Rename the Variation record to Variant; add a precomputed SNV characteristics search

Model/lib/psql/webready/orgSpecific/ChrCopyNumbers_p.psql

@@ -1,22 +1,28 @@
 :CREATE_AND_POPULATE
+          -- Mirrors the ChrCopyNumbers tuningTable in apiTuningManager.xml; any edit here
+          -- MUST be mirrored there. See the port design doc s7.
+          --
+          -- Two changes: the dead PANIO_p join is gone, and the source is
+          -- GenomicSeqAttributes_p rather than TranscriptAttributes_p. Chromosome ploidy
+          -- is sequence-level, and SequencesByPloidy already reads the sequence table.
           SELECT DISTINCT
-	      ta.project_id
-	    , ta.org_abbrev
-	    , current_timestamp as modification_date
-	    , ta.na_sequence_id
-            , ta.chromosome
+              sa.project_id
+            , sa.org_abbrev
+            , sa.organism
+            , sa.taxon_id
+            , current_timestamp as modification_date
+            , sa.source_id
+            , sa.na_sequence_id
+            , sa.chromosome
             , ccn.chr_copy_number AS ploidy
-            , io.input_pan_id
-            , io.output_pan_id
+            , regexp_replace(pan.name, '_Ploidy$', '') AS eda_sample_stable_id
           FROM apidb.ChrCopyNumber ccn
-            , :SCHEMA.TranscriptAttributes_p ta
-            , :SCHEMA.PANIO_p io
-          WHERE ta.na_sequence_id = ccn.na_sequence_id
-            AND ta.chromosome IS NOT NULL
-            AND ccn.protocol_app_node_id = io.output_pan_id
-	    and ta.org_abbrev = ':ORG_ABBREV'
-	    and io.org_abbrev = ':ORG_ABBREV';
-
-
-:DECLARE_PARTITION;
+            , study.protocolappnode pan
+            , :SCHEMA.GenomicSeqAttributes_p sa
+          WHERE ccn.protocol_app_node_id = pan.protocol_app_node_id
+            AND sa.na_sequence_id = ccn.na_sequence_id
+            AND sa.chromosome IS NOT NULL
+            AND sa.org_abbrev = ':ORG_ABBREV';
+
 
+:DECLARE_PARTITION;

Model/lib/psql/webready/orgSpecific/ChrCopyNumbers_p_ix.psql

@@ -1,9 +1,6 @@
-          CREATE index ChrCN_ix 
-          ON :SCHEMA.ChrCopyNumbers_p (org_abbrev, input_pan_id, na_sequence_id)
+          -- Was two indexes, on (org_abbrev, input_pan_id, na_sequence_id) and
+          -- (org_abbrev, output_pan_id). Both pan columns are now eda_sample_stable_id,
+          -- which would make the second a redundant prefix of the first.
+          CREATE index ChrCN_ix
+          ON :SCHEMA.ChrCopyNumbers_p (org_abbrev, eda_sample_stable_id, na_sequence_id)
     ;
-
-
-          CREATE index ChrCN_output 
-          ON :SCHEMA.ChrCopyNumbers_p (org_abbrev, output_pan_id)
-    ;
-

