Merge pull request #133 from singjc/feature/aripf

Feature: Alignment Scoring and Across Run IPF

Author: singjc

pyprophet/data_handling.py

@@ -143,6 +143,7 @@ def prepare_data_table(table,
                        decoy_name="decoy",
                        main_score_name=None,
                        score_columns=None,
+                       level=None
                        ):
     N = len(table)
     if not N:
@@ -219,7 +220,7 @@ def prepare_data_table(table,
 
     tg_ids = table[tg_id_name]
 
-    if not check_for_unique_blocks(tg_ids):
+    if not check_for_unique_blocks(tg_ids) and level != 'alignment':
         raise click.ClickException("" + tg_id_name + " values do not form unique blocks in input file(s).")
 
     tg_map = dict()
@@ -285,6 +286,8 @@ def update_chosen_main_score_in_table(train, score_columns, use_as_main_score):
     train.df.insert(5, temp_col.name, temp_col)
     click.echo(f"Info: Updated main score column from {old_main_score_column} to {use_as_main_score}...")
     return train, tuple(updated_score_columns)
+
+
 class Experiment(object):
 
     @profile

pyprophet/ipf.py

@@ -78,7 +78,8 @@ def type_cast_value(self, ctx, value):
 @click.option('--lfdr_adj', default=1.5, show_default=True, type=float, help='Numeric value that is applied as a multiple of the smoothing bandwidth used in the density estimation.')
 @click.option('--lfdr_eps', default=np.power(10.0,-8), show_default=True, type=float, help='Numeric value that is threshold for the tails of the empirical p-value distribution.')
 # OpenSWATH options
-@click.option('--level', default='ms2', show_default=True, type=click.Choice(['ms1', 'ms2', 'ms1ms2', 'transition']), help='Either "ms1", "ms2", "ms1ms2" or "transition"; the data level selected for scoring. "ms1ms2 integrates both MS1- and MS2-level scores and can be used instead of "ms2"-level results."')
+@click.option('--level', default='ms2', show_default=True, type=click.Choice(['ms1', 'ms2', 'ms1ms2', 'transition', 'alignment']), help='Either "ms1", "ms2", "ms1ms2", "transition", or "alignment"; the data level selected for scoring. "ms1ms2 integrates both MS1- and MS2-level scores and can be used instead of "ms2"-level results."')
+@click.option('--add_alignment_features/--no-add_alignment_features', default=False, show_default=True, help='Add alignment features to scoring.')
 # IPF options
 @click.option('--ipf_max_peakgroup_rank', default=1, show_default=True, type=int, help='Assess transitions only for candidate peak groups until maximum peak group rank.')
 @click.option('--ipf_max_peakgroup_pep', default=0.7, show_default=True, type=float, help='Assess transitions only for candidate peak groups until maximum posterior error probability.')
@@ -92,7 +93,7 @@ def type_cast_value(self, ctx, value):
 # Processing
 @click.option('--threads', default=1, show_default=True, type=int, help='Number of threads used for semi-supervised learning. -1 means all available CPUs.', callback=transform_threads)
 @click.option('--test/--no-test', default=False, show_default=True, help='Run in test mode with fixed seed.')
-def score(infile, outfile, classifier, xgb_autotune, apply_weights, xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, ss_num_iter, ss_main_score, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, level, ipf_max_peakgroup_rank, ipf_max_peakgroup_pep, ipf_max_transition_isotope_overlap, ipf_min_transition_sn, tric_chromprob, threads, test, ss_score_filter, color_palette, main_score_selection_report):
+def score(infile, outfile, classifier, xgb_autotune, apply_weights, xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, ss_num_iter, ss_main_score, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, level, add_alignment_features, ipf_max_peakgroup_rank, ipf_max_peakgroup_pep, ipf_max_transition_isotope_overlap, ipf_min_transition_sn, tric_chromprob, threads, test, ss_score_filter, color_palette, main_score_selection_report):
     """
     Conduct semi-supervised learning and error-rate estimation for MS1, MS2 and transition-level data. 
     """
@@ -112,10 +113,10 @@ def score(infile, outfile, classifier, xgb_autotune, apply_weights, xeval_fracti
     xgb_params_space = {'eta': hp.uniform('eta', 0.0, 0.3), 'gamma': hp.uniform('gamma', 0.0, 0.5), 'max_depth': hp.quniform('max_depth', 2, 8, 1), 'min_child_weight': hp.quniform('min_child_weight', 1, 5, 1), 'subsample': 1, 'colsample_bytree': 1, 'colsample_bylevel': 1, 'colsample_bynode': 1, 'lambda': hp.uniform('lambda', 0.0, 1.0), 'alpha': hp.uniform('alpha', 0.0, 1.0), 'scale_pos_weight': 1.0, 'verbosity': 0, 'objective': 'binary:logitraw', 'nthread': 1, 'eval_metric': 'auc'}
 
     if not apply_weights:
-        PyProphetLearner(infile, outfile, classifier, xgb_hyperparams, xgb_params, xgb_params_space, xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, ss_num_iter, ss_main_score, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, level, ipf_max_peakgroup_rank, ipf_max_peakgroup_pep, ipf_max_transition_isotope_overlap, ipf_min_transition_sn, tric_chromprob, threads, test, ss_score_filter, color_palette, main_score_selection_report).run()
+        PyProphetLearner(infile, outfile, classifier, xgb_hyperparams, xgb_params, xgb_params_space, xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, ss_num_iter, ss_main_score, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, level, add_alignment_features, ipf_max_peakgroup_rank, ipf_max_peakgroup_pep, ipf_max_transition_isotope_overlap, ipf_min_transition_sn, tric_chromprob, threads, test, ss_score_filter, color_palette, main_score_selection_report).run()
     else:
 
