From abb658bc53a3a95493eaeaf5d6a380e287699bb7 Mon Sep 17 00:00:00 2001
From: Eduard Kerkhoven <eduardk@chalmers.se>
Date: Wed, 28 Oct 2026 22:37:52 +0100
Subject: [PATCH] getECstring: split EC numbers on semicolons as well as spaces

A database cell that joins codes with ';' and no space, such as '1.1.1.1;2.2.2.2', gave the single string 'EC1.1.1.12.2.2.2'. The cell is now split on white space and semicolons, as geckopy does.
---
 src/get_enzyme_data/getECstring.m     |  6 ++++--
 test/VerificationMatrix.tsv           |  4 ++--
 test/unit_tests/parityDecisionTests.m | 13 +++++++++++++
 3 files changed, 19 insertions(+), 4 deletions(-)

diff --git a/src/get_enzyme_data/getECstring.m b/src/get_enzyme_data/getECstring.m
index 8f4f2e45..e06955d8 100644
--- a/src/get_enzyme_data/getECstring.m
+++ b/src/get_enzyme_data/getECstring.m
@@ -10,7 +10,8 @@
 % EC_set : char
 %     existing EC string to which the EC numbers are appended.
 % ecNumbers : char
-%     space-separated EC numbers associated to the protein.
+%     EC numbers associated to the protein, separated by white space and/or
+%     semicolons.
 %
 % Returns
 % -------
@@ -25,7 +26,8 @@
 % --------
 % findECInDB
 %In case of several ec numbers for the same protein
-new_EC_set = replace(strsplit(ecNumbers,' '),';','');
+new_EC_set = regexp(strtrim(ecNumbers),'[;\s]+','split');
+new_EC_set = new_EC_set(~cellfun(@isempty,new_EC_set));
 for l = 1:length(new_EC_set)
     EC_set = [EC_set 'EC' new_EC_set{l}];
     if l<length(new_EC_set)
diff --git a/test/VerificationMatrix.tsv b/test/VerificationMatrix.tsv
index d14c6743..f673afb8 100644
--- a/test/VerificationMatrix.tsv
+++ b/test/VerificationMatrix.tsv
@@ -43,10 +43,10 @@ gather_kcats/reviewKcatAssignment.m	Has only been tested by running it manually,
 gather_kcats/ecStatsFromBrenda.m	Not verified against real BRENDA data in this environment -- see its own TODO comment
 gather_kcats/mergeKcats.m	tc0240	
 get_enzyme_data/calculateMW.m	Not explicitly tested - The code has existed for a long time	
-get_enzyme_data/findECInDB.m	Called from getECfromDatabase - tested when that function is tested; also tc0242, tc0243, tc0244, tc0245	
+get_enzyme_data/findECInDB.m	Called from getECfromDatabase - tested when that function is tested; also tc0242, tc0243, tc0244, tc0245, tc0246	
 get_enzyme_data/getECfromDatabase.m	tc0007 - does not test download of the databases - this is tested in the manual workflows for Yeast-GEM and Human-GEM	
 get_enzyme_data/getECfromGEM.m	tc0006	
-get_enzyme_data/getECstring.m	Called from findECInDB - tested when that function is tested.	
+get_enzyme_data/getECstring.m	Called from findECInDB - tested when that function is tested; also tc0246	
 download_databases/loadBRENDAdata.m	Called from fuzzyKcatMatching - tested when that function is tested; also tc0105, tc0220, tc0221, tc0222
 download_databases/fetchBrendaAsset.m	tc0085	
 download_databases/loadDatabases.m	Called from getECfromDatabase - tested when that function is tested.
diff --git a/test/unit_tests/parityDecisionTests.m b/test/unit_tests/parityDecisionTests.m
index 0c11fee0..081c4cca 100644
--- a/test/unit_tests/parityDecisionTests.m
+++ b/test/unit_tests/parityDecisionTests.m
@@ -89,6 +89,19 @@ function testFindECInDBGeneWithoutECIsLeftOut_tc0245(testCase)
     verifyEqual(testCase, EC, '1.1.1.4')
 end
 
+function testFindECInDBSplitsCodesOnSemicolons_tc0246(testCase)
+    % A database cell joins codes with white space, with '; ' or with ';' alone;
+    % each is read as separate codes.
+    DBMW = 10000;
+    geneHashMap = containers.Map({'G1'}, {1});
+    geneIndex = {1};
+    cells = {'1.1.1.1 2.2.2.2', '1.1.1.1; 2.2.2.2;', '1.1.1.1;2.2.2.2', ' 1.1.1.1;;2.2.2.2 '};
+    for i = 1:numel(cells)
+        EC = findECInDB({'G1'}, cells(i), DBMW, geneIndex, geneHashMap);
+        verifyEqual(testCase, EC, '1.1.1.1;2.2.2.2')
+    end
+end
+
 function [fuzzy, dlkcat] = originFiveLists()
     fuzzy.rxns        = {'R1'};
     fuzzy.kcats       = 5;