Merge pull request #201 from wilhelm-lab/release/0.6.1

Release/0.6.1

.cookietemple.yml

@@ -15,5 +15,5 @@ full_name: Victor Giurcoiu
 email: [email protected]
 project_name: oktoberfest
 project_short_description: Public repo oktoberfest
-version: 0.6.0
+version: 0.6.1
 license: MIT

.github/release-drafter.yml

@@ -1,5 +1,5 @@
-name-template: "0.6.0 🌈" # <<COOKIETEMPLE_FORCE_BUMP>>
-tag-template: 0.6.0 # <<COOKIETEMPLE_FORCE_BUMP>>
+name-template: "0.6.1 🌈" # <<COOKIETEMPLE_FORCE_BUMP>>
+tag-template: 0.6.1 # <<COOKIETEMPLE_FORCE_BUMP>>
 exclude-labels:
     - "skip-changelog"
 

.gitignore

@@ -145,6 +145,7 @@ hash.file
 # output files in tutorials folder
 tutorials/
 !tutorials/Oktoberfest Tutorial.ipynb
+!tutorials/Oktoberfest_workshop.ipynb
 
 # example data
 data/

cookietemple.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.6.0
+current_version = 0.6.1
 
 [bumpversion_files_whitelisted]
 init_file = oktoberfest/__init__.py

docs/conf.py

@@ -54,9 +54,9 @@
 # the built documents.
 #
 # The short X.Y version.
-version = "0.6.0"
+version = "0.6.1"
 # The full version, including alpha/beta/rc tags.
-release = "0.6.0"
+release = "0.6.1"
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

docs/config.rst

@@ -22,7 +22,7 @@ Always required
    +----------------------------+--------------------------------------------------------------------------------------------------------------------+
    |     irt                    | Name of the model used for indexed retention time prediction                                                       |
    +----------------------------+--------------------------------------------------------------------------------------------------------------------+
-   | prediction_server          | Server and port for obtaining peptide property predictions; default: "koina.proteomicsdb.org:443"                  |
+   | prediction_server          | Server and port for obtaining peptide property predictions; default: "koina.wilhelmlab.org:443"                    |
    +----------------------------+--------------------------------------------------------------------------------------------------------------------+
    | ssl                        | Use ssl when making requests to the prediction server, can be true or false; default = true                        |
    +----------------------------+--------------------------------------------------------------------------------------------------------------------+

docs/jobs.rst

@@ -65,7 +65,7 @@ Example config file:
             "intensity": "Prosit_2020_intensity_HCD",
             "irt": "Prosit_2019_irt"
         },
-        "prediction_server": "koina.proteomicsdb.org:443",
+        "prediction_server": "koina.wilhelmlab.org:443",
         "numThreads": 1,
         "regressionMethod": "spline",
         "ssl": true,
@@ -138,7 +138,7 @@ Example config file:
             "specialAas": "KR",
             "db": "concat"
         },
-        "prediction_server": "koina.proteomicsdb.org:443",
+        "prediction_server": "koina.wilhelmlab.org:443",
         "numThreads": 1,
         "ssl": true
     }
@@ -190,7 +190,7 @@ Example config file:
             "intensity": "Prosit_2020_intensity_HCD",
             "irt": "Prosit_2019_irt"
         },
-        "prediction_server": "koina.proteomicsdb.org:443",
+        "prediction_server": "koina.wilhelmlab.org:443",
         "numThreads": 1,
         "fdr_estimation_method": "mokapot",
         "allFeatures": false,

docs/predictions.rst

@@ -1,18 +1,18 @@
 Retrieving predictions
 ======================
 
-Oktoberfest relies on retrieving predictions from a `Koina <https://koina.proteomicsdb.org/>`_ server that hosts specific models for peptide property prediction. Users can use any publicly available community server or host their own server.
+Oktoberfest relies on retrieving predictions from a `Koina <https://koina.wilhelmlab.org/>`_ server that hosts specific models for peptide property prediction. Users can use any publicly available community server or host their own server.
 