-        PyProphetWeightApplier(infile, outfile, classifier, xgb_hyperparams, xgb_params, xgb_params_space, xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, ss_num_iter, ss_main_score, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, level, ipf_max_peakgroup_rank, ipf_max_peakgroup_pep, ipf_max_transition_isotope_overlap, ipf_min_transition_sn, tric_chromprob, threads, test, apply_weights, ss_score_filter, color_palette, main_score_selection_report).run()
+        PyProphetWeightApplier(infile, outfile, classifier, xgb_hyperparams, xgb_params, xgb_params_space, xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, ss_num_iter, ss_main_score, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, level, add_alignment_features, ipf_max_peakgroup_rank, ipf_max_peakgroup_pep, ipf_max_transition_isotope_overlap, ipf_min_transition_sn, tric_chromprob, threads, test, apply_weights, ss_score_filter, color_palette, main_score_selection_report).run()
 
 
 # IPF
@@ -132,7 +133,10 @@ def score(infile, outfile, classifier, xgb_autotune, apply_weights, xeval_fracti
 @click.option('--ipf_max_peakgroup_pep', default=0.7, show_default=True, type=float, help='Maximum PEP to consider scored peak groups in IPF.')
 @click.option('--ipf_max_precursor_peakgroup_pep', default=0.4, show_default=True, type=float, help='Maximum BHM layer 1 integrated precursor peakgroup PEP to consider in IPF.')
 @click.option('--ipf_max_transition_pep', default=0.6, show_default=True, type=float, help='Maximum PEP to consider scored transitions in IPF.')
-def ipf(infile, outfile, ipf_ms1_scoring, ipf_ms2_scoring, ipf_h0, ipf_grouped_fdr, ipf_max_precursor_pep, ipf_max_peakgroup_pep, ipf_max_precursor_peakgroup_pep, ipf_max_transition_pep):
+@click.option('--propagate_signal_across_runs/--no-propagate_signal_across_runs', default=False, show_default=True, help='Propagate signal across runs (requires running alignment).')
+@click.option('--ipf_max_alignment_pep', default=1.0, show_default=True, type=float, help='Maximum PEP to consider for good alignments.')
+@click.option('--across_run_confidence_threshold', default=0.5, show_default=True, type=float, help='Maximum PEP to consider for propagating signal across runs for aligned features.')
+def ipf(infile, outfile, ipf_ms1_scoring, ipf_ms2_scoring, ipf_h0, ipf_grouped_fdr, ipf_max_precursor_pep, ipf_max_peakgroup_pep, ipf_max_precursor_peakgroup_pep, ipf_max_transition_pep, propagate_signal_across_runs, ipf_max_alignment_pep, across_run_confidence_threshold):
     """
     Infer peptidoforms after scoring of MS1, MS2 and transition-level data.
     """
@@ -142,7 +146,7 @@ def ipf(infile, outfile, ipf_ms1_scoring, ipf_ms2_scoring, ipf_h0, ipf_grouped_f
     else:
         outfile = outfile
 
-    infer_peptidoforms(infile, outfile, ipf_ms1_scoring, ipf_ms2_scoring, ipf_h0, ipf_grouped_fdr, ipf_max_precursor_pep, ipf_max_peakgroup_pep, ipf_max_precursor_peakgroup_pep, ipf_max_transition_pep)
+    infer_peptidoforms(infile, outfile, ipf_ms1_scoring, ipf_ms2_scoring, ipf_h0, ipf_grouped_fdr, ipf_max_precursor_pep, ipf_max_peakgroup_pep, ipf_max_precursor_peakgroup_pep, ipf_max_transition_pep, propagate_signal_across_runs, ipf_max_alignment_pep, across_run_confidence_threshold)
 
 
 # Peptide-level inference

pyprophet/main.py

@@ -71,7 +71,7 @@ def calculate_params_for_d_score(classifier, experiment):
 
 class Scorer(object):
 
-    def __init__(self, classifier, score_columns, experiment, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, ss_score_filter, color_palette):
+    def __init__(self, classifier, score_columns, experiment, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, ss_score_filter, color_palette, level):
 
         self.classifier = classifier
         self.score_columns = score_columns
@@ -95,6 +95,7 @@ def __init__(self, classifier, score_columns, experiment, group_id, parametric,
         self.tric_chromprob = tric_chromprob
         self.ss_score_filter = ss_score_filter
         self.color_palette = color_palette
+        self.level = level
 
         target_scores = experiment.get_top_target_peaks()["d_score"]
         decoy_scores = experiment.get_top_decoy_peaks()["d_score"]
@@ -122,7 +123,7 @@ def __init__(self, classifier, score_columns, experiment, group_id, parametric,
 
     def score(self, table):
 
