| Previous changeset 0:928db0f952e3 (2021-03-12) Next changeset 2:46b897eb2c8e (2022-03-30) |
|
Commit message:
"planemo upload for repository https://github.com/ohsu-comp-bio/quantification commit 897a7dc7cb43e45d6f0fdfe2b2970e59f20f8853" |
|
modified:
ParseInput.py SingleCellDataExtraction.py macros.xml quantification.xml |
| b |
| diff -r 928db0f952e3 -r aba3655fdef0 ParseInput.py --- a/ParseInput.py Fri Mar 12 00:19:24 2021 +0000 +++ b/ParseInput.py Fri Mar 11 23:35:52 2022 +0000 |
| [ |
| @@ -8,15 +8,39 @@ #if __name__ == '__main__': parser = argparse.ArgumentParser() - parser.add_argument('--masks',nargs='*') - parser.add_argument('--image') - parser.add_argument('--channel_names') - parser.add_argument('--output') + parser.add_argument('--masks',nargs='+', required=True) + parser.add_argument('--image', required=True) + parser.add_argument('--channel_names', required=True) + parser.add_argument('--output', required=True) + parser.add_argument( + '--mask_props', nargs = "+", + help=""" + Space separated list of additional metrics to be calculated for every mask. + This is for metrics that depend only on the cell mask. If the metric depends + on signal intensity, use --intensity-props instead. + See list at https://scikit-image.org/docs/dev/api/skimage.measure.html#regionprops + """ + ) + parser.add_argument( + '--intensity_props', nargs = "+", + help=""" + Space separated list of additional metrics to be calculated for every marker separately. + By default only mean intensity is calculated. + If the metric doesn't depend on signal intensity, use --mask-props instead. + See list at https://scikit-image.org/docs/dev/api/skimage.measure.html#regionprops + Additionally available is gini_index, which calculates a single number + between 0 and 1, representing how unequal the signal is distributed in each region. + See https://en.wikipedia.org/wiki/Gini_coefficient + """ + ) #parser.add_argument('--suffix') args = parser.parse_args() #Create a dictionary object to pass to the next function dict = {'masks': args.masks, 'image': args.image,\ - 'channel_names': args.channel_names,'output':args.output} + 'channel_names': args.channel_names,'output':args.output, + 'intensity_props': set(args.intensity_props if args.intensity_props is not None else []).union(["intensity_mean"]), + 'mask_props': args.mask_props, + } #Print the dictionary object print(dict) #Return the dictionary |
| b |
| diff -r 928db0f952e3 -r aba3655fdef0 SingleCellDataExtraction.py --- a/SingleCellDataExtraction.py Fri Mar 12 00:19:24 2021 +0000 +++ b/SingleCellDataExtraction.py Fri Mar 11 23:35:52 2022 +0000 |
| [ |
| b'@@ -8,83 +8,93 @@\n import numpy as np\n import os\n import skimage.measure as measure\n+import tifffile\n+\n from pathlib import Path\n-import csv\n \n import sys\n \n \n-def MaskChannel(mask_loaded,image_loaded_z):\n+def gini_index(mask, intensity):\n+ x = intensity[mask]\n+ sorted_x = np.sort(x)\n+ n = len(x)\n+ cumx = np.cumsum(sorted_x, dtype=float)\n+ return (n + 1 - 2 * np.sum(cumx) / cumx[-1]) / n\n+\n+def intensity_median(mask, intensity):\n+ return np.median(intensity[mask])\n+\n+def MaskChannel(mask_loaded, image_loaded_z, intensity_props=["intensity_mean"]):\n """Function for quantifying a single channel image\n \n Returns a table with CellID according to the mask and the mean pixel intensity\n for the given channel for each cell"""\n- print(f\'Mask loaded: {mask_loaded.shape}\', file=sys.stderr)\n- print(f\'Image loaded: {image_loaded_z.shape}\', file=sys.stderr)\n- dat = measure.regionprops(mask_loaded, image_loaded_z)\n- n = len(dat)\n- intensity_z = np.empty(n)\n- for i in range(n):\n- intensity_z[i] = dat[i].mean_intensity\n- # Clear reference to avoid memory leak -- see MaskIDs for explanation.\n- dat[i] = None\n- return intensity_z\n+ # Look for regionprops in skimage\n+ builtin_props = set(intensity_props).intersection(measure._regionprops.PROP_VALS)\n+ # Otherwise look for them in this module\n+ extra_props = set(intensity_props).difference(measure._regionprops.PROP_VALS)\n+ dat = measure.regionprops_table(\n+ mask_loaded, image_loaded_z,\n+ properties = tuple(builtin_props),\n+ extra_properties = [globals()[n] for n in extra_props]\n+ )\n+ return dat\n \n \n-def MaskIDs(mask):\n+def MaskIDs(mask, mask_props=None):\n """This function will extract the CellIDs and the XY positions for each\n cell based on that cells centroid\n \n Returns a dictionary object"""\n \n- dat = measure.regionprops(mask)\n- n = len(dat)\n+ all_mask_props = set(["label", "centroid", "area", "major_axis_length", "minor_axis_length", "eccentricity", "solidity", "extent", "orientation"])\n+ if mask_props is not None:\n+ all_mask_props = all_mask_props.union(mask_props)\n \n- # Pre-allocate numpy arrays for all properties we\'ll calculate.