Rfc80/fix patient level filtering for alteration counts endpoints (#1…

…0857) * Add patient_id column to genomic_event_derived * Update sql to convert list of patients to list of samples
cBioPortal · Jun 24, 2024 · 14b7105 · 14b7105
1 parent a484e2e
commit 14b7105
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 8 deletions.
diff --git a/src/main/resources/db-scripts/clickhouse/clickhouse.sql b/src/main/resources/db-scripts/clickhouse/clickhouse.sql
@@ -17,7 +17,8 @@ CREATE TABLE IF NOT EXISTS genomic_event_derived
     driver_tiers_filter       LowCardinality(String),
     cna_alteration            Nullable(Int8),
     cna_cytoband              String,
-    sv_event_info             String
+    sv_event_info             String,
+    patient_unique_id         String
 ) ENGINE = MergeTree
 ORDER BY ( variant_type, hugo_gene_symbol, genetic_profile_stable_id, sample_unique_id);
 

diff --git a/src/main/resources/db-scripts/clickhouse/clickhouse_migration.sql b/src/main/resources/db-scripts/clickhouse/clickhouse_migration.sql
@@ -40,7 +40,8 @@ SELECT concat(cs.cancer_study_identifier, '_', sample.stable_id) AS sample_uniqu
        'NA'                                                      AS drivet_tiers_filter,
        NULL                                                      AS cna_alteration,
        ''                                                        AS cna_cytoband,
-       ''                                                        AS sv_event_info
+       ''                                                        AS sv_event_info,
+       concat(cs.cancer_study_identifier, '_', patient.stable_id) AS patient_unique_id
 FROM mutation
          INNER JOIN mutation_event AS me ON mutation.mutation_event_id = me.mutation_event_id
          INNER JOIN sample_profile sp
@@ -49,6 +50,7 @@ FROM mutation
          LEFT JOIN genetic_profile g ON sp.genetic_profile_id = g.genetic_profile_id
          INNER JOIN cancer_study cs ON g.cancer_study_id = cs.cancer_study_id
          INNER JOIN sample ON mutation.sample_id = sample.internal_id
+         INNER JOIN patient on sample.patient_id = patient.internal_id
          LEFT JOIN gene ON mutation.entrez_gene_id = gene.entrez_gene_id
 UNION ALL
 SELECT concat(cs.cancer_study_identifier, '_', sample.stable_id) AS sample_unique_id,
@@ -64,14 +66,16 @@ SELECT concat(cs.cancer_study_identifier, '_', sample.stable_id) AS sample_uniqu
        'NA'                                                      AS drivet_tiers_filter,
        ce.alteration                                             AS cna_alteration,
        rgg.cytoband                                              AS cna_cytoband,
-       ''                                                        AS sv_event_info
+       ''                                                        AS sv_event_info,
+       concat(cs.cancer_study_identifier, '_', patient.stable_id) AS patient_unique_id
 FROM cna_event ce
          INNER JOIN sample_cna_event sce ON ce.cna_event_id = sce.cna_event_id
          INNER JOIN sample_profile sp ON sce.sample_id = sp.sample_id AND sce.genetic_profile_id = sp.genetic_profile_id
          LEFT JOIN gene_panel gp ON sp.panel_id = gp.internal_id
          INNER JOIN genetic_profile g ON sp.genetic_profile_id = g.genetic_profile_id
          INNER JOIN cancer_study cs ON g.cancer_study_id = cs.cancer_study_id
          INNER JOIN sample ON sce.sample_id = sample.internal_id
+         INNER JOIN patient on sample.patient_id = patient.internal_id
          INNER JOIN gene ON ce.entrez_gene_id = gene.entrez_gene_id
          INNER JOIN reference_genome_gene rgg ON rgg.entrez_gene_id = ce.entrez_gene_id
 UNION ALL
@@ -88,10 +92,12 @@ SELECT concat(cs.cancer_study_identifier, '_', s.stable_id) AS sample_unique_id,
        'NA'                                                 AS drivet_tiers_filter,
        NULL                                                 AS cna_alteration,
        ''                                                   AS cna_cytoband,
-       event_info                                           AS sv_event_info
+       event_info                                           AS sv_event_info,
+       concat(cs.cancer_study_identifier, '_', patient.stable_id) AS patient_unique_id
 FROM structural_variant sv
          INNER JOIN genetic_profile gp ON sv.genetic_profile_id = gp.genetic_profile_id
          INNER JOIN sample s ON sv.sample_id = s.internal_id
+         INNER JOIN patient on s.patient_id = patient.internal_id
          INNER JOIN cancer_study cs ON gp.cancer_study_id = cs.cancer_study_id
          INNER JOIN gene ON sv.site1_entrez_gene_id = gene.entrez_gene_id
          INNER JOIN sample_profile ON s.internal_id = sample_profile.sample_id AND sample_profile.genetic_profile_id = sv.genetic_profile_id
@@ -110,10 +116,12 @@ SELECT concat(cs.cancer_study_identifier, '_', s.stable_id) AS sample_unique_id,
        'NA'                                                 AS drivet_tiers_filter,
        NULL                                                 AS cna_alteration,
        ''                                                   AS cna_cytoband,
-       event_info                                           AS sv_event_info
+       event_info                                           AS sv_event_info,
+       concat(cs.cancer_study_identifier, '_', patient.stable_id) AS patient_unique_id
 FROM structural_variant sv
          INNER JOIN genetic_profile gp ON sv.genetic_profile_id = gp.genetic_profile_id
          INNER JOIN sample s ON sv.sample_id = s.internal_id
+         INNER JOIN patient on s.patient_id = patient.internal_id
          INNER JOIN cancer_study cs ON gp.cancer_study_id = cs.cancer_study_id
          INNER JOIN gene ON sv.site2_entrez_gene_id = gene.entrez_gene_id
          INNER JOIN sample_profile ON s.internal_id = sample_profile.sample_id AND sample_profile.genetic_profile_id = sv.genetic_profile_id

diff --git a/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewMapper.xml b/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewMapper.xml
@@ -281,7 +281,11 @@
         <where>
             stgp.alteration_type = '${alterationType}'
             AND
-            <include refid="applyStudyViewFilter"/>
+            sample_unique_id IN ( <include refid="sampleUniqueIdsFromStudyViewFilter"/>
+            <if test="applyPatientIdFilters == true">
+                INTERSECT <include refid="getSampleIdsFromPatientIds"/>
+            </if>
+            )
         </where>
     </select>
 
@@ -306,8 +310,13 @@
                 select distinct gene_panel_id
                 from sample_to_gene_panel
                 <where>
-                    alteration_type = '${alterationType}' AND
-                    <include refid="applyStudyViewFilter"/>
+                    alteration_type = '${alterationType}'
+                    AND
+                    sample_unique_id IN ( <include refid="sampleUniqueIdsFromStudyViewFilter"/>
+                    <if test="applyPatientIdFilters == true">
+                        INTERSECT <include refid="getSampleIdsFromPatientIds"/>
+                    </if>
+                    )
                 </where>
             )
         </where>