filters)
+ {
+ for (Pattern p : filters)
+ {
+ if (p.matcher(id.toUpperCase()).matches())
+ {
+ return true;
+ }
+ }
+ return false;
+ }
+
+ /**
+ * Records 'wanted' fields defined in the CSQ INFO header (if there is one).
+ * Also records the position of selected fields (Allele, ALLELE_NUM, Feature)
+ * required for processing.
+ *
+ * CSQ fields are declared in the CSQ INFO Description e.g.
*
* Description="Consequence ...from ... VEP. Format: Allele|Consequence|...
*/
- protected void locateCsqFields()
+ protected void parseCsqHeader()
{
- VCFInfoHeaderLine csqInfo = header.getInfoHeaderLine(CSQ);
+ List vepFieldFilters = getFieldMatchers(VEP_FIELDS_PREF,
+ DEFAULT_VEP_FIELDS);
+ vepFieldsOfInterest = new HashMap<>();
+
+ VCFInfoHeaderLine csqInfo = header.getInfoHeaderLine(CSQ_FIELD);
if (csqInfo == null)
{
return;
}
+ /*
+ * parse out the pipe-separated list of CSQ fields; we assume here that
+ * these form the last part of the description, and contain no spaces
+ */
String desc = csqInfo.getDescription();
- int formatPos = desc.indexOf(FORMAT);
- if (formatPos == -1)
- {
- System.err.println("Parse error, failed to find " + FORMAT
- + " in " + desc);
- return;
- }
- desc = desc.substring(formatPos + FORMAT.length());
+ int spacePos = desc.lastIndexOf(" ");
+ desc = desc.substring(spacePos + 1);
if (desc != null)
{
@@ -268,26 +427,69 @@ public class VCFLoader
int index = 0;
for (String field : format)
{
- if (ALLELE_NUM_KEY.equals(field))
+ if (CSQ_CONSEQUENCE_KEY.equals(field))
+ {
+ csqConsequenceFieldIndex = index;
+ }
+ if (CSQ_ALLELE_NUM_KEY.equals(field))
{
csqAlleleNumberFieldIndex = index;
}
- if (ALLELE_KEY.equals(field))
+ if (CSQ_ALLELE_KEY.equals(field))
{
csqAlleleFieldIndex = index;
}
- if (FEATURE_KEY.equals(field))
+ if (CSQ_FEATURE_KEY.equals(field))
{
csqFeatureFieldIndex = index;
}
+
+ if (isFieldWanted(field, vepFieldFilters))
+ {
+ vepFieldsOfInterest.put(index, field);
+ }
+
index++;
}
}
}
/**
+ * Reads the Preference value for the given key, with default specified if no
+ * preference set. The value is interpreted as a comma-separated list of
+ * regular expressions, and converted into a list of compiled patterns ready
+ * for matching. Patterns are forced to upper-case for non-case-sensitive
+ * matching.
+ *
+ * This supports user-defined filters for fields of interest to capture while
+ * processing data. For example, VCF_FIELDS = AF,AC* would mean that VCF INFO
+ * fields with an ID of AF, or starting with AC, would be matched.
+ *
+ * @param key
+ * @param def
+ * @return
+ */
+ private List getFieldMatchers(String key, String def)
+ {
+ String pref = Cache.getDefault(key, def);
+ List patterns = new ArrayList<>();
+ String[] tokens = pref.split(",");
+ for (String token : tokens)
+ {
+ try
+ {
+ patterns.add(Pattern.compile(token.toUpperCase()));
+ } catch (PatternSyntaxException e)
+ {
+ System.err.println("Invalid pattern ignored: " + token);
+ }
+ }
+ return patterns;
+ }
+
+ /**
* Transfers VCF features to sequences to which this sequence has a mapping.
- * If the mapping is 1:3, computes peptide variants from nucleotide variants.
+ * If the mapping is 3:1, computes peptide variants from nucleotide variants.
*
* @param seq
*/
@@ -336,125 +538,260 @@ public class VCFLoader
/**
* Tries to add overlapping variants read from a VCF file to the given
* sequence, and returns the number of variant features added. Note that this
- * requires the sequence to hold information as to its chromosomal positions
- * and reference, in order to be able to map the VCF variants to the sequence.