 Connecting to a community server
 --------------------------------
 
-Our publicly available community server is available at `koina.proteomicsdb.org:443`.
+Our publicly available community server is available at `koina.wilhelmlab.org:443`.
 If you want to connect to it, you need to have the following flags in your config file (default settings):
 
 .. code-block:: json
 
    {
-      "prediction_server": "koina.proteomicsdb.org:443",
+      "prediction_server": "koina.wilhelmlab.org:443",
       "ssl": true,
    }
 
@@ -31,13 +31,13 @@ This is the list of currently supported and tested models for Oktoberfest provid
    +==================================================================================================================+==============================================================================================================================================================================================+
    | Prosit_2019_intensity                                                                                            | Developed for HCD tryptic peptides only. We recommend using the Prosit_2020_intensity_HCD model instead, since it showed slightly superior performance on tryptic peptides as well.          |
    +------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
-   | `Prosit_2020_intensity_HCD <https://koina.proteomicsdb.org/docs#post-/Prosit_2020_intensity_HCD/infer>`_         | Developed for HCD tryptic and non-tryptic peptides. Supported modifications are oxidation and carbamidomethylation. Latest version we recommend to use for HCD.                              |
+   | `Prosit_2020_intensity_HCD <https://koina.wilhelmlab.org/docs#post-/Prosit_2020_intensity_HCD/infer>`_           | Developed for HCD tryptic and non-tryptic peptides. Supported modifications are oxidation and carbamidomethylation. Latest version we recommend to use for HCD.                              |
    +------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
-   | `Prosit_2020_intensity_CID <https://koina.proteomicsdb.org/docs#post-/Prosit_2020_intensity_CID/infer>`_         | Developed for CID tryptic and non-tryptic peptides. Supported modifications are oxidation and carbamidomethylation. Latest version we recommend to use for CID.                              |
+   | `Prosit_2020_intensity_CID <https://koina.wilhelmlab.org/docs#post-/Prosit_2020_intensity_CID/infer>`_           | Developed for CID tryptic and non-tryptic peptides. Supported modifications are oxidation and carbamidomethylation. Latest version we recommend to use for CID.                              |
    +------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
-   | `Prosit_2020_intensity_TMT <https://koina.proteomicsdb.org/docs#post-/Prosit_2020_intensity_TMT/infer>`_         | Developed for HCD and CID, tryptic and non-tryptic peptides. Latest version we commend for TMT labeled peptides in general.                                                                  |
+   | `Prosit_2020_intensity_TMT <https://koina.wilhelmlab.org/docs#post-/Prosit_2020_intensity_TMT/infer>`_           | Developed for HCD and CID, tryptic and non-tryptic peptides. Latest version we commend for TMT labeled peptides in general.                                                                  |
    +------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
-   | `Prosit_2023_intensity_timsTOF <https://koina.proteomicsdb.org/docs#post-/Prosit_2023_intensity_timsTOF/infer>`_ | Developed for timsTOF, tryptic and non-tryptic peptides. Latest version we commend to use for timsTOF.                                                                                       |
+   | `Prosit_2023_intensity_timsTOF <https://koina.wilhelmlab.org/docs#post-/Prosit_2023_intensity_timsTOF/infer>`_   | Developed for timsTOF, tryptic and non-tryptic peptides. Latest version we commend to use for timsTOF.                                                                                       |
    +------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
 
 .. table::
@@ -46,14 +46,14 @@ This is the list of currently supported and tested models for Oktoberfest provid
    +-----------------------------------------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------+
    | iRT models                                                                                    |                             Description                                                                                   |
    +===============================================================================================+===========================================================================================================================+
-   | `Prosit_2019_irt <https://koina.proteomicsdb.org/docs#post-/Prosit_2019_irt/infer>`_          | While developed for tryptic peptides only, we did not observe a drop in prediction performance for non-tryptic peptides.  |
+   | `Prosit_2019_irt <https://koina.wilhelmlab.org/docs#post-/Prosit_2019_irt/infer>`_            | While developed for tryptic peptides only, we did not observe a drop in prediction performance for non-tryptic peptides.  |
    +-----------------------------------------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------+
-   | `Prosit_2020_irt_TMT <https://koina.proteomicsdb.org/docs/#post-/Prosit_2020_irt_TMT/infer>`_ | Developed for TMT labeled peptides.                                                                                       |
+   | `Prosit_2020_irt_TMT <https://koina.wilhelmlab.org/docs/#post-/Prosit_2020_irt_TMT/infer>`_   | Developed for TMT labeled peptides.                                                                                       |
    +-----------------------------------------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------+
 
 Once support for additional models is implemented in Oktoberfest, they will be added here.
 
