Merge pull request #89 from BodenmillerGroup/sort-analysis-stack-chan…

…nels-by-mass

sort analysis stack channels by mass (closes #88)

CHANGELOG.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## [3.2]
+
+ - sort channels by metal tag when creating the ilastik and full stacks
+
 ## [3.1]
 
  - fixed git submodule issue

scripts/download_examples.py

@@ -43,7 +43,7 @@
 # Ilastik project
 ilastik_project = Path("..") / "IMCWorkflow.ilp"
 if not ilastik_project.exists():
-    request.urlretrieve("https://zenodo.org/record/6043544/files/IMCWorkflow.ilp", ilastik_project)
+    request.urlretrieve("https://zenodo.org/record/6449127/files/IMCWorkflow.ilp", ilastik_project)
 
 # Sample metadata
 sample_metadata = Path("..") / "sample_metadata.xlsx"
@@ -52,3 +52,5 @@
 
 # %%
 # !conda list
+
+# %%

scripts/imc_preprocessing.py

@@ -6,14 +6,16 @@
 # !{sys.executable} -m pip install -e {Path.cwd().parent}
 
 # %%
-import imcsegpipe
-import pandas as pd
 import shutil
-
 from pathlib import Path
 from tempfile import TemporaryDirectory
 from typing import List
 
+import pandas as pd
+
+import imcsegpipe
+from imcsegpipe.utils import sort_channels_by_mass
+
 # %% [markdown]
 #
 # # Preprocessing of IMC data for image segmentation
@@ -197,15 +199,19 @@
         imcsegpipe.create_analysis_stacks(
             acquisition_dir=acquisition_dir,
             analysis_dir=final_images_dir,
-            analysis_channels=panel.loc[panel[panel_keep_col] == 1, panel_channel_col].tolist(),
+            analysis_channels=sort_channels_by_mass(
+                panel.loc[panel[panel_keep_col] == 1, panel_channel_col].tolist()
+            ),
             suffix="_full",
             hpf=50.0,
         )
         # Write ilastik stack
         imcsegpipe.create_analysis_stacks(
             acquisition_dir=acquisition_dir,
             analysis_dir=ilastik_dir,
-            analysis_channels=panel.loc[panel[panel_ilastik_col] == 1, panel_channel_col].tolist(),
+            analysis_channels=sort_channels_by_mass(
+                panel.loc[panel[panel_ilastik_col] == 1, panel_channel_col].tolist()
+            ),
             suffix="_ilastik",
             hpf=50.0,
         )
@@ -244,3 +250,5 @@
 
 # %%
 # !conda list
+
+# %%

src/imcsegpipe/__init__.py

@@ -1,16 +1,15 @@
 from ._imcsegpipe import (
-    extract_zip_file,
-    match_txt_files,
-    extract_mcd_file,
     create_analysis_stacks,
     export_to_histocat,
+    extract_mcd_file,
+    extract_zip_file,
+    match_txt_files,
 )
 
-
 __all__ = [
-    "extract_zip_file",
-    "match_txt_files",
-    "extract_mcd_file",
     "create_analysis_stacks",
     "export_to_histocat",
+    "extract_mcd_file",
+    "extract_zip_file",
+    "match_txt_files",
 ]

src/imcsegpipe/_imcsegpipe.py

@@ -1,19 +1,19 @@
-import imageio
 import logging
-import numpy as np
-import pandas as pd
 import re
 import shutil
-import tifffile
-import xtiff
-
 from os import PathLike
 from pathlib import Path
-from readimc import MCDFile, TXTFile
-from readimc.data import Acquisition, Panorama, Slide
 from typing import Dict, List, Optional, Sequence, Union
 from zipfile import ZipFile
 
+import imageio
+import numpy as np
+import pandas as pd
+import tifffile
+import xtiff
+from readimc import MCDFile, TXTFile
+from readimc.data import Acquisition, Panorama, Slide
+
 from .utils import AcquisitionMetadata, filter_hot_pixels, get_acquisition_ome_xml
 
 

src/imcsegpipe/utils.py

@@ -1,12 +1,13 @@
+import re
+from dataclasses import dataclass
+from typing import List, Optional, Sequence
+from xml.etree import ElementTree as ET
+
 import numpy as np
 import xtiff
-
-from dataclasses import dataclass
 from readimc import MCDFile
 from readimc.data import Acquisition
 from scipy.ndimage import maximum_filter
-from typing import Optional, Sequence
-from xml.etree import ElementTree as ET
 
 
 @dataclass
@@ -156,3 +157,7 @@ def filter_hot_pixels(img: np.ndarray, thres: float) -> np.ndarray:
     kernel[0, 1, 1] = False
     max_neighbor_img = maximum_filter(img, footprint=kernel, mode="mirror")
     return np.where(img - max_neighbor_img > thres, max_neighbor_img, img)
+
+
+def sort_channels_by_mass(channels: Sequence[str]) -> List[str]:
+    return sorted(channels, key=lambda channel: int(re.sub("[^0-9]", "", channel) or 0))