Cell Count EDA of NF1 single-cell plate data

0.data_analysis/README.md

@@ -3,3 +3,4 @@ This module focuses on exploring relationships in the cell morphology data.
 We organize this module into the following analyses:
 
 - `plate_correlation_analyses` explores similarities between aggregated well morphology feature, post feature selection, using genotype and plate groups.
+- `plate_cell_count_eda_analyses` explores NF1 plate cell count distributions.

0.data_analysis/plate_cell_count_eda_analyses/nbconverted/single_cell_fs_plate_cell_count_eda.py

@@ -0,0 +1,386 @@
+#!/usr/bin/env python
+# coding: utf-8
+
+# # Cell Count EDA of Single-Cell NF1 Plate Data
+# This analysis focuses on describing single-cell count distributions within the NF1 feature-selected single-cell plate data.
+
+# In[1]:
+
+
+import pathlib
+
+import matplotlib.pyplot as plt
+import pandas as pd
+import seaborn as sns
+
+# ## Find the root of the git repo on the host system
+
+# In[2]:
+
+
+# Get the current working directory
+cwd = pathlib.Path.cwd()
+
+if (cwd / ".git").is_dir():
+    root_dir = cwd
+
+else:
+    root_dir = None
+    for parent in cwd.parents:
+        if (parent / ".git").is_dir():
+            root_dir = parent
+            break
+
+# Check if a Git root directory was found
+if root_dir is None:
+    raise FileNotFoundError("No Git root directory found.")
+
+
+# ## Define paths
+
+# ### Input paths
+
+# In[3]:
+
+
+data_path = pathlib.Path(root_dir / "nf1_painting_repo/3.processing_features/data/single_cell_profiles").resolve(strict=True)
+
+plate3df_path = pathlib.Path(data_path / "Plate_3_sc_feature_selected.parquet").resolve(strict=True)
+plate3pdf_path = pathlib.Path(data_path / "Plate_3_prime_sc_feature_selected.parquet").resolve(strict=True)
+plate5df_path = pathlib.Path(data_path / "Plate_5_sc_feature_selected.parquet").resolve(strict=True)
+
+plate3df = pd.read_parquet(plate3df_path)
+plate3pdf = pd.read_parquet(plate3pdf_path)
+plate5df = pd.read_parquet(plate5df_path)
+
+
+# ### Output paths
+
+# In[4]:
+
+
+fig_path3 = pathlib.Path("plate_3_figures")
+fig_path3p = pathlib.Path("plate_3_prime_figures")
+fig_path5 = pathlib.Path("plate_5_figures")
+
+fig_path3.mkdir(parents=True, exist_ok=True)
+fig_path3p.mkdir(parents=True, exist_ok=True)
+fig_path5.mkdir(parents=True, exist_ok=True)
+
+# Map plate names to path for figure generation
+plate_name_to_path_map = {
+    "plate_3": fig_path3,
+    "plate_3p": fig_path3p,
+    "plate_5": fig_path5
+}
+
+
+# ## Processing Functions
+
+# In[5]:
+
+
+cell_count_col = "Metadata_number_of_singlecells"
+
+def agg_cell_count_to_site(_platedf):
+    """
+    Determine cell counts at the site level.
+
+    Parameters
+    ----------
+    _platedf: Pandas Dataframe
+        Single-cell morphology data for a single plate
+
+    Returns
+    -------
+    site_cell_countdf: Pandas Dataframe
+       Cell counts across sites with genotype and edge well columns
+    """
+
+    site_cell_countdf = (
+        _platedf.groupby(["Metadata_Well", "Metadata_Site"])
+        .agg(
+        {
+            cell_count_col: "count",
+            "Metadata_genotype": get_first_element,
+            "Metadata_Edge_Well": get_first_element
+        })
+    )
+
+    site_cell_countdf.reset_index(inplace=True)
+
+    return site_cell_countdf
+
+
+# In[6]:
+
+
+def agg_cell_count_genotype_edge_well(_df):
+    """
+    Determine cell counts for each unique (genotype, edge well) pair.
+
+    Parameters