+ * requires the sequence to hold information as to its species, chromosomal
+ * positions and reference assembly, in order to be able to map the VCF
+ * variants to the sequence (or not)
*
* @param seq
* @param reader
- * @param isVcfRefGrch37
+ * @param vcfAssembly
* @return
*/
protected int loadSequenceVCF(SequenceI seq, VCFReader reader,
- boolean isVcfRefGrch37)
+ String vcfAssembly)
{
- int count = 0;
- GeneLociI seqCoords = seq.getGeneLoci();
- if (seqCoords == null)
+ VCFMap vcfMap = getVcfMap(seq, vcfAssembly);
+ if (vcfMap == null)
{
return 0;
}
- List seqChromosomalContigs = seqCoords.getMap().getToRanges();
- for (int[] range : seqChromosomalContigs)
+ /*
+ * work with the dataset sequence here
+ */
+ SequenceI dss = seq.getDatasetSequence();
+ if (dss == null)
{
- count += addVcfVariants(seq, reader, range, isVcfRefGrch37);
+ dss = seq;
}
-
- return count;
+ return addVcfVariants(dss, reader, vcfMap, vcfAssembly);
}
/**
- * Queries the VCF reader for any variants that overlap the given chromosome
- * region of the sequence, and adds as variant features. Returns the number of
- * overlapping variants found.
+ * Answers a map from sequence coordinates to VCF chromosome ranges
*
* @param seq
- * @param reader
- * @param range
- * start-end range of a sequence region in its chromosomal
- * coordinates
- * @param isVcfRefGrch37
- * true if the VCF is with reference to GRCh37
+ * @param vcfAssembly
* @return
*/
- protected int addVcfVariants(SequenceI seq, VCFReader reader,
- int[] range, boolean isVcfRefGrch37)
+ private VCFMap getVcfMap(SequenceI seq, String vcfAssembly)
{
+ /*
+ * simplest case: sequence has id and length matching a VCF contig
+ */
+ VCFMap vcfMap = null;
+ if (dictionary != null)
+ {
+ vcfMap = getContigMap(seq);
+ }
+ if (vcfMap != null)
+ {
+ return vcfMap;
+ }
+
+ /*
+ * otherwise, map to VCF from chromosomal coordinates
+ * of the sequence (if known)
+ */
GeneLociI seqCoords = seq.getGeneLoci();
+ if (seqCoords == null)
+ {
+ Cache.log.warn(String.format(
+ "Can't query VCF for %s as chromosome coordinates not known",
+ seq.getName()));
+ return null;
+ }
+ String species = seqCoords.getSpeciesId();
String chromosome = seqCoords.getChromosomeId();
String seqRef = seqCoords.getAssemblyId();
- String species = seqCoords.getSpeciesId();
+ MapList map = seqCoords.getMap();
+
+ if (!vcfSpeciesMatchesSequence(vcfAssembly, species))
+ {
+ return null;
+ }
+
+ if (vcfAssemblyMatchesSequence(vcfAssembly, seqRef))
+ {
+ return new VCFMap(chromosome, map);
+ }
+
+ if (!"GRCh38".equalsIgnoreCase(seqRef) // Ensembl
+ || !vcfAssembly.contains("Homo_sapiens_assembly19")) // gnomAD
+ {
+ return null;
+ }
/*
- * map chromosomal coordinates from GRCh38 (sequence) to
- * GRCh37 (VCF) if necessary
+ * map chromosomal coordinates from sequence to VCF if the VCF
+ * data has a different reference assembly to the sequence
*/
- // TODO generalise for other assemblies and species
- int offset = 0;
- String fromRef = "GRCh38";
- if (fromRef.equalsIgnoreCase(seqRef) && isVcfRefGrch37)
- {
- String toRef = "GRCh37";
- int[] newRange = mapReferenceRange(range, chromosome, "human",
- fromRef, toRef);
+ // TODO generalise for cases other than GRCh38 -> GRCh37 !
+ // - or get the user to choose in a dialog
+
+ List toVcfRanges = new ArrayList<>();
+ List fromSequenceRanges = new ArrayList<>();
+ String toRef = "GRCh37";
+
+ for (int[] range : map.getToRanges())
+ {
+ int[] fromRange = map.locateInFrom(range[0], range[1]);
+ if (fromRange == null)
+ {
+ // corrupted map?!?
