package jalview.io.vcf;
-import htsjdk.samtools.util.CloseableIterator;
-import htsjdk.variant.variantcontext.Allele;
-import htsjdk.variant.variantcontext.VariantContext;
-import htsjdk.variant.vcf.VCFHeader;
-import htsjdk.variant.vcf.VCFHeaderLine;
-import htsjdk.variant.vcf.VCFHeaderLineCount;
-import htsjdk.variant.vcf.VCFInfoHeaderLine;
-
import jalview.analysis.AlignmentUtils;
import jalview.analysis.Dna;
import jalview.api.AlignViewControllerGuiI;
+import jalview.bin.Cache;
import jalview.datamodel.AlignmentI;
import jalview.datamodel.DBRefEntry;
import jalview.datamodel.GeneLociI;
import jalview.datamodel.Mapping;
import jalview.datamodel.SequenceFeature;
import jalview.datamodel.SequenceI;
+import jalview.datamodel.features.FeatureAttributeType;
+import jalview.datamodel.features.FeatureSource;
+import jalview.datamodel.features.FeatureSources;
import jalview.ext.ensembl.EnsemblMap;
import jalview.ext.htsjdk.VCFReader;
import jalview.io.gff.Gff3Helper;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
+import java.util.regex.Pattern;
+import java.util.regex.PatternSyntaxException;
+
+import htsjdk.samtools.util.CloseableIterator;
+import htsjdk.variant.variantcontext.Allele;
+import htsjdk.variant.variantcontext.VariantContext;
+import htsjdk.variant.vcf.VCFHeader;
+import htsjdk.variant.vcf.VCFHeaderLine;
+import htsjdk.variant.vcf.VCFHeaderLineCount;
+import htsjdk.variant.vcf.VCFHeaderLineType;
+import htsjdk.variant.vcf.VCFInfoHeaderLine;
/**
* A class to read VCF data (using the htsjdk) and add variants as sequence
public class VCFLoader
{
/*
+ * Lookup keys, and default values, for Preference entries that describe
+ * patterns for VCF and VEP fields to capture
+ */
+ private static final String VEP_FIELDS_PREF = "VEP_FIELDS";
+
+ private static final String VCF_FIELDS_PREF = "VCF_FIELDS";
+
+ private static final String DEFAULT_VCF_FIELDS = "AF,AC*";
+
+ private static final String DEFAULT_VEP_FIELDS = ".*";// "Allele,Consequence,IMPACT,SWISSPROT,SIFT,PolyPhen,CLIN_SIG";
+
+ /*
* keys to fields of VEP CSQ consequence data
* see https://www.ensembl.org/info/docs/tools/vep/vep_formats.html
*/
- private static final String ALLELE_NUM_KEY = "ALLELE_NUM"; // 0 (ref), 1...
+ private static final String ALLELE_KEY = "Allele";
+ private static final String ALLELE_NUM_KEY = "ALLELE_NUM"; // 0 (ref), 1...
private static final String FEATURE_KEY = "Feature"; // Ensembl stable id
/*
* default VCF INFO key for VEP consequence data
* NB this can be overridden running VEP with --vcf_info_field
- * - we don't handle this case (require CSQ identifier)
+ * - we don't handle this case (require identifier to be CSQ)
*/
- private static final String CSQ = "CSQ";
+ private static final String CSQ_FIELD = "CSQ";
/*
- * separator for fields in consequence data
+ * separator for fields in consequence data is '|'
*/
- private static final String PIPE = "|";
-
- private static final String PIPE_REGEX = "\\" + PIPE;
+ private static final String PIPE_REGEX = "\\|";
/*
* key for Allele Frequency output by VEP
private static final String COMMA = ",";
/*
- * (temporary) flag that determines whether Jalview adds one feature
- * per VCF record, or one per allele (preferred)
- */
- private static final boolean FEATURE_PER_ALLELE = true;
-
- /*
* the feature group assigned to a VCF variant in Jalview
*/
private static final String FEATURE_GROUP_VCF = "VCF";
private VCFHeader header;
/*
- * the position (0...) of the ALLELE_NUM field in each block of
+ * the position (0...) of field in each block of
* CSQ (consequence) data (if declared in the VCF INFO header for CSQ)
* see http://www.ensembl.org/info/docs/tools/vep/vep_formats.html
*/
+ private int csqAlleleFieldIndex = -1;
private int csqAlleleNumberFieldIndex = -1;
-
private int csqFeatureFieldIndex = -1;
+ /*
+ * a unique identifier under which to save metadata about feature
+ * attributes (selected INFO field data)
+ */
+ private String sourceId;
+
+ /*
+ * The INFO IDs of data that is both present in the VCF file, and
+ * also matched by any filters for data of interest
+ */
+ List<String> vcfFieldsOfInterest;
+
+ /*
+ * The field offsets and identifiers for VEP (CSQ) data that is both present
+ * in the VCF file, and also matched by any filters for data of interest
+ * for example 0 -> Allele, 1 -> Consequence, ..., 36 -> SIFT, ...
