Revert "Removed cLisi metric from pipeline (#145)"

CHANGELOG.md

@@ -157,8 +157,6 @@
 
 * Change inconsistent peaks metrics to consistent peaks (PR #119). 
 
-* Removed cLisi metric as it is not informative (PR #145).
-
 ## MINOR CHANGES
 
 * Enabled unit tests (PR #2).

src/metrics/lisi/config.vsh.yaml

@@ -48,6 +48,48 @@ info:
       # Whether a higher value represents a 'better' solution (required)
       maximize: true
 
+
+    - name: cLisi
+      label: cLisi
+      summary: "Cell type Lisi score."
+      description: |
+        Compute the cell type local inverse simpson index (cLISI) for each cell, then return the median 
+        as the final score. 
+      references:
+        doi: 
+          - 10.1038/s41592-019-0619-0
+        bibtex:
+          - |
+            @article{korsunsky2019fast,
+            title={Fast, sensitive and accurate integration of single-cell data with Harmony},
+            author={Korsunsky, Ilya and Millard, Nghia and Fan, Jean and Slowikowski, Kamil and Zhang, Fan and Wei, Kevin and Baglaenko, Yuriy and Brenner, Michael and Loh, Po-ru and Raychaudhuri, Soumya},
+            journal={Nature methods},
+            volume={16},
+            number={12},
+            pages={1289--1296},
+            year={2019},
+            publisher={Nature Publishing Group US New York}
+            }
+      links:
+        # URL to the documentation for this metric (required).
+        documentation: https://scib-metrics.readthedocs.io/en/latest/api.html
+        # URL to the code repository for this metric (required).
+        repository: https://github.com/YosefLab/scib-metrics
+      # The minimum possible value for this metric (required)
+      min: -0.0001
+      # The maximum possible value for this metric (required)
+      max: 1.0001
+      # Whether a higher value represents a 'better' solution (required)
+      maximize: true
+
+
+# Component-specific parameters (optional)
+# arguments:
+#   - name: "--n_neighbors"
+#     type: "integer"
+#     default: 5
+#     description: Number of neighbors to use.
+
 # Resources required to run the component
 resources:
   # The script of your component (required)

src/metrics/lisi/helper.py

@@ -77,9 +77,6 @@ def compute_ilisi(adata: ad.AnnData, n_batches: int):
 
 def compute_clisi(integrated: ad.AnnData):
     """
-    No longer used in the current version of the script, but kept in case
-    it is needed in the future.
-
     Compute cLISI for an integrated AnnData object.
 
     Args:

src/metrics/lisi/script.py

@@ -61,9 +61,20 @@
 
 ilisi = np.mean([ilisi_s1, ilisi_s2])
 
+print("Compute cLisi for split 1", flush=True)
+clisi_s1, clisi_s1_per_cell = lisi_helper.compute_clisi(integrated_s1)
+
+print("Compute cLisi for split 2", flush=True)
+clisi_s2, clisi_s2_per_cell = lisi_helper.compute_clisi(integrated_s2)
+
+clisi = np.mean([clisi_s1, clisi_s2])
+
 print(f"iLisi split 1: {ilisi_s1}", flush=True)
 print(f"iLisi split 2: {ilisi_s2}", flush=True)
 print(f"Mean iLisi: {ilisi}", flush=True)
+print(f"cLisi split 1: {clisi_s1}", flush=True)
+print(f"cLisi split 2: {clisi_s2}", flush=True)
+print(f"Mean cLisi: {clisi}", flush=True)
 
 print("Write output AnnData to file", flush=True)
 # uns layout:
@@ -74,8 +85,8 @@
 uns = {
     "dataset_id": integrated_s1.uns["dataset_id"],
     "method_id": integrated_s1.uns["method_id"],
-    "metric_ids": ["iLisi"],
-    "metric_values": [ilisi],
+    "metric_ids": ["iLisi", "cLisi"],
+    "metric_values": [ilisi, clisi],
     "ilisi_per_split": {
         "split_1": {
             "score": ilisi_s1,
@@ -88,6 +99,18 @@
             "cell_ids": np.array(integrated_s2.obs_names),
         },
     },
+    "clisi_per_split": {
+        "split_1": {
+            "score": clisi_s1,
+            "per_cell": clisi_s1_per_cell,
+            "cell_ids": np.array(integrated_s1.obs_names),
+        },
+        "split_2": {
+            "score": clisi_s2,
+            "per_cell": clisi_s2_per_cell,
+            "cell_ids": np.array(integrated_s2.obs_names),
+        },
+    },
 }
 
 output = ad.AnnData(uns=uns)