+ continue;
+ }
+
+ int[] newRange = mapReferenceRange(range, chromosome, "human", seqRef,
+ toRef);
if (newRange == null)
{
- System.err.println(String.format(
- "Failed to map %s:%s:%s:%d:%d to %s", species, chromosome,
- fromRef, range[0], range[1], toRef));
- return 0;
+ Cache.log.error(
+ String.format("Failed to map %s:%s:%s:%d:%d to %s", species,
+ chromosome, seqRef, range[0], range[1], toRef));
+ continue;
+ }
+ else
+ {
+ toVcfRanges.add(newRange);
+ fromSequenceRanges.add(fromRange);
+ }
+ }
+
+ return new VCFMap(chromosome,
+ new MapList(fromSequenceRanges, toVcfRanges, 1, 1));
+ }
+
+ /**
+ * If the sequence id matches a contig declared in the VCF file, and the
+ * sequence length matches the contig length, then returns a 1:1 map of the
+ * sequence to the contig, else returns null
+ *
+ * @param seq
+ * @return
+ */
+ private VCFMap getContigMap(SequenceI seq)
+ {
+ String id = seq.getName();
+ SAMSequenceRecord contig = dictionary.getSequence(id);
+ if (contig != null)
+ {
+ int len = seq.getLength();
+ if (len == contig.getSequenceLength())
+ {
+ MapList map = new MapList(new int[] { 1, len },
+ new int[]
+ { 1, len }, 1, 1);
+ return new VCFMap(id, map);
}
- offset = newRange[0] - range[0];
- range = newRange;
}
+ return null;
+ }
+
+ /**
+ * Answers true if we determine that the VCF data uses the same reference
+ * assembly as the sequence, else false
+ *
+ * @param vcfAssembly
+ * @param seqRef
+ * @return
+ */
+ private boolean vcfAssemblyMatchesSequence(String vcfAssembly,
+ String seqRef)
+ {
+ // TODO improve on this stub, which handles gnomAD and
+ // hopes for the best for other cases
+
+ if ("GRCh38".equalsIgnoreCase(seqRef) // Ensembl
+ && vcfAssembly.contains("Homo_sapiens_assembly19")) // gnomAD
+ {
+ return false;
+ }
+ return true;
+ }
- boolean forwardStrand = range[0] <= range[1];
+ /**
+ * Answers true if the species inferred from the VCF reference identifier
+ * matches that for the sequence
+ *
+ * @param vcfAssembly
+ * @param speciesId
+ * @return
+ */
+ boolean vcfSpeciesMatchesSequence(String vcfAssembly, String speciesId)
+ {
+ // PROBLEM 1
+ // there are many aliases for species - how to equate one with another?
+ // PROBLEM 2
+ // VCF ##reference header is an unstructured URI - how to extract species?
+ // perhaps check if ref includes any (Ensembl) alias of speciesId??
+ // TODO ask the user to confirm this??
+
+ if (vcfAssembly.contains("Homo_sapiens") // gnomAD exome data example
+ && "HOMO_SAPIENS".equals(speciesId)) // Ensembl species id
+ {
+ return true;
+ }
+
+ if (vcfAssembly.contains("c_elegans") // VEP VCF response example
+ && "CAENORHABDITIS_ELEGANS".equals(speciesId)) // Ensembl
+ {
+ return true;
+ }
+
+ // this is not a sustainable solution...
+
+ return false;
+ }
+
+ /**
+ * Queries the VCF reader for any variants that overlap the mapped chromosome
+ * ranges of the sequence, and adds as variant features. Returns the number of
+ * overlapping variants found.