+ */
+ Map<Integer, String> vepFieldsOfInterest;
+
/**
* Constructor given an alignment context
*
al = alignment;
// map of species!chromosome!fromAssembly!toAssembly to {fromRange, toRange}
- assemblyMappings = new HashMap<String, Map<int[], int[]>>();
+ assemblyMappings = new HashMap<>();
}
/**
reader = new VCFReader(filePath);
header = reader.getFileHeader();
- VCFHeaderLine ref = header
- .getOtherHeaderLine(VCFHeader.REFERENCE_KEY);
+
+ sourceId = filePath;
+
+ saveMetadata(sourceId);
/*
- * note offset of CSQ ALLELE_NUM field if it is declared
+ * get offset of CSQ ALLELE_NUM and Feature if declared
*/
- locateCsqFields();
+ parseCsqHeader();
- // check if reference is wrt assembly19 (GRCh37)
- // todo may need to allow user to specify reference assembly?
- boolean isRefGrch37 = (ref != null && ref.getValue().contains(
- "assembly19"));
+ VCFHeaderLine ref = header
+ .getOtherHeaderLine(VCFHeader.REFERENCE_KEY);
+ String vcfAssembly = ref.getValue();
int varCount = 0;
int seqCount = 0;
*/
for (SequenceI seq : al.getSequences())
{
- int added = loadVCF(seq, reader, isRefGrch37);
+ int added = loadSequenceVCF(seq, reader, vcfAssembly);
if (added > 0)
{
seqCount++;
}
/**
- * Records the position of fields for ALLELE_NUM and Feature defined in the
- * CSQ INFO header (if there is one). CSQ fields are declared in the CSQ INFO
- * Description e.g.
+ * Reads metadata (such as INFO field descriptions and datatypes) and saves
+ * them for future reference
+ *
+ * @param theSourceId
+ */
+ void saveMetadata(String theSourceId)
+ {
+ List<Pattern> vcfFieldPatterns = getFieldMatchers(VCF_FIELDS_PREF,
+ DEFAULT_VCF_FIELDS);
+ vcfFieldsOfInterest = new ArrayList<>();
+
+ FeatureSource metadata = new FeatureSource(theSourceId);
+
+ for (VCFInfoHeaderLine info : header.getInfoHeaderLines())
+ {
+ String attributeId = info.getID();
+ String desc = info.getDescription();
+ VCFHeaderLineType type = info.getType();
+ FeatureAttributeType attType = null;
+ switch (type)
+ {
+ case Character:
+ attType = FeatureAttributeType.Character;
+ break;
+ case Flag:
+ attType = FeatureAttributeType.Flag;
+ break;
+ case Float:
+ attType = FeatureAttributeType.Float;
+ break;
+ case Integer:
+ attType = FeatureAttributeType.Integer;
+ break;
+ case String:
+ attType = FeatureAttributeType.String;
+ break;
+ }
+ metadata.setAttributeName(attributeId, desc);
+ metadata.setAttributeType(attributeId, attType);
+
+ if (isFieldWanted(attributeId, vcfFieldPatterns))
+ {
+ vcfFieldsOfInterest.add(attributeId);
+ }
+ }
+
+ FeatureSources.getInstance().addSource(theSourceId, metadata);
+ }
+
+ /**
+ * Answers true if the field id is matched by any of the filter patterns, else
+ * false. Matching is against regular expression patterns, and is not
+ * case-sensitive.
+ *
+ * @param id
+ * @param filters
+ * @return
+ */
+ private boolean isFieldWanted(String id, List<Pattern> 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.
+ * <p>
+ * CSQ fields are declared in the CSQ INFO Description e.g.
* <p>
* Description="Consequence ...from ... VEP. Format: Allele|Consequence|...