\n- labels = np.empty(n, int)\n- xcoords = np.empty(n)\n- ycoords = np.empty(n)\n- area = np.empty(n, int)\n- minor_axis_length = np.empty(n)\n- major_axis_length = np.empty(n)\n- eccentricity = np.empty(n)\n- solidity = np.empty(n)\n- extent = np.empty(n)\n- orientation = np.empty(n)\n+ dat = measure.regionprops_table(\n+ mask,\n+ properties=all_mask_props\n+ )\n \n- for i in range(n):\n- labels[i] = dat[i].label\n- xcoords[i] = dat[i].centroid[1]\n- ycoords[i] = dat[i].centroid[0]\n- area[i] = dat[i].area\n- major_axis_length[i] = dat[i].major_axis_length\n- minor_axis_length[i] = dat[i].minor_axis_length\n- eccentricity[i] = dat[i].eccentricity\n- solidity[i] = dat[i].solidity\n- extent[i] = dat[i].extent\n- orientation[i] = dat[i].orientation\n- # By clearing the reference to each RegionProperties object, we allow it\n- # and its cache to be garbage collected immediately. Otherwise memory\n- # usage creeps up needlessly while this function is executing.\n- dat[i] = None\n+ name_map = {\n+ "CellID": "label",\n+ "X_centroid": "centroid-1",\n+ "Y_centroid": "centroid-0",\n+ "Area": "area",\n+ "MajorAxisLength": "major_axis_length",\n+ "MinorAxisLength": "minor_axis_length",\n+ "Eccentricity": "eccentricity",\n+ "Solidity": "solidity",\n+ "Extent": "extent",\n+ "Orientation": "orientation",\n+ }\n+ for new_name, old_name in name_map.items():\n+ dat[new_name] = dat[old_name]\n+ for old_name in set(name_map.values()):\n+ del dat[old_name]\n+\n+ return dat\n \n- IDs = {\n- "CellID": labels,\n- "X_centroid": xcoords'..b'fer:\n- #If header available\n- #channel_names_loaded = pd.read_csv(channel_names)\n- #channel_names_loaded_list = list(channel_names_loaded.marker_name)\n- #else:\n- #print("negative")\n- #old one column version\n- #channel_names_loaded = pd.read_csv(channel_names,header=None)\n- #Add a column index for ease\n- #channel_names_loaded.columns = ["marker"]\n- #channel_names_loaded = list(channel_names_loaded.marker.values)\n-\n #Read csv channel names\n channel_names_loaded = pd.read_csv(channel_names)\n- #Check for size of columns\n- if channel_names_loaded.shape[1] > 1:\n+ #Check for the presence of `marker_name` column\n+ if \'marker_name\' in channel_names_loaded:\n #Get the marker_name column if more than one column (CyCIF structure)\n channel_names_loaded_list = list(channel_names_loaded.marker_name)\n- else:\n- #old one column version -- re-read the csv file and add column name\n+ #Consider the old one-marker-per-line plain text format\n+ elif channel_names_loaded.shape[1] == 1:\n+ #re-read the csv file and add column name\n channel_names_loaded = pd.read_csv(channel_names, header = None)\n- #Add a column index for ease and for standardization\n- channel_names_loaded.columns = ["marker"]\n- channel_names_loaded_list = list(channel_names_loaded.marker)\n+ channel_names_loaded_list = list(channel_names_loaded.iloc[:,0])\n+ else:\n+ raise Exception(\'%s must contain the marker_name column\'%channel_names)\n \n+ #Contrast against the number of markers in the image\n+ if len(channel_names_loaded_list) != n_channels(image):\n+ raise Exception("The number of channels in %s doesn\'t match the image"%channel_names)\n+ \n #Check for unique marker names -- create new list to store new names\n channel_names_loaded_checked = []\n for idx,val in enumerate(channel_names_loaded_list):\n@@ -238,9 +230,6 @@\n #Otherwise, leave channel name\n channel_names_loaded_checked.append(val)\n \n- #Clear small memory amount by clearing old channel names\n- channel_names_loaded, channel_names_loaded_list = None, None\n-\n #Read the masks\n masks_loaded = {}\n #iterate through mask paths and read images to add to dictionary object\n@@ -249,22 +238,30 @@\n m_name = m_full_name.split(\'.\')[0]\n masks_loaded.update({str(m_name):skimage.io.imread(m,plugin=\'tifffile\')})\n \n- scdata_z = MaskZstack(masks_loaded,image,channel_names_loaded_checked)\n+ scdata_z = MaskZstack(masks_loaded,image,channel_names_loaded_checked, mask_props=mask_props, intensity_props=intensity_props)\n #Write the singe cell data to a csv file using the image name\n \n im_full_name = os.path.basename(image)\n im_name = im_full_name.split(\'.