+ *
+ * @param seq
+ * @param reader
+ * @param map
+ * mapping from sequence to VCF coordinates
+ * @param vcfAssembly
+ * the '##reference' identifier for the VCF reference assembly
+ * @return
+ */
+ protected int addVcfVariants(SequenceI seq, VCFReader reader,
+ VCFMap map, String vcfAssembly)
+ {
+ boolean forwardStrand = map.map.isToForwardStrand();
/*
- * query the VCF for overlaps
- * (convert a reverse strand range to forwards)
+ * query the VCF for overlaps of each contiguous chromosomal region
*/
int count = 0;
- MapList mapping = seqCoords.getMap();
- int fromLocus = Math.min(range[0], range[1]);
- int toLocus = Math.max(range[0], range[1]);
- CloseableIterator variants = reader.query(chromosome,
- fromLocus, toLocus);
- while (variants.hasNext())
+ for (int[] range : map.map.getToRanges())
{
- /*
- * get variant location in sequence chromosomal coordinates
- */
- VariantContext variant = variants.next();
-
- /*
- * we can only process SNP variants (which can be reported
- * as part of a MIXED variant record
- */
- if (!variant.isSNP() && !variant.isMixed())
+ int vcfStart = Math.min(range[0], range[1]);
+ int vcfEnd = Math.max(range[0], range[1]);
+ CloseableIterator variants = reader
+ .query(map.chromosome, vcfStart, vcfEnd);
+ while (variants.hasNext())
{
- // continue;
- }
+ VariantContext variant = variants.next();
- int start = variant.getStart() - offset;
- int end = variant.getEnd() - offset;
+ int[] featureRange = map.map.locateInFrom(variant.getStart(),
+ variant.getEnd());
- /*
- * convert chromosomal location to sequence coordinates
- * - null if a partially overlapping feature
- */
- int[] seqLocation = mapping.locateInFrom(start, end);
- if (seqLocation != null)
- {
- count += addAlleleFeatures(seq, variant, seqLocation[0],
- seqLocation[1], forwardStrand);
+ if (featureRange != null)
+ {
+ int featureStart = Math.min(featureRange[0], featureRange[1]);
+ int featureEnd = Math.max(featureRange[0], featureRange[1]);
+ count += addAlleleFeatures(seq, variant, featureStart, featureEnd,
+ forwardStrand);
+ }
}
+ variants.close();
}
- variants.close();
-
return count;
}
@@ -541,9 +878,9 @@ public class VCFLoader
/**
* Inspects one allele and attempts to add a variant feature for it to the
- * sequence. We extract as much as possible of the additional data associated
- * with this allele to store in the feature's key-value map. Answers the
- * number of features added (0 or 1).
+ * sequence. The additional data associated with this allele is extracted to
+ * store in the feature's key-value map. Answers the number of features added (0
+ * or 1).
*
* @param seq
* @param variant
@@ -561,34 +898,60 @@ public class VCFLoader
String reference = variant.getReference().getBaseString();
Allele alt = variant.getAlternateAllele(altAlleleIndex);
String allele = alt.getBaseString();
- if (allele.length() != 1)
+
+ /*
+ * insertion after a genomic base, if on reverse strand, has to be
+ * converted to insertion of complement after the preceding position
+ */
+ int referenceLength = reference.length();
+ if (!forwardStrand && allele.length() > referenceLength
+ && allele.startsWith(reference))
{
- /*
- * not a SNP variant
- */
- // return 0;
+ featureStart -= referenceLength;
+ featureEnd = featureStart;
+ char insertAfter = seq.getCharAt(featureStart - seq.getStart());
+ reference = Dna.reverseComplement(String.valueOf(insertAfter));
+ allele = allele.substring(referenceLength) + reference;
}
/*
* build the ref,alt allele description e.g. "G,A", using the base
* complement if the sequence is on the reverse strand
*/
- // TODO check how structural variants are shown on reverse strand
StringBuilder sb = new StringBuilder();
sb.append(forwardStrand ? reference : Dna.reverseComplement(reference));
