Example Usage

Here we will provide examples of different ways to use both the INQuant and Calibration classes. Example data can be genereated with the data_generation.py script.

The example data generator will generate a predictions file, 3 mzML files named file_1.mzml, file_2.mzml, and file_3.mzml, and a proteome file named example_proteome.fasta. The data will be generated in the directory, where the script is located, and examples will assume that the inquant_tools module is located in the same directory.

Calibration

At the moment, calibration must be run within the INQuant class, this will be deprecated when the calibration is implemented in the InstaNovo pipeline. As of now, calibration can be run standalone or combined with the quantification.

Standalone:

from inquant_tools import INQuant, Calibration

with INQuant(predictions_file="predictions.csv",
                 mzml_file_list = ["file_1.mzml", "file_2.mzml", "file_3.mzml"],
                 confidence_filter = 0
                ) as INQ:
    # Load predictions using INQuant
    predictions_dict = INQ.load_predictions()

    # Create an instance of Calibration and pass the INQuant predictions
    calibration = Calibration(predictions_dict)

    mzml_files = []
    for mzml_id, mzml_values in INQ.mzml_file_dict.items():
        path = mzml_values['file path']
        name = mzml_values['file name']
        calibration.load_mzml(path)
        calibration.calibration(mzml_id, set_training_seed=42)
        new_predictions, new_mzml = calibration.write_calibrated_file(INQ.experiment_path, mzml_name=mzml_id, predictions_name=INQ.experiment_name)
        mzml_files.append(new_mzml)

    new_tolerance = calibration.get_ppm_tolerance()

    print("New tolerance: ", new_tolerance)

The output is the maximum difference of error between the theoritical and calibrated m/z values in ppm. This can be used as the ppm_tolerance, in INQuant. The script will also generate new mzML files with the calibrated m/z values, until this is deprecated, when the calibration is implemented in the InstaNovo pipeline.

INQuant

Using the run function:

No user arguments:

from inquant_tools import INQuant

with INQuant(predictions_file="predictions.csv",
             mzml_file_list = ["file_1.mzml", "file_2.mzml", "file_3.mzml"],
             proteome_file="example_proteome.fasta"
            ) as INQ:
    INQ.run()

INQuant will promt the user for the column names of the predictions file, if these do not correspond to the default names. The run function combines MBR, quantification, protein alignment and inference, and output of files. All of these can be configured with the user arguments or run seperately.

Combined calibration and quantification:

from inquant_tools import INQuant, Calibration

mzml_file_list=[f"file_{num}.mzML" for num in [1,2,3]]
predictions = "predictions.csv"

with INQuant(
    predictions_file=predictions,
    mzml_file_list=mzml_file_list,
    script_purpose='Calibration example and runthrough of INQuant with auto generated example data',
    confidence_filter=0.0
    ) as INQ:
    # Load predictions using INQuant
    predictions_dict = INQ.load_predictions()

    # Create an instance of Calibration and pass the INQuant instance
    calibration = Calibration(predictions_dict)

    mzml_files = []
    for mzml_id, mzml_values in INQ.mzml_file_dict.items():
        path = mzml_values['file path']
        name = mzml_values['file name']
        calibration.load_mzml(path)
        calibration.calibration(mzml_id, set_training_seed=42)
        new_mzml = calibration.write_calibrated_mzml_file(INQ.experiment_path, mzml_name=mzml_id, predictions_name=INQ.experiment_name)
        mzml_files.append(new_mzml)
    
    new_tolerance = calibration.get_ppm_tolerance()

    INQ.ppm_tolerance = new_tolerance

    INQ.mzml_file_list = mzml_files

    INQ.run(mbr = True, 
            overwrite_all=True)