Model/lib/psql/webready/orgSpecific/GeneCopyNumbers_p.psql

@@ -1,30 +1,59 @@
 :CREATE_AND_POPULATE
+            -- Mirrors the GeneCopyNumbers tuningTable in apiTuningManager.xml, which
+            -- carries the searches until this file's next workflow run. Any edit here
+            -- MUST be mirrored there. See
+            -- docs/superpowers/specs/2026-08-06-genetic-variation-searches-port-design.md s7.
+            --
+            -- The PANIO_p join was removed: study.Input has no rows, so PANIO_p is empty
+            -- and this table came out empty. Organism identity now comes from
+            -- TranscriptAttributes_p and sample identity from the protocolappnode name.
+            --
+            -- ref_cn is computed from the annotation (same ortholog group, same chromosome)
+            -- and deliberately IGNORES apidb.genecopynumber.ref_copy_number, which the
+            -- loader computes per sample and which therefore undercounts for partially
+            -- covered samples. That made ref_cn vary per gene, and hit_medians GROUPs BY it,
+            -- so WDK rejected the answer with a row-count mismatch.
+            WITH grp AS (
+              SELECT DISTINCT ta.gene_source_id, ta.chromosome, oga.group_id
+              FROM :SCHEMA.TranscriptAttributes_p ta
+              JOIN apidb.orthologgroupaasequence oga ON oga.aa_sequence_id = ta.aa_sequence_id
+              WHERE ta.org_abbrev = ':ORG_ABBREV'
+            ), refcn AS (
+              SELECT a.gene_source_id, count(DISTINCT b.gene_source_id) AS ref_cn
+              FROM grp a
+              JOIN grp b ON b.group_id = a.group_id AND b.chromosome = a.chromosome
+              GROUP BY 1
+            )
             SELECT DISTINCT ':PROJECT_ID' as project_id
             , ':ORG_ABBREV' as org_abbrev
+            , ta.organism
+            , ta.taxon_id
             , current_timestamp as modification_date
             , ta.source_id
             , ta.gene_source_id
-            , REGEXP_REPLACE(pan.name, '_[A-Za-z0-9]+ (.+)$', '') AS strain
+            , regexp_replace(pan.name, '_GeneCNV$', '') AS eda_sample_stable_id
             , gcn.haploid_number AS raw_estimate
-            , gcn.ref_copy_number AS ref_cn
+            -- 1 = the gene itself, for a gene in no ortholog group (67 such in pfal).
+            -- CAUTION: ref_cn=1 is also the NORMAL result - 89% of pfal genes have no
+            -- same-chromosome paralog - so if this table is built before the orthomcl
+            -- load has run, it is silently indistinguishable from a correct one for most
+            -- rows. The tuningManager copy guards this with an externalDependency on
+            -- apidb.OrthologGroupAaSequence; here it is a workflow ORDERING requirement.
+            , COALESCE(r.ref_cn, 1) AS ref_cn
             , CASE WHEN (gcn.haploid_number < 0.01) THEN 0
                 WHEN (0.01 < gcn.haploid_number AND gcn.haploid_number < 1.85) THEN 1
                 ELSE round(gcn.haploid_number) END AS haploid_number
             , ta.chromosome
             , ta.na_sequence_id
-            , io.input_pan_id
-            , io.output_pan_id
             FROM apidb.genecopynumber gcn
-            , study.protocolappnode pan
-            , :SCHEMA.TranscriptAttributes_p ta
-            , :SCHEMA.PANIO_p io
-            WHERE gcn.protocol_app_node_id = pan.protocol_app_node_id
-            AND gcn.na_feature_id = ta.gene_na_feature_id
-            AND gcn.protocol_app_node_id = io.output_pan_id
-            AND (ta.gene_type = 'protein coding' or ta.gene_type = 'protein coding gene')
-            AND ta.org_abbrev = ':ORG_ABBREV'
-            AND io.org_abbrev = ':ORG_ABBREV';
-
-
-:DECLARE_PARTITION;
+            JOIN study.protocolappnode pan
+                 ON pan.protocol_app_node_id = gcn.protocol_app_node_id
+            JOIN :SCHEMA.TranscriptAttributes_p ta
+                 ON ta.gene_na_feature_id = gcn.na_feature_id
+            LEFT JOIN refcn r
+                 ON r.gene_source_id = ta.gene_source_id
+            WHERE ta.gene_type IN ('protein coding', 'protein coding gene')
+              AND ta.org_abbrev = ':ORG_ABBREV';
+
 
+:DECLARE_PARTITION;

Model/lib/psql/webready/orgSpecific/GeneCopyNumbers_p_ix.psql

@@ -1,4 +1,3 @@
-            CREATE INDEX GeneCN_ix 
-            ON :SCHEMA.GeneCopyNumbers_p (org_abbrev, input_pan_id, na_sequence_id)
+            CREATE INDEX GeneCN_ix
+            ON :SCHEMA.GeneCopyNumbers_p (org_abbrev, eda_sample_stable_id, na_sequence_id)
     ;
-