sb.append(COMMA);
sb.append(forwardStrand ? allele : Dna.reverseComplement(allele));
String alleles = sb.toString(); // e.g. G,A
+ /*
+ * pick out the consequence data (if any) that is for the current allele
+ * and feature (transcript) that matches the current sequence
+ */
+ String consequence = getConsequenceForAlleleAndFeature(variant, CSQ_FIELD,
+ altAlleleIndex, csqAlleleFieldIndex,
+ csqAlleleNumberFieldIndex, seq.getName().toLowerCase(),
+ csqFeatureFieldIndex);
+
+ /*
+ * pick out the ontology term for the consequence type
+ */
String type = SequenceOntologyI.SEQUENCE_VARIANT;
+ if (consequence != null)
+ {
+ type = getOntologyTerm(seq, variant, altAlleleIndex,
+ consequence);
+ }
+
float score = getAlleleFrequency(variant, altAlleleIndex);
SequenceFeature sf = new SequenceFeature(type, alleles, featureStart,
featureEnd, score, FEATURE_GROUP_VCF);
+ sf.setSource(sourceId);
sf.setValue(Gff3Helper.ALLELES, alleles);
- addAlleleProperties(variant, seq, sf, altAlleleIndex);
+ addAlleleProperties(variant, seq, sf, altAlleleIndex, consequence);
seq.addSequenceFeature(sf);
@@ -596,6 +959,165 @@ public class VCFLoader
}
/**
+ * Determines the Sequence Ontology term to use for the variant feature type in
+ * Jalview. The default is 'sequence_variant', but a more specific term is used
+ * if:
+ *
+ * - VEP (or SnpEff) Consequence annotation is included in the VCF
+ * - sequence id can be matched to VEP Feature (or SnpEff Feature_ID)
+ *
+ *
+ * @param seq
+ * @param variant
+ * @param altAlleleIndex
+ * @param consequence
+ * @return
+ * @see http://www.sequenceontology.org/browser/current_svn/term/SO:0001060
+ */
+ String getOntologyTerm(SequenceI seq, VariantContext variant,
+ int altAlleleIndex, String consequence)
+ {
+ String type = SequenceOntologyI.SEQUENCE_VARIANT;
+
+ if (csqAlleleFieldIndex == -1) // && snpEffAlleleFieldIndex == -1
+ {
+ /*
+ * no Consequence data so we can't refine the ontology term
+ */
+ return type;
+ }
+
+ /*
+ * can we associate Consequence data with this allele and feature (transcript)?
+ * if so, prefer the consequence term from that data
+ */
+ if (consequence != null)
+ {
+ String[] csqFields = consequence.split(PIPE_REGEX);
+ if (csqFields.length > csqConsequenceFieldIndex)
+ {
+ type = csqFields[csqConsequenceFieldIndex];
+ }
+ }
+ else
+ {
+ // todo the same for SnpEff consequence data matching if wanted
+ }
+
+ /*
+ * if of the form (e.g.) missense_variant&splice_region_variant,
+ * just take the first ('most severe') consequence
+ */
+ if (type != null)
+ {
+ int pos = type.indexOf('&');
+ if (pos > 0)
+ {
+ type = type.substring(0, pos);
+ }
+ }
+ return type;
+ }
+
+ /**
+ * Returns matched consequence data if it can be found, else null.
+ *
+ * - inspects the VCF data for key 'vcfInfoId'
+ * - splits this on comma (to distinct consequences)
+ * - returns the first consequence (if any) where
+ *
+ * - the allele matches the altAlleleIndex'th allele of variant
+ * - the feature matches the sequence name (e.g. transcript id)
+ *
+ *
+ * If matched, the consequence is returned (as pipe-delimited fields).
+ *
+ * @param variant
+ * @param vcfInfoId
+ * @param altAlleleIndex
+ * @param alleleFieldIndex
+ * @param alleleNumberFieldIndex
+ * @param seqName
+ * @param featureFieldIndex
+ * @return
+ */
+ private String getConsequenceForAlleleAndFeature(VariantContext variant,
+ String vcfInfoId, int altAlleleIndex, int alleleFieldIndex,
+ int alleleNumberFieldIndex,
+ String seqName, int featureFieldIndex)
+ {
+ if (alleleFieldIndex == -1 || featureFieldIndex == -1)
+ {
+ return null;
+ }
+ Object value = variant.getAttribute(vcfInfoId);
+
+ if (value == null || !(value instanceof List>))
+ {
+ return null;
+ }
+
+ /*
+ * inspect each consequence in turn (comma-separated blocks