-        prepared_table, __ = prepare_data_table(table, self.ss_score_filter, tg_id_name=self.group_id, score_columns=self.score_columns)
+        prepared_table, __ = prepare_data_table(table, self.ss_score_filter, tg_id_name=self.group_id, score_columns=self.score_columns, level=self.level)
         texp = Experiment(prepared_table)
         score = self.classifier.score(texp, True)
         texp["r_score"] = score
@@ -180,7 +181,7 @@ class HolyGostQuery(object):
         See below how PyProphet parameterises this class.
     """
 
-    def __init__(self, semi_supervised_learner, classifier, ss_num_iter, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, threads, test, ss_score_filter, color_palette):
+    def __init__(self, semi_supervised_learner, classifier, ss_num_iter, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, threads, test, ss_score_filter, color_palette, level):
         assert isinstance(semi_supervised_learner,
                           AbstractSemiSupervisedLearner)
         self.semi_supervised_learner = semi_supervised_learner
@@ -203,9 +204,10 @@ def __init__(self, semi_supervised_learner, classifier, ss_num_iter, group_id, p
         self.test = test
         self.ss_score_filter = ss_score_filter
         self.color_palette = color_palette
+        self.level = level
 
     def _setup_experiment(self, table):
-        prepared_table, score_columns = prepare_data_table(table, self.ss_score_filter, tg_id_name=self.group_id)
+        prepared_table, score_columns = prepare_data_table(table, self.ss_score_filter, tg_id_name=self.group_id, level=self.level)
         experiment = Experiment(prepared_table)
         experiment.log_summary()
         return experiment, score_columns
@@ -339,7 +341,7 @@ def _build_result(self, table, final_classifier, score_columns, experiment):
             for key, value in reversed(sorted(mapped.items(), key=operator.itemgetter(1))):
                 click.echo("Info: Importance of %s: %s" % (key, value))
 
-        scorer = Scorer(final_classifier, score_columns, experiment, self.group_id, self.parametric, self.pfdr, self.pi0_lambda, self.pi0_method, self.pi0_smooth_df, self.pi0_smooth_log_pi0, self.lfdr_truncate, self.lfdr_monotone, self.lfdr_transformation, self.lfdr_adj, self.lfdr_eps, self.tric_chromprob, self.ss_score_filter, self.color_palette)
+        scorer = Scorer(final_classifier, score_columns, experiment, self.group_id, self.parametric, self.pfdr, self.pi0_lambda, self.pi0_method, self.pi0_smooth_df, self.pi0_smooth_log_pi0, self.lfdr_truncate, self.lfdr_monotone, self.lfdr_transformation, self.lfdr_adj, self.lfdr_eps, self.tric_chromprob, self.ss_score_filter, self.color_palette, self.level)
 
         scored_table = scorer.score(table)
 
@@ -354,8 +356,8 @@ def _build_result(self, table, final_classifier, score_columns, experiment):
 @profile
 def PyProphet(classifier, xgb_hyperparams, xgb_params, xgb_params_space, xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, ss_num_iter, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, threads, test, ss_score_filter, color_palette, main_score_selection_report, outfile, level, ss_use_dynamic_main_score):
     if classifier == "LDA":
-        return HolyGostQuery(StandardSemiSupervisedLearner(LDALearner(), xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, test, main_score_selection_report, outfile, level, ss_use_dynamic_main_score), classifier, ss_num_iter, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, threads, test, ss_score_filter, color_palette)
+        return HolyGostQuery(StandardSemiSupervisedLearner(LDALearner(), xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, test, main_score_selection_report, outfile, level, ss_use_dynamic_main_score), classifier, ss_num_iter, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, threads, test, ss_score_filter, color_palette, level)
     elif classifier == "XGBoost":
-        return HolyGostQuery(StandardSemiSupervisedLearner(XGBLearner(xgb_hyperparams, xgb_params, xgb_params_space, threads), xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, test, main_score_selection_report, outfile, level, ss_use_dynamic_main_score), classifier, ss_num_iter, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, threads, test, ss_score_filter, color_palette)
+        return HolyGostQuery(StandardSemiSupervisedLearner(XGBLearner(xgb_hyperparams, xgb_params, xgb_params_space, threads), xeval_fraction, xeval_num_iter, ss_initial_fdr, ss_iteration_fdr, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, test, main_score_selection_report, outfile, level, ss_use_dynamic_main_score), classifier, ss_num_iter, group_id, parametric, pfdr, pi0_lambda, pi0_method, pi0_smooth_df, pi0_smooth_log_pi0, lfdr_truncate, lfdr_monotone, lfdr_transformation, lfdr_adj, lfdr_eps, tric_chromprob, threads, test, ss_score_filter, color_palette, level)
     else:
         raise click.ClickException("Classifier not supported.")