Save progress for breaking up time points

pf2/figures/figureA16.py

@@ -29,9 +29,11 @@ def makeFigure():
 
     sns.violinplot(data=pat_df, x="Status", y="icu_day", hue="Status", hue_order=order, order=order, cut=0, ax=ax[0])
 
+    print(pat_df)
     pat_df["Day"] = pat_df.groupby("patient_id")["icu_day"].transform(
         lambda x: "1TP" if x.nunique() == 1 else ("2TP" if x.nunique() == 2 else ">=3TP")
     )
+    print(pat_df)
 
     count_df = (
             pat_df.groupby(["Status", "Day"], observed=True).size().reset_index(name="Sample Count")

pf2/figures/figureA18.py

@@ -0,0 +1,132 @@
+"""
+Figure A9:
+"""
+
+import numpy as np
+import pandas as pd
+import anndata
+from sklearn.metrics import accuracy_score
+import seaborn as sns
+from ..data_import import convert_to_patients, import_meta
+from ..predict import predict_mortality
+from .common import subplotLabel, getSetup
+from sklearn.metrics import RocCurveDisplay
+from sklearn.metrics import accuracy_score, roc_auc_score
+from pf2.figures.commonFuncs.plotGeneral import bal_combine_bo_covid, rotate_xaxis
+from ..data_import import add_obs, condition_factors_meta
+
+
+def makeFigure():
+    """Get a list of the axis objects and create a figure."""
+    ax, f = getSetup((6, 6), (3, 3))
+    subplotLabel(ax)
+
+    X = anndata.read_h5ad("/opt/northwest_bal/full_fitted.h5ad")
+
+    meta = import_meta()
+    conversions = convert_to_patients(X)
+
+    patient_factor = pd.DataFrame(
+        X.uns["Pf2_A"],
+        index=conversions,
+        columns=np.arange(X.uns["Pf2_A"].shape[1]) + 1,
+    )
+    meta = meta.loc[patient_factor.index, :]
+
+    time_point_counts = meta.groupby("patient_id").size().reset_index(name="TP")
+    meta = meta.merge(time_point_counts, on="patient_id", how="left")
+
+    meta["TP"] = meta["TP"].transform(
+        lambda x: "1TP" if x == 1 else ("2TP" if x == 2 else ">=3TP")
+    )
+    meta = meta.set_index("patient_id")
+    roc_auc = [False, True]
+    axs = 0
+    for i in range(2):
+        for j, timepoint in enumerate(["1TP", "2TP"]):
+            plsr_acc_df = pd.DataFrame([])
+            for k in range(3):
+                meta_timepoint = meta.loc[meta["TP"] == timepoint, :]
+                patient_factor_timepoint = patient_factor.loc[
+                    meta["TP"] == timepoint, :
+                ]
+                df = plsr_acc_proba(
+                    patient_factor_timepoint, meta_timepoint, n_components=k + 1, roc_auc=roc_auc[i]
+                )
+
+                df["Component"] = k + 1
+                plsr_acc_df = pd.concat([plsr_acc_df, df], axis=0)
+
+            plsr_acc_df = plsr_acc_df.melt(
+                id_vars="Component", var_name="Category", value_name="Accuracy"
+            )
+            print(plsr_acc_df)
+            sns.barplot(
+                data=plsr_acc_df, x="Component", y="Accuracy", hue="Category", ax=ax[axs]
+            )
+            ax[axs].set(title=timepoint)
+            axs += 1
+            if roc_auc[i] is True:
+                ax[i].set(ylim=[0, 1], ylabel="AUC ROC")
+            else:
+                ax[i].set(ylim=[0, 1], ylabel="Prediction Accuracy")
+
+        # plot_plsr_auc_roc(patient_factor, meta, ax[2])
+
+    return f
+
+
+def plsr_acc_proba(patient_factor_matrix, meta_data, n_components=2, roc_auc=True):
+    """Runs PLSR and obtains average prediction accuracy"""
+
+    acc_df = pd.DataFrame(columns=["Overall", "C19", "nC19"])
+
+    probabilities, labels = predict_mortality(
+        patient_factor_matrix, n_components=n_components, meta=meta_data, proba=True
+    )
+
+    probabilities = probabilities.round().astype(int)
+    meta_data = meta_data.loc[~meta_data.index.duplicated()].loc[labels.index]
+
+    if roc_auc:
+        score = roc_auc_score
+    else:
+        score = accuracy_score
+
+    covid_acc = score(
+        labels.loc[meta_data.loc[:, "patient_category"] == "COVID-19"],
+        probabilities.loc[meta_data.loc[:, "patient_category"] == "COVID-19"],
+    )
+    nc_acc = score(
+        labels.loc[meta_data.loc[:, "patient_category"] != "COVID-19"],
+        probabilities.loc[meta_data.loc[:, "patient_category"] != "COVID-19"],
+    )
+    acc = score(labels, probabilities)
+
+    acc_df.loc[0, :] = [acc, covid_acc, nc_acc]
+
+    return acc_df
+
+
+def plot_plsr_auc_roc(patient_factor_matrix, meta_data, ax):
+    """Runs PLSR and plots ROC AUC based on actual and prediction labels"""
+    probabilities, labels = predict_mortality(
+        patient_factor_matrix, meta_data, proba=True
+    )
+    meta_data = meta_data.loc[~meta_data.index.duplicated()].loc[labels.index]
+
+    RocCurveDisplay.from_predictions(
+        labels.loc[meta_data.loc[:, "patient_category"] == "COVID-19"],
+        probabilities.loc[meta_data.loc[:, "patient_category"] == "COVID-19"],
+        ax=ax,
+        name="C19",
+    )
+    RocCurveDisplay.from_predictions(
+        labels.loc[meta_data.loc[:, "patient_category"] != "COVID-19"],
+        probabilities.loc[meta_data.loc[:, "patient_category"] != "COVID-19"],
+        ax=ax,
+        name="nC19",
+    )
+    RocCurveDisplay.from_predictions(
+        labels, probabilities, plot_chance_level=True, ax=ax, name="Overall"
+    )

pf2/predict.py

@@ -3,7 +3,8 @@
 from sklearn.metrics import accuracy_score
 from sklearn.model_selection import StratifiedKFold
 
-SKF = StratifiedKFold(n_splits=10)
+SKF = StratifiedKFold(n_splits=5)
+# DONT FORGET TO CHANGE THIS
 
 
 def run_plsr(