+ * extracted by htsjdk)
+ */
+ List consequences = (List) value;
+
+ for (String consequence : consequences)
+ {
+ String[] csqFields = consequence.split(PIPE_REGEX);
+ if (csqFields.length > featureFieldIndex)
+ {
+ String featureIdentifier = csqFields[featureFieldIndex];
+ if (featureIdentifier.length() > 4
+ && seqName.indexOf(featureIdentifier.toLowerCase()) > -1)
+ {
+ /*
+ * feature (transcript) matched - now check for allele match
+ */
+ if (matchAllele(variant, altAlleleIndex, csqFields,
+ alleleFieldIndex, alleleNumberFieldIndex))
+ {
+ return consequence;
+ }
+ }
+ }
+ }
+ return null;
+ }
+
+ private boolean matchAllele(VariantContext variant, int altAlleleIndex,
+ String[] csqFields, int alleleFieldIndex,
+ int alleleNumberFieldIndex)
+ {
+ /*
+ * if ALLELE_NUM is present, it must match altAlleleIndex
+ * NB first alternate allele is 1 for ALLELE_NUM, 0 for altAlleleIndex
+ */
+ if (alleleNumberFieldIndex > -1)
+ {
+ if (csqFields.length <= alleleNumberFieldIndex)
+ {
+ return false;
+ }
+ String alleleNum = csqFields[alleleNumberFieldIndex];
+ return String.valueOf(altAlleleIndex + 1).equals(alleleNum);
+ }
+
+ /*
+ * else consequence allele must match variant allele
+ */
+ if (alleleFieldIndex > -1 && csqFields.length > alleleFieldIndex)
+ {
+ String csqAllele = csqFields[alleleFieldIndex];
+ String vcfAllele = variant.getAlternateAllele(altAlleleIndex)
+ .getBaseString();
+ return csqAllele.equals(vcfAllele);
+ }
+ return false;
+ }
+
+ /**
* Add any allele-specific VCF key-value data to the sequence feature
*
* @param variant
@@ -603,9 +1125,12 @@ public class VCFLoader
* @param sf
* @param altAlelleIndex
* (0, 1..)
+ * @param consequence
+ * if not null, the consequence specific to this sequence (transcript
+ * feature) and allele
*/
protected void addAlleleProperties(VariantContext variant, SequenceI seq,
- SequenceFeature sf, final int altAlelleIndex)
+ SequenceFeature sf, final int altAlelleIndex, String consequence)
{
Map atts = variant.getAttributes();
@@ -617,9 +1142,17 @@ public class VCFLoader
* extract Consequence data (if present) that we are able to
* associated with the allele for this variant feature
*/
- if (CSQ.equals(key))
+ if (CSQ_FIELD.equals(key))
+ {
+ addConsequences(variant, seq, sf, consequence);
+ continue;
+ }
+
+ /*
+ * filter out fields we don't want to capture
+ */
+ if (!vcfFieldsOfInterest.contains(key))
{
- addConsequences(variant, seq, sf, altAlelleIndex);
continue;
}
@@ -669,30 +1202,24 @@ public class VCFLoader
/**
* Inspects CSQ data blocks (consequences) and adds attributes on the sequence
- * feature for the current allele (and transcript if applicable)
+ * feature.
*
- * Allele matching: if field ALLELE_NUM is present, it must match
- * altAlleleIndex. If not present, then field Allele value must match the VCF
- * Allele.
- *
- * Transcript matching: if sequence name can be identified to at least one of
- * the consequences' Feature values, then select only consequences that match
- * the value (i.e. consequences for the current transcript sequence). If not,
- * take all consequences (this is the case when adding features to the gene
- * sequence).
+ * If myConsequence
is not null, then this is the specific
+ * consequence data (pipe-delimited fields) that is for the current allele and
+ * transcript (sequence) being processed)
*
* @param variant
* @param seq
* @param sf
- * @param altAlelleIndex
- * (0, 1..)
+ * @param myConsequence
*/
protected void addConsequences(VariantContext variant, SequenceI seq,
- SequenceFeature sf, int altAlelleIndex)
+ SequenceFeature sf, String myConsequence)
{
- Object value = variant.getAttribute(CSQ);
+ Object value = variant.getAttribute(CSQ_FIELD);
+ // TODO if CSQ not present, try ANN (for SnpEff consequence data)?
