improve calibration speed

Author: rodvrees

im2deep/calibrate.py

@@ -59,25 +59,13 @@ def get_ccs_shift(
     """
     LOGGER.debug(f"Using charge state {use_charge_state} for CCS shift calculation.")
 
-    tmp_df = cal_df.copy(deep=True)
-    tmp_ref_df = reference_dataset.copy(deep=True)
-
-    tmp_df["sequence"] = tmp_df["peptidoform"].apply(lambda x: x.proforma.split("\\")[0])
-    tmp_df["charge"] = tmp_df["peptidoform"].apply(lambda x: x.precursor_charge)
-    tmp_ref_df["sequence"] = tmp_ref_df["peptidoform"].apply(
-        lambda x: Peptidoform(x).proforma.split("\\")[0]
-    )
-    tmp_ref_df["charge"] = tmp_ref_df["peptidoform"].apply(
-        lambda x: Peptidoform(x).precursor_charge
-    )
-
-    reference_tmp = tmp_ref_df[tmp_ref_df["charge"] == use_charge_state]
-    df_tmp = tmp_df[tmp_df["charge"] == use_charge_state]
+    reference_tmp = reference_dataset[reference_dataset["charge"] == use_charge_state]
+    df_tmp = cal_df[cal_df["charge"] == use_charge_state]
     both = pd.merge(
         left=reference_tmp,
         right=df_tmp,
         right_on=["sequence", "charge"],
-        left_on=["sequence", "charge"],
+        left_on=["peptidoform", "charge"],
         how="inner",
         suffixes=("_ref", "_data"),
     )
@@ -90,7 +78,7 @@ def get_ccs_shift(
 
     # How much CCS in calibration data is larger than reference CCS, so predictions
     # need to be increased by this amount
-    return 0 if both.shape[0] == 0 else np.mean(both["ccs_observed"] - both["CCS"])
+    return 0 if both.empty else np.mean(both["ccs_observed"] - both["CCS"])
 
 
 def get_ccs_shift_per_charge(cal_df: pd.DataFrame, reference_dataset: pd.DataFrame) -> ndarray:
@@ -111,25 +99,11 @@ def get_ccs_shift_per_charge(cal_df: pd.DataFrame, reference_dataset: pd.DataFra
         CCS shift factors per charge state.
 
     """
-    tmp_df = cal_df.copy(deep=True)
-    tmp_ref_df = reference_dataset.copy(deep=True)
-
-    tmp_df["sequence"] = tmp_df["peptidoform"].apply(lambda x: x.proforma.split("\\")[0])
-    tmp_df["charge"] = tmp_df["peptidoform"].apply(lambda x: x.precursor_charge)
-    tmp_ref_df["sequence"] = tmp_ref_df["peptidoform"].apply(
-        lambda x: Peptidoform(x).proforma.split("\\")[0]
-    )
-    tmp_ref_df["charge"] = tmp_ref_df["peptidoform"].apply(
-        lambda x: Peptidoform(x).precursor_charge
-    )
-
-    reference_tmp = tmp_ref_df
-    df_tmp = tmp_df
     both = pd.merge(
-        left=reference_tmp,
-        right=df_tmp,
+        left=reference_dataset,
+        right=cal_df,
         right_on=["sequence", "charge"],
-        left_on=["sequence", "charge"],
+        left_on=["peptidoform", "charge"],
         how="inner",
         suffixes=("_ref", "_data"),
     )
@@ -159,7 +133,6 @@ def calculate_ccs_shift(
         CCS shift factor.
 
     """
-    cal_df["charge"] = cal_df["peptidoform"].apply(lambda x: x.precursor_charge)
     cal_df = cal_df[cal_df["charge"] < 7]  # predictions do not go higher for IM2Deep
 
     if not per_charge:
@@ -207,37 +180,35 @@ def linear_calibration(
 
     """
     LOGGER.info("Calibrating CCS values using linear calibration...")
+    calibration_dataset['sequence'] = calibration_dataset['peptidoform'].apply(lambda x: x.proforma.split("\\")[0])
+    calibration_dataset['charge'] = calibration_dataset['peptidoform'].apply(lambda x: x.precursor_charge)
+    # reference_dataset['sequence'] = reference_dataset['peptidoform'].apply(lambda x: x.split('/')[0])
+    reference_dataset['charge'] = reference_dataset['peptidoform'].apply(lambda x: int(x.split('/')[1]))
+
     if per_charge:
+        LOGGER.info('Getting general shift factor')
         general_shift = calculate_ccs_shift(
             calibration_dataset,
             reference_dataset,
             per_charge=False,
             use_charge_state=use_charge_state,
         )
+        LOGGER.info('Getting shift factors per charge state')
         shift_factor_dict = calculate_ccs_shift(
             calibration_dataset, reference_dataset, per_charge=True
         )
-        for charge in preds_df["charge"].unique():
-            if charge not in shift_factor_dict:
-                LOGGER.info(
-                    "No overlapping precursors for charge state {}. Using overall shift factor for precursors with that charge.".format(
-                        charge
-                    )
-                )
-                shift_factor_dict[charge] = general_shift
-        LOGGER.info("Shift factors per charge: {}".format(shift_factor_dict))
-        preds_df["predicted_ccs"] = preds_df.apply(
-            lambda x: x["predicted_ccs"] + shift_factor_dict[x["charge"]], axis=1
-        )
+
+        preds_df['shift'] = preds_df['charge'].map(shift_factor_dict).fillna(general_shift)
+        preds_df['predicted_ccs'] = preds_df['predicted_ccs'] + preds_df['shift']
+
     else:
         shift_factor = calculate_ccs_shift(
             calibration_dataset,
             reference_dataset,
             per_charge=False,
             use_charge_state=use_charge_state,
         )
-        preds_df["predicted_ccs"] = preds_df.apply(
-            lambda x: x["predicted_ccs"] + shift_factor, axis=1
-        )
+        preds_df['predicted_ccs'] += shift_factor
+
     LOGGER.info("CCS values calibrated.")
     return preds_df