Merge pull request #47 from klarman-cell-observatory/boli

Remove zero categories when using hashing or genetic pooling, modifie…

ext_modules/fast_funcs.pyx

@@ -20,8 +20,7 @@ cpdef tuple split_barcode_channel(str[:] arr):
     for i in range(size):
         res = arr[i].rsplit(sep = '-', maxsplit = 1)
         bview[i] = res[0]
-        if len(res) > 1:
-            cview[i] = res[1]
+        cview[i] = res[1]
 
     return (barcodes, channels)
 

pegasusio/qc_utils.py

@@ -1,4 +1,5 @@
 import numpy as np
+import pandas as pd
 
 from pegasusio import UnimodalData
 
@@ -158,6 +159,10 @@ def apply_qc_filters(unidata: UnimodalData):
         cols = ["passed_qc"]
         if unidata.uns.get("__del_demux_type", False):
             cols.append("demux_type")
+            if "assignment" in unidata.obs:
+                # remove categories that contain no elements
+                series = unidata.obs["assignment"].value_counts(sort = False)
+                unidata.obs["assignment"] = pd.Categorical(unidata.obs["assignment"], categories = series[series > 0].index.astype(str))
             # del unidata.uns["__del_demux_type"]
 
         unidata.obs.drop(columns=cols, inplace=True)

pegasusio/unimodal_data.py

@@ -345,10 +345,10 @@ def _inplace_subset_var(self, index: List[bool]) -> None:
 
 
     def separate_channels(self) -> None:
-        """ Separate channel information from barcodekeys, only used for 10x v2, v3 h5 and mtx.
+        """ Separate channel information from barcodekeys, used for 10x v2, v3 h5 and mtx as well as Optimus loom.
         """
-        if self.barcode_metadata.shape[0] == 0:
-            return None # no data
+        if self.barcode_metadata.shape[0] == 0 or self.barcode_metadata.index[0].find("-") < 0:
+            return None # no data or no dash to remove
 
         try:
             from pegasusio.cylib.funcs import split_barcode_channel
@@ -357,10 +357,7 @@ def separate_channels(self) -> None:
 
         barcodes, channels = split_barcode_channel(self.barcode_metadata.index.values)
 
-        if channels[0] is None:
-            return None # no need to separate channel information and the file should not be generated by cellranger
-
-        if (channels != "1").sum() > 0:
+        if np.unique(channels).size > 1:
             # we have multiple channels
             self.barcode_metadata["Channel"] = channels
             barcodes = np.array([x + "-" + y for x, y in zip(channels, barcodes)], dtype = object)