+    ----------
+    _df: Pandas Dataframe
+        Single-cell morphology data for a single plate
+
+    Returns
+    -------
+    _df: Pandas Dataframe
+        Number of single-cells for each unique (genotype, edge well) pair.
+    """
+
+    _df = _df.groupby(["Metadata_genotype", "Metadata_Edge_Well"]).agg(
+        {
+        "Metadata_number_of_singlecells": "count",
+        }).reset_index()
+
+    return _df
+
+
+# In[7]:
+
+
+def combine_edge_well_data(_genotype_cell_countdf):
+    """
+    Parameters
+    ----------
+    _genotype_cell_countdf: Pandas Dataframe
+        Single-cell morphology data aggregated to the well level with well cell counts.
+
+    Returns
+    -------
+    Pandas dataframe with the well data duplicated for plotting cell count distributions.
+    """
+
+    total_well_type = _genotype_cell_countdf.groupby(["Metadata_genotype"]).agg(
+        {
+            "Metadata_number_of_singlecells": "sum"
+        }
+    ).reset_index()
+
+    total_well_type["Metadata_Edge_Well"] = "All"
+
+    return pd.concat([_genotype_cell_countdf, total_well_type], axis=0)
+
+
+# In[8]:
+
+
+def set_edge_well(_well):
+    """
+    Parameters
+    ----------
+    _well: String
+        The well name.
+
+    Returns
+    -------
+    A string specifying the category of well {Edge Well, Inner Well}
+    """
+
+    return "Edge Well" if _well in edge_wells else "Inner Well"
+
+
+# In[9]:
+
+
+# Custom function to retrieve the first element of metadata
+def get_first_element(_pdseries):
+    return _pdseries.iloc[0]
+
+
+# In[10]:
+
+
+def calc_missing_columns_to_cells(_platedf):
+    """
+    Calculates the number of cells with a matching frequency of missing columns.
+
+    Parameters
+    ----------
+    _platedf: Pandas Dataframe
+        Single-cell morphology data for a single plate.
+
+    Returns
+    -------
+    na_rows: Pandas Dataframe
+        Specifies the number of cells with a matching frequency of missing columns.
+        If no columns are missing by any cells, this is empty.
+    """
+
+    na_rows = _platedf.isna().sum(axis=1).value_counts()
+    na_rows = na_rows.reset_index(name="Number of Cells")
+    na_rows.rename(columns={"index": "Number of Missing Columns"}, inplace=True)
+    na_rows.drop(0, inplace=True)
+
+    return na_rows
+
+
+# ## Visualization Functions
+
+# In[11]:
+
+
+def visualize_missing_columns_to_cells(_platedf, _plate_name):
+    """
+    Visualize the number of cells with a matching frequency of missing columns in a bar plot.
+
+    Parameters
+    ----------
+    _platedf: Pandas Dataframe
+        Single-cell morphology data for a single plate
+
+    _plate_name: String
+        Name of the plate for labeling the data
+    """
+
+    na_rows = calc_missing_columns_to_cells(_platedf)
+
+    if na_rows.empty:
+        print("None of the cells have missing columns")
+
+    else:
+        print(na_rows)
+
+
+# In[12]:
+
+
+def visualize_cell_counts_per_edge_wells(_edge_well_df, _plate_name):
+    """
+    Visualize the number of cells per edge well category and genotype in a bar plot
+
+    Parameters
+    ----------
+    _edge_well_df: Pandas Dataframe
+        Contains the cell count per edge well category and genotype
+
+    _plate_name: String
+        Name of the plate for labeling the data
+    """
+    f, ax = plt.subplots(figsize=(20, 10))
+
+    sns.barplot(data=_edge_well_df, x="Metadata_genotype", y="Metadata_number_of_singlecells", hue="Metadata_Edge_Well")