 Hosting and adding your own models
 ----------------------------------
 
-In case you are planning to host your own private or public instance of Koina or want us to host your model, please refer to the official `Koina documentation <https://koina.proteomicsdb.org/docs#overview>`_.
+In case you are planning to host your own private or public instance of Koina or want us to host your model, please refer to the official `Koina documentation <https://koina.wilhelmlab.org/docs#overview>`_.

oktoberfest/__init__.py

@@ -5,7 +5,7 @@
 __author__ = """The Oktoberfest development team (Wilhelmlab at Technical University of Munich)"""
 __copyright__ = f"Copyright {datetime.now():%Y}, Wilhelmlab at Technical University of Munich"
 __license__ = "MIT"
-__version__ = "0.6.0"
+__version__ = "0.6.1"
 
 import logging.handlers
 import sys

oktoberfest/data/spectra.py

@@ -214,6 +214,12 @@ def from_hdf5(cls: Type[SpectraT], input_file: Union[str, Path]) -> SpectraT:
         sparse_raw_intensities = hdf5.read_file(input_file, f"sparse_{hdf5.INTENSITY_RAW_KEY}")
         if not sparse_raw_intensities.empty:
             spectra.add_matrix_from_hdf5(sparse_raw_intensities, FragmentType.RAW)
+        try:
+            sparse_pred_intensities = hdf5.read_file(input_file, f"sparse_{hdf5.INTENSITY_PRED_KEY}")
+            if not sparse_pred_intensities.empty:
+                spectra.add_matrix_from_hdf5(sparse_pred_intensities, FragmentType.PRED)
+        except Exception as e:
+            logger.warning(e)
         sparse_raw_mzs = hdf5.read_file(input_file, f"sparse_{hdf5.MZ_RAW_KEY}")
         if not sparse_raw_mzs.empty:
             spectra.add_matrix_from_hdf5(sparse_raw_mzs, FragmentType.MZ)

oktoberfest/plotting/plotting.py

@@ -118,11 +118,16 @@ def joint_plot(
         height=10,
         joint_kws={"rasterized": True, "edgecolor": "none", "s": 10},
     )
+    jplot.ax_joint.axhline(y=0, c="red")
+    jplot.ax_joint.axvline(x=0, c="red")
+    jplot.ax_marg_y.axhline(y=0, c="red")
+    jplot.ax_marg_x.axvline(x=0, c="red")
+
     jplot.ax_joint.set_ylabel("Score\n(peptide property prediction)")
     jplot.ax_joint.set_xlabel("Score\n(search engine)")
     jplot.fig.suptitle(f"Score distribution ({level.capitalize()})", y=0.99)
     plt.savefig(filename, dpi=300)
-    plt.plot()
+    plt.show()
     plt.close()
 
 
@@ -196,7 +201,7 @@ def plot_gain_loss(prosit_target: pd.DataFrame, andromeda_target: pd.DataFrame,
     ax.spines["top"].set_visible(False)
     ax.spines["bottom"].set_visible(False)
     # grid
-    ax.set_ylabel("number of lost-common-shared targets below 1% FDR")
+    ax.set_ylabel(f"number of target {level.lower()}s below 1% FDR")
     ax.set_axisbelow(True)
     ax.yaxis.grid(color="black")
     ax.tick_params(axis="y", which="major")
@@ -207,7 +212,7 @@ def plot_gain_loss(prosit_target: pd.DataFrame, andromeda_target: pd.DataFrame,
     legend_label = ["Common", "Gained", "Lost"]
     plt.legend(legend_label, ncol=1, bbox_to_anchor=([1.2, 0.5, 0, 0]), frameon=False)
     plt.savefig(filename, dpi=300, bbox_inches="tight")
-    plt.plot()
+    plt.show()
     plt.close()
 
 
@@ -236,7 +241,11 @@ def plot_violin_sa_ce(sa_ce_df: pd.DataFrame, filename: Union[str, Path]):
     """
     fig, ax = plt.subplots(figsize=(8, 8))
     sns.violinplot(data=sa_ce_df, x="COLLISION_ENERGY", y="SPECTRAL_ANGLE", ax=ax, color="#1f77b4")
-    ax.axvline(x=sa_ce_df["COLLISION_ENERGY"][sa_ce_df["SPECTRAL_ANGLE"].idxmax()], color="red")
+    ax.axvline(
+        x=sa_ce_df["COLLISION_ENERGY"][sa_ce_df["SPECTRAL_ANGLE"].idxmax()] - sa_ce_df["COLLISION_ENERGY"].min(),
+        color="red",
+    )
+    plt.xticks(rotation=90)
     plt.grid()
     plt.savefig(filename, dpi=300)
     plt.plot()

oktoberfest/predict/koina.py

@@ -34,7 +34,7 @@ class Koina:
     def __init__(
         self,
         model_name: str,
-        server_url: str = "koina.proteomicsdb.org:443",
+        server_url: str = "koina.wilhelmlab.org:443",
         ssl: bool = True,
         targets: Optional[List[str]] = None,
         disable_progress_bar: bool = False,
@@ -49,7 +49,7 @@ def __init__(
         and that the specified model is available on the server.
 