*/
- protected void locateCsqFields()
+ protected void parseCsqHeader()
{
- VCFInfoHeaderLine csqInfo = header.getInfoHeaderLine(CSQ);
+ List<Pattern> 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 spacePos = desc.lastIndexOf(" ");
+ desc = desc.substring(spacePos + 1);
+
if (desc != null)
{
String[] format = desc.split(PIPE_REGEX);
{
csqAlleleNumberFieldIndex = index;
}
+ if (ALLELE_KEY.equals(field))
+ {
+ csqAlleleFieldIndex = index;
+ }
if (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.
+ * <p>
+ * 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<Pattern> getFieldMatchers(String key, String def)
+ {
+ String pref = Cache.getDefault(key, def);
+ List<Pattern> 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
*/
/**
* 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 loadVCF(SequenceI seq, VCFReader reader,
- boolean isVcfRefGrch37)
+ protected int loadSequenceVCF(SequenceI seq, VCFReader reader,
+ String vcfAssembly)
{
int count = 0;
GeneLociI seqCoords = seq.getGeneLoci();
if (seqCoords == null)
{
+ System.out.println(String.format(
+ "Can't query VCF for %s as chromosome coordinates not known",
+ seq.getName()));
+ return 0;
+ }
+
+ if (!vcfSpeciesMatchesSequence(vcfAssembly, seqCoords.getSpeciesId()))
+ {
return 0;
}
List<int[]> seqChromosomalContigs = seqCoords.getMap().getToRanges();
for (int[] range : seqChromosomalContigs)
{
- count += addVcfVariants(seq, reader, range, isVcfRefGrch37);
+ count += addVcfVariants(seq, reader, range, vcfAssembly);
}
return count;
}
/**
+ * 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 given chromosome
* region of the sequence, and adds as variant features. Returns the number of
* overlapping variants found.
* @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
+ * the '##reference' identifier for the VCF reference assembly
* @return
*/
protected int addVcfVariants(SequenceI seq, VCFReader reader,
- int[] range, boolean isVcfRefGrch37)
+ int[] range, String vcfAssembly)
{
GeneLociI seqCoords = seq.getGeneLoci();
String species = seqCoords.getSpeciesId();
/*
- * 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
+ // TODO generalise for non-human species
+ // - or get the user to choose in a dialog
+
int offset = 0;
- String fromRef = "GRCh38";
- if (fromRef.equalsIgnoreCase(seqRef) && isVcfRefGrch37)
+ if ("GRCh38".equalsIgnoreCase(seqRef) // Ensembl
+ && vcfAssembly.contains("Homo_sapiens_assembly19")) // gnomAD
{
String toRef = "GRCh37";
int[] newRange = mapReferenceRange(range, chromosome, "human",
- fromRef, toRef);
+ 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));
+ seqRef, range[0], range[1], toRef));
return 0;
}
offset = newRange[0] - range[0];
*/
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())
- {
- continue;
- }
-
int start = variant.getStart() - offset;
int end = variant.getEnd() - offset;
/*
* convert chromosomal location to sequence coordinates
+ * - may be reverse strand (convert to forward for sequence feature)
* - null if a partially overlapping feature
*/
int[] seqLocation = mapping.locateInFrom(start, end);
if (seqLocation != null)
{
- count += addVariantFeature(seq, variant, seqLocation[0],
- seqLocation[1], forwardStrand);
+ int featureStart = Math.min(seqLocation[0], seqLocation[1]);
+ int featureEnd = Math.max(seqLocation[0], seqLocation[1]);
+ count += addAlleleFeatures(seq, variant, featureStart, featureEnd,
+ forwardStrand);
}
}
}
/**
- * Inspects the VCF variant record, and adds variant features to the sequence.
- * Only SNP variants are added, not INDELs. Returns the number of features
- * added.
- * <p>
- * If the sequence maps to the reverse strand of the chromosome, reference and
- * variant bases are recorded as their complements (C/G, A/T).
