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.features.FeatureSource;
import jalview.datamodel.features.FeatureSources;
import jalview.ext.ensembl.EnsemblMap;
+import jalview.ext.htsjdk.HtsContigDb;
import jalview.ext.htsjdk.VCFReader;
import jalview.io.gff.Gff3Helper;
import jalview.io.gff.SequenceOntologyI;
import jalview.util.MappingUtils;
import jalview.util.MessageManager;
+import java.io.File;
import java.io.IOException;
import java.util.ArrayList;
import java.util.HashMap;
*/
public class VCFLoader
{
+ private static final String DEFAULT_SPECIES = "homo_sapiens";
+
/**
* A class to model the mapping from sequence to VCF coordinates. Cases include
* <ul>
/*
* Lookup keys, and default values, for Preference entries that describe
- * patterns for VCF and VEP fields to capture
+ * patterns for VCF and VEP fields to capture
*/
private static final String VEP_FIELDS_PREF = "VEP_FIELDS";
private static final String DEFAULT_VEP_FIELDS = ".*";// "Allele,Consequence,IMPACT,SWISSPROT,SIFT,PolyPhen,CLIN_SIG";
/*
+ * Lookup keys, and default values, for Preference entries that give
+ * mappings from tokens in the 'reference' header to species or assembly
+ */
+ private static final String VCF_ASSEMBLY = "VCF_ASSEMBLY";
+
+ private static final String DEFAULT_VCF_ASSEMBLY = "assembly19=GRCh37,hs37=GRCh37,grch37=GRCh37,grch38=GRCh38";
+
+ private static final String VCF_SPECIES = "VCF_SPECIES"; // default is human
+
+ private static final String DEFAULT_REFERENCE = "grch37"; // fallback default is human GRCh37
+
+ /*
* keys to fields of VEP CSQ consequence data
* see https://www.ensembl.org/info/docs/tools/vep/vep_formats.html
*/
private static final String PIPE_REGEX = "\\|";
/*
- * key for Allele Frequency output by VEP
- * see http://www.ensembl.org/info/docs/tools/vep/vep_formats.html
- */
- private static final String ALLELE_FREQUENCY_KEY = "AF";
-
- /*
* delimiter that separates multiple consequence data blocks
*/
private static final String COMMA = ",";
private static final String EXCL = "!";
/*
- * the alignment we are associating VCF data with
+ * the VCF file we are processing
*/
- private AlignmentI al;
+ protected String vcfFilePath;
/*
* mappings between VCF and sequence reference assembly regions, as
*/
private Map<String, Map<int[], int[]>> assemblyMappings;
+ private VCFReader reader;
+
/*
* holds details of the VCF header lines (metadata)
*/
private VCFHeader header;
/*
+ * species (as a valid Ensembl term) the VCF is for
+ */
+ private String vcfSpecies;
+
+ /*
+ * genome assembly version (as a valid Ensembl identifier) the VCF is for
+ */
+ private String vcfAssembly;
+
+ /*
* a Dictionary of contigs (if present) referenced in the VCF file
*/
private SAMSequenceDictionary dictionary;
Map<Integer, String> vepFieldsOfInterest;
/**
- * Constructor given an alignment context
+ * Constructor given a VCF file
*
* @param alignment
*/
- public VCFLoader(AlignmentI alignment)
+ public VCFLoader(String vcfFile)
{
- al = alignment;
+ try
+ {
+ initialise(vcfFile);
+ } catch (IOException e)
+ {
+ System.err.println("Error opening VCF file: " + e.getMessage());
+ }
// map of species!chromosome!fromAssembly!toAssembly to {fromRange, toRange}
assemblyMappings = new HashMap<>();
}
/**
- * Starts a new thread to query and load VCF variant data on to the alignment
+ * Starts a new thread to query and load VCF variant data on to the given
+ * sequences
* <p>
* This method is not thread safe - concurrent threads should use separate
* instances of this class.
*
- * @param filePath
+ * @param seqs
* @param gui
*/
- public void loadVCF(final String filePath,
- final AlignViewControllerGuiI gui)
+ public void loadVCF(SequenceI[] seqs, final AlignViewControllerGuiI gui)
{
if (gui != null)
{
new Thread()
{
-
@Override
public void run()
{
- VCFLoader.this.doLoad(filePath, gui);
+ VCFLoader.this.doLoad(seqs, gui);
}
-
}.start();
}
/**
- * Loads VCF on to an alignment - provided it can be related to one or more
- * sequence's chromosomal coordinates
+ * Reads the specified contig sequence and adds its VCF variants to it
*
- * @param filePath
- * @param gui
- * optional callback handler for messages
+ * @param contig
+ * the id of a single sequence (contig) to load
+ * @return
*/
- protected void doLoad(String filePath, AlignViewControllerGuiI gui)
+ public SequenceI loadVCFContig(String contig)
{
- VCFReader reader = null;
- try
+ VCFHeaderLine headerLine = header.getOtherHeaderLine(VCFHeader.REFERENCE_KEY);
+ if (headerLine == null)
{
- // long start = System.currentTimeMillis();
- reader = new VCFReader(filePath);
-
- header = reader.getFileHeader();
-
- try
- {
- dictionary = header.getSequenceDictionary();
- } catch (SAMException e)
- {
- // ignore - thrown if any contig line lacks length info
- }
+ Cache.log.error("VCF reference header not found");
+ return null;
+ }
+ String ref = headerLine.getValue();
+ if (ref.startsWith("file://"))
+ {
+ ref = ref.substring(7);
