MEA with custom kinase-substrate sets (#42)

* Adding the kinase-substrate sets to the results

* Adding a function to run MEA with custom kinase-substrate sets

* Fixing misspelling

* Converting kinase name in custom sets to uppercase for consistency

* Fixing description misspelling

* Bumping version

Author: yaront

src/kinase_library/__init__.py

@@ -23,7 +23,7 @@
 
 #%%
 
-__version__ = "1.2.0"
+__version__ = "1.3.0"
 
 #%% Loading scored phosphoproteome one time per session
 

src/kinase_library/enrichment/mea.py

@@ -136,8 +136,8 @@ def mea(self, kin_type, kl_method, kl_thresh,
 
         Returns
         -------
-        enrichemnt_results : pd.DataFrame
-            pd.Dataframe with results of MEA for the specified KL method and threshold.
+        enrichment_results : pd.DataFrame
+            pd.DataFrame with results of MEA for the specified KL method and threshold.
         """
 
         exceptions.check_kl_method(kl_method)
@@ -197,11 +197,86 @@ def mea(self, kin_type, kl_method, kl_thresh,
         enrichment_data['FDR'] = enrichment_data['FDR'].replace(0,enrichment_data['FDR'][enrichment_data['FDR'] != 0].min()).astype(float) #Setting FDR of zero to lowest FDR in data
         sorted_enrichment_data = enrichment_data.sort_values('Kinase').set_index('Kinase').reindex(data.get_kinase_list(kin_type, non_canonical=non_canonical))
 
-        enrichemnt_results = MeaEnrichmentResults(enrichment_results=sorted_enrichment_data, pps_data=self, gseapy_obj=prerank_results,
+        enrichment_results = MeaEnrichmentResults(enrichment_results=sorted_enrichment_data, pps_data=self, kin_sub_sets=kin_sub_sets, gseapy_obj=prerank_results,
                                                    kin_type=kin_type, kl_method=kl_method, kl_thresh=kl_thresh, tested_kins=kinases,
                                                    data_att=data_att, kl_comp_direction=kl_comp_direction)
 
-        return enrichemnt_results
+        return enrichment_results
+
+
+    def mea_custom(self, custom_kin_sets,
+                   kinases=None, kin_type='custom',
+                   weight=1, threads=4, min_size=1, max_size=100000,
+                   permutation_num=1000, seed=112123,
+                   gseapy_verbose=False):
+        """
+        Kinase enrichment analysis based on pre-ranked GSEA substrates list using custom kinase-substrate sets.
+
+        Parameters
+        ----------
+        custom_kin_sets : dict
+            A dictionary of custom kinase-substrate sets where keys are kinase names and values are lists of substrates.
+        kinases : list, optional
+            If provided, kinase enrichment will only be calculated for the specified kinases from custom_kin_sets. The default is None, which uses all kinases in custom_kin_sets.
+        kin_type : str, optional
+            A label to identify the type of custom kinase sets being used. The default is 'custom'.
+        **GSEApy parameters: weight, threads, min_size, max_size, permutation_num, seed, gseapy_verbose
+
+        Returns
+        -------
+        enrichment_results : pd.DataFrame
+            pd.DataFrame with results of MEA for the custom kinase-substrate sets.
+        """
+
+        if not isinstance(custom_kin_sets, dict) or not custom_kin_sets:
+            raise ValueError('custom_kin_sets must be a non-empty dictionary with kinase names as keys and substrate lists as values.')
+
+        custom_kin_sets = {k.upper(): v for k, v in custom_kin_sets.items()}
+
+        if kinases is None:
+            kinases = list(custom_kin_sets.keys())
+        elif isinstance(kinases, str):
+            kinases = [kinases]
+
+        kinases = [x.upper() for x in kinases]
+
+        filtered_kin_sets = {k: v for k, v in custom_kin_sets.items() if k in kinases}
+        if not filtered_kin_sets:
+            raise ValueError('No kinases from the provided list were found in custom_kin_sets.')
+
+        ranked_subs = self.dp_data_pps.data.set_index(_global_vars.default_seq_col)[self.rank_col].sort_values(ascending=False)
+
+        prerank_results = gp.prerank(rnk=ranked_subs,
+                             gene_sets=filtered_kin_sets,
+                             weight=weight,
+                             threads=threads,
+                             min_size=min_size,
+                             max_size=max_size,
+                             permutation_num=permutation_num,
+                             seed=seed,
+                             verbose=gseapy_verbose)
+
+        res_col_converter = {'Term': 'Kinase', 'ES': 'ES', 'NES': 'NES', 'NOM p-val': 'p-value', 'FDR q-val': 'FDR', 'Tag %': 'Subs fraction', 'Lead_genes': 'Leading substrates'}
+
+        enrichment_data = prerank_results.res2d.drop(['Name', 'FWER p-val', 'Gene %'], axis=1).rename(columns=res_col_converter)
+        enrichment_data['p-value'] = enrichment_data['p-value'].replace(0,1/permutation_num).astype(float) #Setting p-value of zero to 1/(# of permutations)
+        enrichment_data['FDR'] = enrichment_data['FDR'].replace(0,enrichment_data['FDR'][enrichment_data['FDR'] != 0].min()).astype(float) #Setting FDR of zero to lowest FDR in data
+        sorted_enrichment_data = enrichment_data.sort_values('Kinase').set_index('Kinase')
+
+        enrichment_results = MeaEnrichmentResults(
+            enrichment_results=sorted_enrichment_data,
+            pps_data=self,
+            kin_sub_sets=filtered_kin_sets,
+            gseapy_obj=prerank_results,
+            kin_type=kin_type,
+            kl_method='custom',
+            kl_thresh=None,
+            tested_kins=kinases,
+            data_att='custom',
+            kl_comp_direction=None
+        )
+
+        return enrichment_results
 
 #%%
 
@@ -215,6 +290,8 @@ class MeaEnrichmentResults(object):
         Dataframe containing Kinase Library enrichment results.
     pps_data : kl.EnrichmentData
         Object initialized from the foreground and background dataframes used to calculate provided enrichment_results.
+    kin_sub_sets : dict
+        Kinase-substrate sets used for the enrichment.
     kin_type : str
         Kinase type ('ser_thr' or 'tyrosine').
     kl_method : str
@@ -231,12 +308,13 @@ class MeaEnrichmentResults(object):
         Dictates if kinases above or below the specified threshold are used ('higher','lower').
     """
 
-    def __init__(self, enrichment_results, pps_data, gseapy_obj,
-                 kin_type, kl_method, kl_thresh, tested_kins,
-                 data_att, kl_comp_direction):
+    def __init__(self, enrichment_results, pps_data, kin_sub_sets,
+                 gseapy_obj, kin_type, kl_method, kl_thresh,
+                 tested_kins, data_att, kl_comp_direction):
 
         self.enrichment_results = enrichment_results
         self.pps_data = pps_data
+        self.kin_sub_sets = kin_sub_sets
         self.gseapy_obj = gseapy_obj
         self.kin_type = kin_type
         self.kl_method = kl_method