+
+    plt.ylabel('Well Cell Count')
+    plt.xlabel('Genotype')
+    plt.title('Number of cells per Genotype and Edge Well')
+    plt.savefig(plate_name_to_path_map[_plate_name] / f"{_plate_name}_genotype_cell_counts_per_well_type.png")
+    plt.close()
+
+
+# In[13]:
+
+
+def visualize_cell_count_per_site(_cell_countdf, _plate_name):
+    """
+    Visualize the cell count in a violin plot
+
+    Parameters
+    ----------
+    _cell_countdf: Pandas Dataframe
+        Conatins the cell count data
+
+    _plate_name: String
+        Name of the plate for labeling the data
+    """
+
+    f, ax = plt.subplots(figsize=(20, 10))
+
+    sns.violinplot(x=cell_count_col, y="Metadata_genotype", data=_cell_countdf)
+    plt.ylabel("Genotype")
+    plt.xlabel("Site Cell Count")
+    plt.title("Site Cell Count for each Genotype")
+    plt.savefig(plate_name_to_path_map[_plate_name] / f"{_plate_name}_violin_site_cell_counts.png")
+
+
+# In[14]:
+
+
+def visualize_cell_count_per_edge_well_genotype(_cell_countdf, _plate_name):
+    """
+    Visualize the cell count in a violin plot
+
+    Parameters
+    ----------
+    _cell_countdf: Pandas Dataframe
+        Conatins the cell count data
+
+    _plate_name: String
+        Name of the plate for labeling the data
+    """
+
+    f, ax = plt.subplots(figsize=(20, 10))
+
+    sns.violinplot(x=cell_count_col, y="Metadata_genotype", data=_cell_countdf)
+    plt.ylabel("Genotype")
+    plt.xlabel("Site Cell Count")
+    plt.title("Site Cell Count for each Genotype")
+    plt.savefig(plate_name_to_path_map[_plate_name] / f"{_plate_name}_violin_site_cell_counts.png")
+
+
+# ## Visualize Number of Missing Columns
+
+# In[15]:
+
+
+visualize_missing_columns_to_cells(plate3df, "plate_3")
+visualize_missing_columns_to_cells(plate3df, "plate_3p")
+visualize_missing_columns_to_cells(plate3df, "plate_5")
+
+
+# ### Specify Edge Wells
+
+# In[16]:
+
+
+start_row = 'B'
+end_row = 'G'
+
+start_col = 1
+end_col = 12
+
+edge_wells = [
+    f"{chr(wletter)}{wnum}" for wletter in range(ord(start_row), ord(end_row) + 1)
+    for wnum in [start_col, end_col]
+]
+
+plate3df["Metadata_Edge_Well"] = plate3df["Metadata_Well"].apply(set_edge_well)
+plate3pdf["Metadata_Edge_Well"] = plate3pdf["Metadata_Well"].apply(set_edge_well)
+plate5df["Metadata_Edge_Well"] = plate5df["Metadata_Well"].apply(set_edge_well)
+
+
+# ## Aggregate and Visualize Cell Counts
+
+# ### Cell Counts per Site
+
+# In[17]:
+
+
+site_cell_count3df = agg_cell_count_to_site(plate3df)
+site_cell_count3pdf = agg_cell_count_to_site(plate3pdf)
+site_cell_count5df = agg_cell_count_to_site(plate5df)
+
+visualize_cell_count_per_site(site_cell_count3df, "plate_3")
+visualize_cell_count_per_site(site_cell_count3pdf, "plate_3p")
+visualize_cell_count_per_site(site_cell_count5df, "plate_5")
+
+
+# ### Cell Counts per Edge Well Category
+
+# In[18]:
+
+
+genotype_cell_count3df = agg_cell_count_genotype_edge_well(plate3df)
+genotype_cell_count3pdf = agg_cell_count_genotype_edge_well(plate3pdf)
+genotype_cell_count5df = agg_cell_count_genotype_edge_well(plate5df)
+
+edge_well_3df = combine_edge_well_data(genotype_cell_count3df)
+edge_well_3pdf = combine_edge_well_data(genotype_cell_count3pdf)
+edge_well_5df = combine_edge_well_data(genotype_cell_count5df)
+
+visualize_cell_counts_per_edge_wells(edge_well_3df, "plate_3")
+visualize_cell_counts_per_edge_wells(edge_well_3pdf, "plate_3p")
+visualize_cell_counts_per_edge_wells(edge_well_5df, "plate_5")
+