added n_components and K as user-controlled parameters

ddqc/ddqc.py

@@ -134,15 +134,16 @@ def _boxplot_sorted(df, column, by, hline_x=None, log=False):
 # n_genes_lower_bound - maximum cutoff for n genes
 # percent_mito_upper_bound - minimum cutoff for percent mito
 # return_df_qc: return a dataframe with cluster labels and thresholds for each metric
-def ddqc_metrics(data: MultimodalData, res=1.3, method="mad", threshold=2, basic_n_genes=100, basic_percent_mito=80,
-                 mito_prefix="MT-",
+def ddqc_metrics(data: MultimodalData, res=1.3, n_components=50, K=20, method="mad", threshold=2, basic_n_genes=100,
+                 basic_percent_mito=80, mito_prefix="MT-",
                  ribo_prefix="^RP[SL][[:digit:]]|^RPLP[[:digit:]]|^RPSA", do_counts=True, do_genes=True, do_mito=True,
                  do_ribo=False, n_genes_lower_bound=200, percent_mito_upper_bound=10, random_state=29,
                  return_df_qc=False,
                  display_plots=True) -> Union[None, pd.DataFrame]:
     assert isinstance(data, MultimodalData)
     obs_copy, var_copy, uns_copy = _cluster_data(data, basic_n_genes, basic_percent_mito, mito_prefix, ribo_prefix,
-                                                 resolution=res, random_state=random_state)
+                                                 resolution=res, n_components=n_components, K=K,
+                                                 random_state=random_state)
 
     df_qc = pd.DataFrame({"cluster_labels": data.obs["louvain_labels"].values}, index=data.obs_names)
 

setup.py

@@ -3,7 +3,7 @@
 
 setup(
     name='ddqc',
-    version='0.1.0',
+    version='0.1.1',
     description='Description',
     url='https://github.com/ayshwaryas/ddqc',
     author='Author',

tutorials/ddqc_tutorial.ipynb

@@ -108,7 +108,7 @@
     "    - **Percent of mitochondrial transctipts**: keep cells that have percent_mito less than median + 2 MADs\n",
     "- In order to prevent the removal of healthy cells in clusters with high median n_genes and low percent_mito there are additional bounds for those thresholds:\n",
     "    - Cluster-level threshold for n_genes can't be greater than 200 (default). If it is greater, it will be set to 200.\n",
-    "    - Cluster-level threshold for percent_mito can't be lower than 10 (default). If it is lower, it will be set to 200."
+    "    - Cluster-level threshold for percent_mito can't be lower than 10 (default). If it is lower, it will be set to 10."
    ]
   },
   {
@@ -124,6 +124,8 @@
    "source": [
     "If you want to customize the filtering you can use the following parameters:\n",
     "- `res`: float - clustering resolution (default: 1.3)\n",
+    "- `n_components` - number of PCA components (default: 50)\n",
+    "- `K` - k to be used by `neighbors` Pegasus function (default: 20)\n",
     "- `method`: string - statistic on which the threshold would be calculated (default: \"mad\", available options \"mad\", \"outlier\")\n",
     "- `threshold`: float - parameter for the selected method (default: 2)\n",
     "- `basic_n_genes`: int - parameter for the initial QC n_genes filtering (default: 100)\n",
@@ -635,7 +637,11 @@
   {
    "cell_type": "code",
    "execution_count": 7,
-   "metadata": {},
+   "metadata": {
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -653,7 +659,11 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   },
    "outputs": [],
    "source": []
   }
@@ -679,4 +689,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 4
-}
+}