dir,
str(round_to_nearest_half(1 *
config.pixel_interval *
config.pixels_to_distance)),
config.data_resolution_units))
pipe.visualizer.expansion_line_plots_all_points(n_expansions=1)
assert path.exists("%s/%s%s Expansion" % ("%s/Line Plots All Points Average" % temp_dir,
str(round_to_nearest_half(1 *
config.pixel_interval *
config.pixels_to_distance)),
config.data_resolution_units))
def test_read(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=1, resolution=(2048, 2048))
config = Config()
loader = MIBILoader(config)
example_feed = MIBIDataFeed(feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=1)
loader.add_feed(example_feed)
all_feeds_metadata, all_feeds_data, all_feeds_mask, marker_names = loader.read(
)
self.assertEqual(len(all_feeds_data), 1)
def test_vessel_asymmetry_analyzer(self):
with tempfile.TemporaryDirectory() as temp_dir:
create_test_data(temp_dir, n_points=2, resolution=(2048, 2048))
config = Config()
example_feed = MIBIDataFeed(feed_data_loc="%s/data" % temp_dir,
feed_mask_loc="%s/masks" % temp_dir,
feed_name="Test",
n_points=2)
pipe = MIBIPipeline(config,
temp_dir,
csv_loc="data/dummy_test_data.csv",
max_inward_expansions=1,
max_outward_expansions=1,
expansions=[1],
n_workers=1,
run_async=False)
pipe.add_feed(example_feed)
pipe.load_preprocess_data()
analyzer = ShapeQuantificationAnalyzer(
pipe.config, pipe.all_expansions_features, pipe.marker_names,
pipe.all_feeds_contour_data, pipe.all_feeds_metadata,
pipe.all_feeds_data)
analyzer.analyze(temp_dir,
mask_type="expansion_only",
marker_settings="all_markers",
shape_quantification_method={
"Name": "Solidity",
"Metric": solidity
},
img_shape=(2048, 2048))
self.assertIn("Solidity", pipe.all_expansions_features.columns)
def test_normalize_expression_data(self):
config = Config()
marker_names = [
"HH3",
"CD45",
"HLADR",
"Iba1",
"CD47",
"ABeta42",
"polyubiK48",
"PHFTau",
"8OHGuanosine",
"SMA",
"CD31",
"CollagenIV",
"TrkA",
"GLUT1",
"Desmin",
"vWF",
"CD105",
"S100b",
"GlnSyn",
"Cx30",
"EAAT2",
"CD44",
"GFAP",
"Cx43",
"CD56",
"Synaptophysin",
"VAMP2",
"PSD95",
"MOG",
"MAG",
"Calretinin",
"Parvalbumin",
"MAP2",
"Gephyrin",
]
test_raw_data = pd.read_csv("data/dummy_test_data_unnormalized.csv",
index_col=[0, 1, 2, 3],
skipinitialspace=True)
scaling_factor = config.scaling_factor
transformation = config.transformation_type
normalization = config.normalization_type
n_markers = config.n_markers
hh3_data = test_raw_data["HH3"].to_numpy()
hh3_data = hh3_data * scaling_factor
hh3_data = arcsinh(hh3_data)
percentile_99 = np.percentile(hh3_data, 99, axis=0)
normalized_hh3 = hh3_data / percentile_99
all_expansions_features = normalize_expression_data(
config,
test_raw_data,
marker_names,
transformation=transformation,
normalization=normalization,
scaling_factor=scaling_factor,
n_markers=n_markers)
test_normalized_hh3 = all_expansions_features["HH3"].to_numpy()
assert_array_equal(normalized_hh3, test_normalized_hh3)