- *
- * @param seq
- * @param variant
- * @param featureStart
- * @param featureEnd
- * @param forwardStrand
- */
- protected int addVariantFeature(SequenceI seq, VariantContext variant,
- int featureStart, int featureEnd, boolean forwardStrand)
- {
- byte[] reference = variant.getReference().getBases();
- if (reference.length != 1)
- {
- /*
- * sorry, we don't handle INDEL variants
- */
- return 0;
- }
-
- if (FEATURE_PER_ALLELE)
- {
- return addAlleleFeatures(seq, variant, featureStart, featureEnd,
- forwardStrand);
- }
-
- /*
- * for now we extract allele frequency as feature score; note
- * this attribute is String for a simple SNP, but List<String> if
- * multiple alleles at the locus; we extract for the simple case only
- */
- float score = getAlleleFrequency(variant, 0);
-
- StringBuilder sb = new StringBuilder();
- sb.append(forwardStrand ? (char) reference[0] : complement(reference));
-
- /*
- * inspect alleles and record SNP variants (as the variant
- * record could be MIXED and include INDEL and SNP alleles)
- * warning: getAlleles gives no guarantee as to the order
- * in which they are returned
- */
- for (Allele allele : variant.getAlleles())
- {
- if (!allele.isReference())
- {
- byte[] alleleBase = allele.getBases();
- if (alleleBase.length == 1)
- {
- sb.append(COMMA).append(
- forwardStrand ? (char) alleleBase[0]
- : complement(alleleBase));
- }
- }
- }
- String alleles = sb.toString(); // e.g. G,A,C
-
- String type = SequenceOntologyI.SEQUENCE_VARIANT;
-
- SequenceFeature sf = new SequenceFeature(type, alleles, featureStart,
- featureEnd, score, FEATURE_GROUP_VCF);
-
- sf.setValue(Gff3Helper.ALLELES, alleles);
-
- Map<String, Object> atts = variant.getAttributes();
- for (Entry<String, Object> att : atts.entrySet())
- {
- sf.setValue(att.getKey(), att.getValue());
- }
- seq.addSequenceFeature(sf);
-
- return 1;
- }
-
- /**
* A convenience method to get the AF value for the given alternate allele
* index
*
}
/**
- * Adds one variant feature for each SNP allele in the VCF variant record, and
+ * Adds one variant feature for each allele in the VCF variant record, and
* returns the number of features added.
*
* @param seq
/**
* Inspects one allele and attempts to add a variant feature for it to the
- * sequence. Only SNP variants are added as features. 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. 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).
*
* @param seq
* @param variant
* @param altAlleleIndex
+ * (0, 1..)
* @param featureStart
* @param featureEnd
* @param forwardStrand
int altAlleleIndex, int featureStart, int featureEnd,
boolean forwardStrand)
{
- byte[] reference = variant.getReference().getBases();
+ String reference = variant.getReference().getBaseString();
Allele alt = variant.getAlternateAllele(altAlleleIndex);
- byte[] allele = alt.getBases();
- if (allele.length != 1)
- {
- /*
- * not a SNP variant
- */
- return 0;
- }
+ String allele = alt.getBaseString();
/*
- * build the ref,alt allele description e.g. "G,A"
+ * 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 ? (char) reference[0] : complement(reference));
+ sb.append(forwardStrand ? reference : Dna.reverseComplement(reference));
sb.append(COMMA);
- sb.append(forwardStrand ? (char) allele[0] : complement(allele));
+ sb.append(forwardStrand ? allele : Dna.reverseComplement(allele));
String alleles = sb.toString(); // e.g. G,A
String type = SequenceOntologyI.SEQUENCE_VARIANT;
SequenceFeature sf = new SequenceFeature(type, alleles, featureStart,
featureEnd, score, FEATURE_GROUP_VCF);
+ sf.setSource(sourceId);
sf.setValue(Gff3Helper.ALLELES, alleles);
* @param seq
* @param sf
* @param altAlelleIndex
+ * (0, 1..)
*/
- protected void addAlleleProperties(VariantContext variant,
- SequenceI seq,
+ protected void addAlleleProperties(VariantContext variant, SequenceI seq,
SequenceFeature sf, final int altAlelleIndex)
{
Map<String, Object> atts = variant.getAttributes();
* extract Consequence data (if present) that we are able to
* associated with the allele for this variant feature
*/
- if (CSQ.equals(key) && csqAlleleNumberFieldIndex > -1)
+ if (CSQ_FIELD.equals(key))
{
- addConsequences(att.getValue(), seq, sf, altAlelleIndex + 1);
- return;
+ addConsequences(variant, seq, sf, altAlelleIndex);
+ continue;
+ }
+
+ /*
+ * filter out fields we don't want to capture
+ */
+ if (!vcfFieldsOfInterest.contains(key))
+ {
+ continue;
}
/*
* Inspects CSQ data blocks (consequences) and adds attributes on the sequence
* feature for the current allele (and transcript if applicable)
* <p>
- * Allele matching: we require field ALLELE_NUM to match altAlleleIndex. If
- * the CSQ data does not include ALLELE_NUM values then no data is added to
- * the variant 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.
