SuperSCC.label_transfer.predict_label

SuperSCC.label_transfer.predict_label(query_data, models, wk_dir='/home/docs/checkouts/readthedocs.org/user_builds/superscc/checkouts/latest/docs/source', normalization_method='Min-Max', magic_based_imputation=False, pred_confidence_cutoff=None, filname=None, save=True, logger=None)[source]

A function to predict label and score the prediction by using pre-trained model.

Parameters:

  • query_data – A log normalized expression matrix. Rows are cells; Columns are features.

  • models – A regular expression to match the name of pre-trained models. Otherwise, a dict returned by load_pick_file function.

  • wk_dir – A string to specify the directory where pre-trained model files should be searched.

  • normalization_method – A string to decide how to normalize query data. Default is “Min-Max”. Other available words including “Standardization”. This parameter should be same as the one when running the model training, thereby guaranteeing same normalization on query and reference.

  • magic_based_imputation – A Bool value to decide whether MAGIC-based imputation should be employed. When False, the reference features lost in the query would be padded with 0. Default is False.

  • pred_confidence_cutoff – A float to decide the prediction accuracy cutoff. If the prediction score is lower than the cut off, the predicted cell label will return ‘uncertain’. Default is None. This argument

  • filename – A string decide the name of output. Default is None.

  • save – A Bool value to decide whether the result will be written into disk. Default is True.

  • logger – A log_file object to write log information into disk. Default is None.