\')[0]\n- scdata_z.to_csv(str(Path(os.path.join(str(output),str(im_name+".csv")))),index=False)\n+\n+ # iterate through each mask and export csv with mask name as suffix\n+ for k,v in scdata_z.items():\n+ # export the csv for this mask name\n+ scdata_z[k].to_csv(\n+ str(Path(os.path.join(str(output),\n+ str(im_name+"_{}"+".csv").format(k)))),\n+ index=False\n+ )\n \n \n-def MultiExtractSingleCells(masks,image,channel_names,output):\n+def MultiExtractSingleCells(masks,image,channel_names,output, mask_props=None, intensity_props=["intensity_mean"]):\n """Function for iterating over a list of z_stacks and output locations to\n export single-cell data from image masks"""\n \n print("Extracting single-cell data for "+str(image)+\'...\')\n \n #Run the ExtractSingleCells function for this image\n- ExtractSingleCells(masks,image,channel_names,output)\n+ ExtractSingleCells(masks,image,channel_names,output, mask_props=mask_props, intensity_props=intensity_props)\n \n #Print update\n im_full_name = os.path.basename(image)\n' |
| b |
| diff -r 928db0f952e3 -r aba3655fdef0 macros.xml --- a/macros.xml Fri Mar 12 00:19:24 2021 +0000 +++ b/macros.xml Fri Mar 11 23:35:52 2022 +0000 |
| [ |
| @@ -2,12 +2,13 @@ <macros> <xml name="requirements"> <requirements> - <requirement type="package" version="3.6.10">python</requirement> - <requirement type="package" version="0.17.2">scikit-image</requirement> - <requirement type="package" version="2.10.0">h5py</requirement> - <requirement type="package" version="1.0.4">pandas</requirement> - <requirement type="package" version="1.18.5">numpy</requirement> - <requirement type="package" version="1.0.1">pathlib</requirement> + <container type="docker">labsyspharm/quantification:@VERSION@</container> + <requirement type="package" version="3.9">python</requirement> + <requirement type="package" version="0.18.0">scikit-image</requirement> + <requirement type="package">h5py</requirement> + <requirement type="package">pandas</requirement> + <requirement type="package">numpy</requirement> + <requirement type="package">pathlib</requirement> </requirements> </xml> @@ -19,6 +20,13 @@ </citations> </xml> - <token name="@VERSION@">1.3.1</token> - <token name="@CMD_BEGIN@">python ${__tool_directory__}/CommandSingleCellExtraction.py</token> + <token name="@VERSION@">1.5.1</token> + <token name="@CMD_BEGIN@"><![CDATA[ + QUANT_PATH=""; + if [ -f "/app/CommandSingleCellExtraction.py" ]; then + export QUANT_PATH="/app/CommandSingleCellExtraction.py"; + else + export QUANT_PATH="${__tool_directory__}/CommandSingleCellExtraction.py"; + fi; + ]]></token> </macros> |
| b |
| diff -r 928db0f952e3 -r aba3655fdef0 quantification.xml --- a/quantification.xml Fri Mar 12 00:19:24 2021 +0000 +++ b/quantification.xml Fri Mar 11 23:35:52 2022 +0000 |
| [ |
| @@ -18,6 +18,7 @@ @CMD_BEGIN@ + python \$QUANT_PATH --masks '${primary_mask.name}'.ome.tiff #if $supp_masks @@ -28,9 +29,17 @@ --image '${image.name}'.ome.tiff --output ./tool_out + + #if $mask_props + --mask_props $mask_props + #end if + #if $intensity_props + --intensity_props $intensity_props + #end if + --channel_names '$channel_names'; - mv ./tool_out/*.csv ./tool_out/quantified.csv; + cp tool_out/*cellMasks.csv cellMasks.csv ]]></command> <inputs> @@ -38,11 +47,16 @@ <param name="primary_mask" type="data" format="tiff" label="Primary Cell Mask"/> <param name="supp_masks" type="data" multiple="true" optional="true" format="tiff" label="Additional Cell Masks"/> <param name="channel_names" type="data" format="csv" label="Marker Channels"/> + <param name="mask_props" type="text" label="Mask Metrics" help="Space separated list of additional metrics to be calculated for every mask."/> + <param name="intensity_props" type="text" label="Intensity Metrics" help="Space separated list of additional metrics to be calculated for every marker separately."/> </inputs> <outputs> - <data format="csv" name="quant_out" from_work_dir="./tool_out/quantified.csv" label="${tool.name} on ${on_string}"/> - </outputs> + <data format="csv" name="cellmask" from_work_dir="cellMasks.csv" label="CellMaskQuant"/> + <collection type="list" name="quantification" label="${tool.name} on ${on_string}"> + <discover_datasets pattern="__designation_and_ext__" format="csv" directory="tool_out/" visible="true"/> + </collection> + </outputs> <help><![CDATA[ # Single cell quantification Module for single-cell data extraction given a segmentation mask and multi-channel image. The CSV structure is aligned with histoCAT output. |