         :param model_name: The name of the Koina model to be used for inference.
-        :param server_url: The URL of the inference server. Defaults to "koina.proteomicsdb.org:443".
+        :param server_url: The URL of the inference server. Defaults to "koina.wilhelmlab.org:443".
         :param ssl: Indicates whether to use SSL for communication with the server. Defaults to True.
         :param targets: An optional list of targets to predict. If this is None, all model targets are
             predicted and received.
@@ -100,7 +100,7 @@ def _is_server_ready(self):
             if not self.client.is_server_live():
                 raise ValueError("Server not yet started.")
         except InferenceServerException as e:
-            if self.url == "koina.proteomicsdb.org:443":
+            if self.url in ["koina.wilhelmlab.org:443", "koina.proteomicsdb.org:443"]:
                 if self.ssl:
                     raise InferenceServerException(
                         "The public koina network seems to be inaccessible at the moment. "

oktoberfest/preprocessing/preprocessing.py

@@ -180,7 +180,7 @@ def filter_peptides(peptides: pd.DataFrame, min_length: int, max_length: int, ma
         & (~peptides["MODIFIED_SEQUENCE"].str.contains(r"\(ac\)"))
         & (~peptides["MODIFIED_SEQUENCE"].str.contains(r"\(Acetyl \(Protein N-term\)\)"))
         & (~peptides["MODIFIED_SEQUENCE"].str.contains(r"\[UNIMOD\:21\]"))
-        & (~peptides["SEQUENCE"].str.contains("U|X"))
+        & (~peptides["SEQUENCE"].str.contains(r"B|\*|\.|U|X|Z"))
     ]
 
 

oktoberfest/runner.py

@@ -351,6 +351,14 @@ def generate_spectral_lib(config_path: Union[str, Path]):
             # Create a pool for producer processes
             predictor_pool = pool.Pool(config.num_threads)
 
+            consumer_process = Process(
+                target=speclib.async_write,
+                args=(
+                    shared_queue,
+                    writing_progress,
+                ),
+            )
+
             try:
                 results = []
                 for i in batches:
@@ -376,13 +384,7 @@ def generate_spectral_lib(config_path: Union[str, Path]):
                     total=n_batches, desc="Writing library", postfix={"successful": 0, "missing": 0}
                 ) as writer_pbar:
                     # Start the consumer process
-                    consumer_process = Process(
-                        target=speclib.async_write,
-                        args=(
-                            shared_queue,
-                            writing_progress,
-                        ),
-                    )
+
                     consumer_process.start()
                     with tqdm(
                         total=n_batches, desc="Getting predictions", postfix={"successful": 0, "failed": 0}
@@ -475,23 +477,28 @@ def _calculate_features(spectra_file: Path, config: Config):
     if calc_feature_step.is_done():
         return
 
-    predict_kwargs = {
-        "server_url": config.prediction_server,
-        "ssl": config.ssl,
-    }
+    predict_step = ProcessStep(config.output, "predict." + spectra_file.stem)
+    if not predict_step.is_done():
 
-    pred_intensities = pr.predict(
-        data=library.spectra_data,
-        model_name=config.models["intensity"],
-        **predict_kwargs,
-    )
+        predict_kwargs = {
+            "server_url": config.prediction_server,
+            "ssl": config.ssl,
+        }
 
-    pred_irts = pr.predict(data=library.spectra_data, model_name=config.models["irt"], **predict_kwargs)
+        pred_intensities = pr.predict(
+            data=library.spectra_data,
+            model_name=config.models["intensity"],
+            **predict_kwargs,
+        )
 
-    library.add_matrix(pd.Series(pred_intensities["intensities"].tolist(), name="intensities"), FragmentType.PRED)
-    library.add_column(pred_irts["irt"], name="PREDICTED_IRT")
+        pred_irts = pr.predict(data=library.spectra_data, model_name=config.models["irt"], **predict_kwargs)
 
-    library.write_pred_as_hdf5(config.output / "data" / spectra_file.with_suffix(".mzml.pred.hdf5").name).join()
+        library.add_matrix(pd.Series(pred_intensities["intensities"].tolist(), name="intensities"), FragmentType.PRED)
+        library.add_column(pred_irts["irt"], name="PREDICTED_IRT")
+
+        library.write_pred_as_hdf5(config.output / "data" / spectra_file.with_suffix(".mzml.pred.hdf5").name).join()
+
+        predict_step.mark_done()
 
     # produce percolator tab files
     fdr_dir = config.output / "results" / config.fdr_estimation_method