JAL-1793 apply URL decoding when saving VCF attribute values (%3D -> =)

diff --git a/src/jalview/io/vcf/VCFLoader.java b/src/jalview/io/vcf/VCFLoader.java
index 1ac9ad7..d461811 100644 (file)
--- a/src/jalview/io/vcf/VCFLoader.java
+++ b/src/jalview/io/vcf/VCFLoader.java
@@ -22,6 +22,8 @@ import jalview.util.MessageManager;
 
 import java.io.File;
 import java.io.IOException;
+import java.io.UnsupportedEncodingException;
+import java.net.URLDecoder;
 import java.util.ArrayList;
 import java.util.HashMap;
 import java.util.List;
@@ -50,6 +52,8 @@ import htsjdk.variant.vcf.VCFInfoHeaderLine;
  */
 public class VCFLoader
 {
+  private static final String UTF_8 = "UTF-8";
+
   private static final String DEFAULT_SPECIES = "homo_sapiens";
 
   /**
@@ -1189,14 +1193,6 @@ public class VCFLoader
       }
 
       /*
-       * filter out fields we don't want to capture
-       */
-      if (!vcfFieldsOfInterest.contains(key))
-      {
-        continue;
-      }
-
-      /*
        * we extract values for other data which are allele-specific; 
        * these may be per alternate allele (INFO[key].Number = 'A') 
        * or per allele including reference (INFO[key].Number = 'R') 
@@ -1235,6 +1231,16 @@ public class VCFLoader
       String value = getAttributeValue(variant, key, index);
       if (value != null)
       {
+        /*
+         * VCF spec requires encoding of special characters e.g. '='
+         * so decode them here before storing
+         */
+        try
+        {
+          value = URLDecoder.decode(value, UTF_8);
+        } catch (UnsupportedEncodingException e)
+        {
+        }
         sf.setValue(key, value);
       }
     }
@@ -1288,6 +1294,16 @@ public class VCFLoader
             String id = vepFieldsOfInterest.get(i);
             if (id != null)
             {
+              /*
+               * VCF spec requires encoding of special characters e.g. '='
+               * so decode them here before storing
+               */
+              try
+              {
+                field = URLDecoder.decode(field, UTF_8);
+              } catch (UnsupportedEncodingException e)
+              {
+              }
               csqValues.put(id, field);
             }
           }

diff --git a/test/jalview/io/vcf/VCFLoaderTest.java b/test/jalview/io/vcf/VCFLoaderTest.java
index 20cabbd..fb7a4e4 100644 (file)
--- a/test/jalview/io/vcf/VCFLoaderTest.java
+++ b/test/jalview/io/vcf/VCFLoaderTest.java
@@ -504,6 +504,7 @@ public class VCFLoaderTest
     // gene features include Consequence for all transcripts
     Map map = (Map) sf.getValue("CSQ");
     assertEquals(map.size(), 9);
+    assertEquals(map.get("PolyPhen"), "Bad");
 
     sf = geneFeatures.get(1);
     assertEquals(sf.getBegin(), 5);
@@ -513,6 +514,7 @@ public class VCFLoaderTest
     assertEquals(sf.getValue("alleles"), "C,T");
     map = (Map) sf.getValue("CSQ");
     assertEquals(map.size(), 9);
+    assertEquals(map.get("PolyPhen"), "Bad++"); // %3B%3B decoded
 
     sf = geneFeatures.get(2);
     assertEquals(sf.getBegin(), 9);

diff --git a/test/jalview/io/vcf/testVcf.dat b/test/jalview/io/vcf/testVcf.dat
deleted file mode 100644 (file)
index 77e070c..0000000
--- a/test/jalview/io/vcf/testVcf.dat
+++ /dev/null
@@ -1,13 +0,0 @@
-##fileformat=VCFv4.2
-##INFO=<ID=AC,Number=A,Type=Integer,Description="Allele count in genotypes, for each ALT allele, in the same order as listed">
-##INFO=<ID=AF,Number=A,Type=Float,Description="Allele Frequency, for each ALT allele, in the same order as listed">
-##INFO=<ID=AF_Female,Number=R,Type=Float,Description="Allele Frequency among Female genotypes, for each ALT allele, in the same order as listed">
-##INFO=<ID=AN,Number=1,Type=Integer,Description="Total number of alleles in called genotypes">
-##INFO=<ID=CSQ,Number=.,Type=String,Description="Consequence annotations from Ensembl VEP. Format: Allele|Consequence|IMPACT|SYMBOL|Gene|Feature_type|Feature|BIOTYPE|PolyPhen">
-##reference=/Homo_sapiens/GRCh38
-#CHROM POS     ID      REF     ALT     QUAL    FILTER  INFO
-5      45051610        .       C       A       81.96   RF;AC0  AC=1;AF=0.1;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=A|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,A|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad
-5      45051614        .       C       T       1666.64 RF      AC=1;AF=0.2;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=T|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,T|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad
-5      45051618        .       CGG     C       41.94   AC0     AC=1;AF=0.3;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=C|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,C|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad,CSQ=CGT|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,CGT|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad
-5      45051622        .       C       G,T     224.23  RF;AC0  AC=1,2;AF=0.4,0.5;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=G|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,G|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad,T|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,T|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad
-5      45051626        .       A       AC,G    433.35  RF;AC0  AC=3,4;AF=0.6,0.7;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=G|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,G|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad,AC|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,AC|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad

diff --git a/test/jalview/io/vcf/testVcf.vcf b/test/jalview/io/vcf/testVcf.vcf
index 77e070c..8a16a90 100644 (file)
--- a/test/jalview/io/vcf/testVcf.vcf
+++ b/test/jalview/io/vcf/testVcf.vcf
@@ -7,7 +7,7 @@
 ##reference=/Homo_sapiens/GRCh38
 #CHROM POS     ID      REF     ALT     QUAL    FILTER  INFO
 5      45051610        .       C       A       81.96   RF;AC0  AC=1;AF=0.1;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=A|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,A|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad
-5      45051614        .       C       T       1666.64 RF      AC=1;AF=0.2;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=T|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,T|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad
+5      45051614        .       C       T       1666.64 RF      AC=1;AF=0.2;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=T|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,T|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad%2B%2B
 5      45051618        .       CGG     C       41.94   AC0     AC=1;AF=0.3;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=C|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,C|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad,CSQ=CGT|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,CGT|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad
 5      45051622        .       C       G,T     224.23  RF;AC0  AC=1,2;AF=0.4,0.5;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=G|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,G|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad,T|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,T|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad
 5      45051626        .       A       AC,G    433.35  RF;AC0  AC=3,4;AF=0.6,0.7;AN=0;AF_Female=2;AB_MEDIAN=6.00000e-01;CSQ=G|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,G|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad,AC|missense_variant|MODIFIER|WASH7P|gene3|Transcript|transcript3|rna|Benign,AC|downstream_gene_variant|MODIFIER|WASH7P|gene3|Transcript|transcript4|mrna|Bad