Merge branch 'main' into update_n_inconsistent_peak

Author: ghar1821

CHANGELOG.md

@@ -64,6 +64,7 @@
 
 * Added Seurat rPCA (PR #95).
 
+* Added processing scripts for CLL dataset (PR #106).
 
 ## MAJOR CHANGES
 
@@ -165,4 +166,4 @@
 
 * Fix bug in EMD vertical where sample combination was malformed (PR #113)
 
-
+* Fix lisi inconsistent naming (PR #117) for issue #116.

scripts/create_resources/human_cll_mass_cytometry.sh

@@ -0,0 +1,36 @@
+#!/bin/bash
+
+# get the root of the directory
+REPO_ROOT=$(git rev-parse --show-toplevel)
+
+# ensure that the command below is run from the root of the repository
+cd "$REPO_ROOT"
+
+set -e
+
+RAW_DIR=resources_raw/human_cll_mass_cytometry/
+DATASET_ID=human_cll_mass_cytometry
+OUTPUT_DIR=resources/datasets_raw/$DATASET_ID/
+
+mkdir -p $OUTPUT_DIR
+
+# create raw dataset files
+python << HERE
+import anndata as ad
+
+adata = ad.read_h5ad("$RAW_DIR/human_cll_mass_cytometry.h5ad")
+
+# make sure the output is compressed
+adata.write_h5ad("$OUTPUT_DIR/common_dataset.h5ad", compression='gzip')
+HERE
+
+cat > $OUTPUT_DIR/state.yaml << HERE
+id: $DATASET_ID
+output_dataset: !file common_dataset.h5ad
+HERE
+
+# only run this if you have access to the openproblems-data bucket
+aws s3 sync --profile op \
+  resources/datasets_raw/human_cll_mass_cytometry \
+  s3://openproblems-data/resources/task_cyto_batch_integration/datasets_raw/human_cll_mass_cytometry/ \
+  --delete --dryrun

src/metrics/lisi/config.vsh.yaml

@@ -9,8 +9,6 @@ __merge__: ../../api/comp_metric.yaml
 # Can contain only lowercase letters or underscores.
 name: lisi
 
-
-
 # Metadata for your component
 info:
   metrics:

src/metrics/lisi/script.py

@@ -1,20 +1,19 @@
+import sys
+
 import anndata as ad
-import scib_metrics as sm
 import numpy as np
-import sys
+import scib_metrics as sm
 
 ## VIASH START
 # Note: this section is auto-generated by viash at runtime. To edit it, make changes
 # in config.vsh.yaml and then run `viash config inject config.vsh.yaml`.
 par = {
-  'input_unintegrated': 'resources_test/.../unintegrated.h5ad',
-  'input_integrated_split1': 'resources_test/.../integrated_split1.h5ad',
-  'input_integrated_split2': 'resources_test/.../integrated_split2.h5ad',
-  'output': 'output.h5ad'
-}
-meta = {
-  'name': 'lisi'
+    "input_unintegrated": "resources_test/.../unintegrated.h5ad",
+    "input_integrated_split1": "resources_test/.../integrated_split1.h5ad",
+    "input_integrated_split2": "resources_test/.../integrated_split2.h5ad",
+    "output": "output.h5ad",
 }
+meta = {"name": "lisi"}
 ## VIASH END
 
 sys.path.append(meta["resources_dir"])
@@ -23,10 +22,10 @@
     subset_markers_tocorrect,
 )
 
-print('Reading input files', flush=True)
-input_unintegrated = ad.read_h5ad(par['input_unintegrated'])
-input_integrated_split1 = ad.read_h5ad(par['input_integrated_split1'])
-input_integrated_split2 = ad.read_h5ad(par['input_integrated_split2'])
+print("Reading input files", flush=True)
+input_unintegrated = ad.read_h5ad(par["input_unintegrated"])
+input_integrated_split1 = ad.read_h5ad(par["input_integrated_split1"])
+input_integrated_split2 = ad.read_h5ad(par["input_integrated_split2"])
 
 print("Formatting input files", flush=True)
 integrated_s1, integrated_s2 = get_obs_var_for_integrated(
@@ -35,12 +34,14 @@
 integrated_s1 = subset_markers_tocorrect(integrated_s1)
 integrated_s2 = subset_markers_tocorrect(integrated_s2)
 
-print('Compute metrics', flush=True)
+print("Compute metrics", flush=True)
 n_batches = len(integrated_s1.obs.batch.unique())
 n_celltypes = len(integrated_s1.obs.cell_type.unique())
 
 print("Compute iLisi and cLisi for split 1", flush=True)
-knn = sm.nearest_neighbors.pynndescent(integrated_s1.layers['integrated'], n_neighbors=100, random_state=0)
+knn = sm.nearest_neighbors.pynndescent(
+    integrated_s1.layers["integrated"], n_neighbors=100, random_state=0
+)
 
 ilisi_s1_per_cell = sm.lisi_knn(knn, integrated_s1.obs.batch)
 ilisi_s1 = (np.nanmedian(ilisi_s1_per_cell) - 1) / (n_batches - 1)
@@ -49,7 +50,9 @@
 clisi_s1 = (n_celltypes - np.nanmedian(clisi_s1_per_cell)) / (n_celltypes - 1)
 
 print("Compute iLisi and cLisi for split 2", flush=True)
-knn = sm.nearest_neighbors.pynndescent(integrated_s2.layers['integrated'], n_neighbors=100, random_state=0)
+knn = sm.nearest_neighbors.pynndescent(
+    integrated_s2.layers["integrated"], n_neighbors=100, random_state=0
+)
 ilisi_s2_per_cell = sm.lisi_knn(knn, integrated_s2.obs.batch)
 ilisi_s2 = (np.nanmedian(ilisi_s2_per_cell) - 1) / (n_batches - 1)
 
@@ -58,21 +61,20 @@
 
 ilisi = np.mean([ilisi_s1, ilisi_s2])
 clisi = np.mean([clisi_s1, clisi_s2])
-uns_metric_ids = [ 'ilisi', 'clisi' ]
-uns_metric_values = [ ilisi, clisi ]
+uns_metric_ids = ["iLisi", "cLisi"]
+uns_metric_values = [ilisi, clisi]
 
 print("Write output AnnData to file", flush=True)
 output = ad.AnnData(
-  uns={
+    uns={
         "dataset_id": integrated_s1.uns["dataset_id"],
         "method_id": integrated_s1.uns["method_id"],
         "metric_ids": uns_metric_ids,
         "metric_values": uns_metric_values,
         "ilisi_s1_index": ilisi_s1_per_cell,
         "ilisi_s2_index": ilisi_s2_per_cell,
         "clisi_s1_index": clisi_s1_per_cell,
-        "clisi_s2_index": clisi_s2_per_cell
+        "clisi_s2_index": clisi_s2_per_cell,
     }
-
 )
-output.write_h5ad(par['output'], compression='gzip')
+output.write_h5ad(par["output"], compression="gzip")