* <p>
* Transcript matching: if sequence name can be identified to at least one of
* the consequences' Feature values, then select only consequences that match
* take all consequences (this is the case when adding features to the gene
* sequence).
*
- * @param value
+ * @param variant
* @param seq
* @param sf
* @param altAlelleIndex
- * (1=first alternative allele...)
+ * (0, 1..)
*/
- protected void addConsequences(Object value, SequenceI seq,
- SequenceFeature sf,
- int altAlelleIndex)
+ protected void addConsequences(VariantContext variant, SequenceI seq,
+ SequenceFeature sf, int altAlelleIndex)
{
- if (!(value instanceof ArrayList<?>))
+ Object value = variant.getAttribute(CSQ_FIELD);
+
+ if (value == null || !(value instanceof ArrayList<?>))
{
return;
}
/*
* if CSQ data includes 'Feature', and any value matches the sequence name,
- * then restrict consequence data to the matching value (transcript)
+ * then restrict consequence data to only the matching value (transcript)
* i.e. just pick out consequences for the transcript the variant feature is on
*/
String seqName = seq.getName()== null ? "" : seq.getName().toLowerCase();
- boolean matchFeature = false;
- String matchFeatureValue = null;
+ String matchFeature = null;
if (csqFeatureFieldIndex > -1)
{
for (String consequence : consequences)
if (featureIdentifier.length() > 4
&& seqName.indexOf(featureIdentifier.toLowerCase()) > -1)
{
- matchFeature = true;
- matchFeatureValue = featureIdentifier;
+ matchFeature = featureIdentifier;
}
}
}
}
- StringBuilder sb = new StringBuilder(128);
- boolean found = false;
+ /*
+ * inspect CSQ consequences; where possible restrict to the consequence
+ * associated with the current transcript (Feature)
+ */
+ Map<String, String> csqValues = new HashMap<>();
for (String consequence : consequences)
{
String[] csqFields = consequence.split(PIPE_REGEX);
- /*
- * check consequence is for the current transcript
- */
- if (matchFeature)
+ if (includeConsequence(csqFields, matchFeature, variant,
+ altAlelleIndex))
{
- if (csqFields.length <= csqFeatureFieldIndex)
- {
- continue;
- }
- String featureIdentifier = csqFields[csqFeatureFieldIndex];
- if (!featureIdentifier.equals(matchFeatureValue))
+ /*
+ * inspect individual fields of this consequence, copying non-null
+ * values which are 'fields of interest'
+ */
+ int i = 0;
+ for (String field : csqFields)
{
- continue; // consequence is for a different transcript
+ if (field != null && field.length() > 0)
+ {
+ String id = vepFieldsOfInterest.get(i);
+ if (id != null)
+ {
+ csqValues.put(id, field);
+ }
+ }
+ i++;
}
}
+ }
+
+ if (!csqValues.isEmpty())
+ {
+ sf.setValue(CSQ_FIELD, csqValues);
+ }
+ }
- if (csqFields.length > csqAlleleNumberFieldIndex)
+ /**
+ * Answers true if we want to associate this block of consequence data with
+ * the specified alternate allele of the VCF variant.
+ * <p>
+ * 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.
+ * <p>
+ * 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)
{
- String alleleNum = csqFields[csqAlleleNumberFieldIndex];
- if (String.valueOf(altAlelleIndex).equals(alleleNum))
- {
- if (found)
- {
- sb.append(COMMA);
- }
- found = true;
- sb.append(consequence);
- }
+ 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);
}
- if (found)
+ /*
+ * else consequence allele must match variant allele
+ */
+ if (csqAlleleFieldIndex > -1 && csqFields.length > csqAlleleFieldIndex)
{
- sf.setValue(CSQ, sb.toString());
+ String csqAllele = csqFields[csqAlleleFieldIndex];
+ String vcfAllele = variant.getAlternateAllele(altAlelleIndex)
+ .getBaseString();
+ return csqAllele.equals(vcfAllele);
}
+
+ return false;
}
/**
* 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)
{
git://source.jalview.org / jalview.git / blobdiff