- if (value == null || !(value instanceof ArrayList>))
+ if (value == null || !(value instanceof List>))
{
return;
}
@@ -700,118 +1227,41 @@ public class VCFLoader
List consequences = (List) value;
/*
- * if CSQ data includes 'Feature', and any value matches the sequence name,
- * then restrict consequence data to only the matching value (transcript)
- * i.e. just pick out consequences for the transcript the variant feature is on
+ * inspect CSQ consequences; restrict to the consequence
+ * associated with the current transcript (Feature)
*/
- String seqName = seq.getName()== null ? "" : seq.getName().toLowerCase();
- String matchFeature = null;
- if (csqFeatureFieldIndex > -1)
+ Map csqValues = new HashMap<>();
+
+ for (String consequence : consequences)
{
- for (String consequence : consequences)
+ if (myConsequence == null || myConsequence.equals(consequence))
{
String[] csqFields = consequence.split(PIPE_REGEX);
- if (csqFields.length > csqFeatureFieldIndex)
+
+ /*
+ * inspect individual fields of this consequence, copying non-null
+ * values which are 'fields of interest'
+ */
+ int i = 0;
+ for (String field : csqFields)
{
- String featureIdentifier = csqFields[csqFeatureFieldIndex];
- if (featureIdentifier.length() > 4
- && seqName.indexOf(featureIdentifier.toLowerCase()) > -1)
+ if (field != null && field.length() > 0)
{
- matchFeature = featureIdentifier;
+ String id = vepFieldsOfInterest.get(i);
+ if (id != null)
+ {
+ csqValues.put(id, field);
+ }
}
+ i++;
}
}
}
- StringBuilder sb = new StringBuilder(128);
- boolean found = false;
-
- for (String consequence : consequences)
+ if (!csqValues.isEmpty())
{
- String[] csqFields = consequence.split(PIPE_REGEX);
-
- if (includeConsequence(csqFields, matchFeature, variant,
- altAlelleIndex))
- {
- if (found)
- {
- sb.append(COMMA);
- }
- found = true;
- sb.append(consequence);
- }
- }
-
- if (found)
- {
- sf.setValue(CSQ, sb.toString());
- }
- }
-
- /**
- * Answers true if we want to associate this block of consequence data with
- * the specified alternate allele of the VCF variant.
- *
- * If consequence data includes the ALLELE_NUM field, then this has to match
- * altAlleleIndex. Otherwise the Allele field of the consequence data has to
- * match the allele value.
- *
- * Optionally (if matchFeature is not null), restrict to only include
- * consequences whose Feature value matches. This allows us to attach
- * consequences to their respective transcripts.
- *
- * @param csqFields
- * @param matchFeature
- * @param variant
- * @param altAlelleIndex
- * (0, 1..)
- * @return
- */
- protected boolean includeConsequence(String[] csqFields,
- String matchFeature, VariantContext variant, int altAlelleIndex)
- {
- /*
- * check consequence is for the current transcript
- */
- if (matchFeature != null)
- {
- if (csqFields.length <= csqFeatureFieldIndex)
- {
- return false;
- }
- String featureIdentifier = csqFields[csqFeatureFieldIndex];
- if (!featureIdentifier.equals(matchFeature))
- {
- return false; // consequence is for a different transcript
- }
- }
-
- /*
- * if ALLELE_NUM is present, it must match altAlleleIndex
- * NB first alternate allele is 1 for ALLELE_NUM, 0 for altAlleleIndex
- */
- if (csqAlleleNumberFieldIndex > -1)
- {
- if (csqFields.length <= csqAlleleNumberFieldIndex)
- {
- return false;
- }
- String alleleNum = csqFields[csqAlleleNumberFieldIndex];
- return String.valueOf(altAlelleIndex + 1).equals(alleleNum);
- }
-
- /*
- * else consequence allele must match variant allele
- */
- if (csqAlleleFieldIndex > -1 && csqFields.length > csqAlleleFieldIndex)
- {
- String csqAllele = csqFields[csqAlleleFieldIndex];
- String vcfAllele = variant.getAlternateAllele(altAlelleIndex)
- .getBaseString();
- return csqAllele.equals(vcfAllele);
+ sf.setValue(CSQ_FIELD, csqValues);
}
-
- return false;
}
/**
@@ -865,8 +1315,8 @@ public class VCFLoader
* call (e.g.) http://rest.ensembl.org/map/human/GRCh38/17:45051610..45109016:1/GRCh37
*/
EnsemblMap mapper = new EnsemblMap();
- int[] mapping = mapper.getMapping(species, chromosome, fromRef, toRef,
- queryRange);
+ int[] mapping = mapper.getAssemblyMapping(species, chromosome, fromRef,
+ toRef, queryRange);
if (mapping == null)
{