Parameterize standardization

inferelator_velocity/programs.py

@@ -44,6 +44,7 @@ def program_select(
     n_programs=2,
     n_comps=None,
     layer="X",
+    count_layer=None,
     normalize=True,
     mcv_loss_arr=None,
     n_jobs=-1,
@@ -109,7 +110,7 @@ def program_select(
 
     # CREATE A NEW DATA OBJECT FOR THIS ANALYSIS #
 
-    d = copy_count_layer(data, layer)
+    d = copy_count_layer(data, layer, counts_layer=count_layer)
 
     # PREPROCESSING / NORMALIZATION #
 

inferelator_velocity/times.py

@@ -47,7 +47,8 @@ def program_times(
     program_var_key=PROGRAM_KEY,
     programs=None,
     n_comps=None,
-    verbose=False
+    verbose=False,
+    standardize_count_data=True
 ):
     """
     Calcuate times for each cell based on known cluster time values
@@ -148,7 +149,8 @@ def program_times(
             return_components=True,
             verbose=verbose,
             n_comps=n_comps if n_comps is None else n_comps[prog],
-            wrap_time=wrap_time[prog] if wrap_time is not None else None
+            wrap_time=wrap_time[prog] if wrap_time is not None else None,
+            normalize_data=standardize_count_data
         )
 
         # Add keys to the .uns object

inferelator_velocity/utils/misc.py

@@ -164,7 +164,7 @@ def get_bins(data, n_bins=None, centers=None, width=None):
     return centers, half_width
 
 
-def copy_count_layer(data, layer):
+def copy_count_layer(data, layer, counts_layer=None):
 
     lref = data.X if layer == 'X' else data.layers[layer]
 
@@ -177,9 +177,26 @@ def copy_count_layer(data, layer):
             "these results will be nonsense."
         )
 
-    d = ad.AnnData(lref.astype(float))
-    d.layers['counts'] = lref.copy()
-    d.var = data.var.copy()
+    d = ad.AnnData(
+        lref,
+        var=data.var
+    )
+
+    if counts_layer is None:
+        d.layers['counts'] = lref.copy()
+    else:
+        d.layers['counts'] = data.layers[counts_layer]
+
+    if not pat.is_integer_dtype(d.X.dtype):
+        warnings.warn(
+            "Count data is expected, "
+            f"but {d.layers['counts'].dtype} data has been passed. "
+            "This data will be normalized and processed "
+            "as count data. If it is not count data, "
+            "these results will be nonsense."
+        )
+
+    d.X = d.X.astype(float)
 
     return d