+ }
+ setSpeciesAndAssembly(ref);
- sourceId = filePath;
+ SequenceI seq = null;
+ File dbFile = new File(ref);
- saveMetadata(sourceId);
+ if (dbFile.exists())
+ {
+ HtsContigDb db = new HtsContigDb("", dbFile);
+ seq = db.getSequenceProxy(contig);
+ loadSequenceVCF(seq);
+ db.close();
+ }
+ else
+ {
+ Cache.log.error("VCF reference not found: " + ref);
+ }
- /*
- * get offset of CSQ ALLELE_NUM and Feature if declared
- */
- parseCsqHeader();
+ return seq;
+ }
+ /**
+ * Loads VCF on to one or more sequences
+ *
+ * @param seqs
+ * @param gui
+ * optional callback handler for messages
+ */
+ protected void doLoad(SequenceI[] seqs, AlignViewControllerGuiI gui)
+ {
+ try
+ {
VCFHeaderLine ref = header
.getOtherHeaderLine(VCFHeader.REFERENCE_KEY);
- String vcfAssembly = ref.getValue();
+ String reference = ref == null ? null : ref.getValue();
+
+ setSpeciesAndAssembly(reference);
int varCount = 0;
int seqCount = 0;
/*
* query for VCF overlapping each sequence in turn
*/
- for (SequenceI seq : al.getSequences())
+ for (SequenceI seq : seqs)
{
- int added = loadSequenceVCF(seq, reader, vcfAssembly);
+ int added = loadSequenceVCF(seq);
if (added > 0)
{
seqCount++;
}
if (gui != null)
{
- // long elapsed = System.currentTimeMillis() - start;
String msg = MessageManager.formatMessage("label.added_vcf",
varCount, seqCount);
gui.setStatus(msg);
}
/**
+ * Attempts to determine and save the species and genome assembly version to
+ * which the VCF data applies. This may be done by parsing the {@code reference}
+ * header line, configured in a property file, or (potentially) confirmed
+ * interactively by the user.
+ * <p>
+ * The saved values should be identifiers valid for Ensembl's REST service
+ * {@code map} endpoint, so they can be used (if necessary) to retrieve the
+ * mapping between VCF coordinates and sequence coordinates.
+ *
+ * @param reference
+ * @see https://rest.ensembl.org/documentation/info/assembly_map
+ * @see https://rest.ensembl.org/info/assembly/human?content-type=text/xml
+ * @see https://rest.ensembl.org/info/species?content-type=text/xml
+ */
+ protected void setSpeciesAndAssembly(String reference)
+ {
+ if (reference == null)
+ {
+ Cache.log.error("No VCF ##reference found, defaulting to "
+ + DEFAULT_REFERENCE + ":" + DEFAULT_SPECIES);
+ reference = DEFAULT_REFERENCE; // default to GRCh37 if not specified
+ }
+ reference = reference.toLowerCase();
+
+ /*
+ * for a non-human species, or other assembly identifier,
+ * specify as a Jalview property file entry e.g.
+ * VCF_ASSEMBLY = hs37=GRCh37,assembly19=GRCh37
+ * VCF_SPECIES = c_elegans=celegans
+ * to map a token in the reference header to a value
+ */
+ String prop = Cache.getDefault(VCF_ASSEMBLY, DEFAULT_VCF_ASSEMBLY);
+ for (String token : prop.split(","))
+ {
+ String[] tokens = token.split("=");
+ if (tokens.length == 2)
+ {
+ if (reference.contains(tokens[0].trim().toLowerCase()))
+ {
+ vcfAssembly = tokens[1].trim();
+ break;
+ }
+ }
+ }
+
+ vcfSpecies = DEFAULT_SPECIES;
+ prop = Cache.getProperty(VCF_SPECIES);
+ if (prop != null)
+ {
+ for (String token : prop.split(","))
+ {
+ String[] tokens = token.split("=");
+ if (tokens.length == 2)
+ {
+ if (reference.contains(tokens[0].trim().toLowerCase()))
+ {
+ vcfSpecies = tokens[1].trim();
+ break;
+ }
+ }
+ }
+ }
+ }
+
+ /**
+ * Opens the VCF file and parses header data
+ *
+ * @param filePath
+ * @throws IOException
+ */
+ private void initialise(String filePath) throws IOException
+ {
+ vcfFilePath = filePath;
+
+ reader = new VCFReader(filePath);
+
+ header = reader.getFileHeader();
+
+ try
+ {
+ dictionary = header.getSequenceDictionary();
+ } catch (SAMException e)
+ {
+ // ignore - thrown if any contig line lacks length info
+ }
+
+ sourceId = filePath;
+
+ saveMetadata(sourceId);
+
+ /*
+ * get offset of CSQ ALLELE_NUM and Feature if declared
+ */
+ parseCsqHeader();
+ }
+
+ /**
* Reads metadata (such as INFO field descriptions and datatypes) and saves
* them for future reference
*
}
/**
- * 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 species, chromosomal
- * positions and reference assembly, in order to be able to map the VCF
- * variants to the sequence (or not)
+ * Tries to add overlapping variants read from a VCF file to the given sequence,
+ * and returns the number of variant features added
*
* @param seq
- * @param reader
- * @param vcfAssembly
* @return
*/
- protected int loadSequenceVCF(SequenceI seq, VCFReader reader,
- String vcfAssembly)
+ protected int loadSequenceVCF(SequenceI seq)
{
- VCFMap vcfMap = getVcfMap(seq, vcfAssembly);
+ VCFMap vcfMap = getVcfMap(seq);
if (vcfMap == null)
{
return 0;
{
dss = seq;
}
- return addVcfVariants(dss, reader, vcfMap, vcfAssembly);
+ return addVcfVariants(dss, vcfMap);
}
/**
* Answers a map from sequence coordinates to VCF chromosome ranges
*
* @param seq
- * @param vcfAssembly
* @return
*/
- private VCFMap getVcfMap(SequenceI seq, String vcfAssembly)
+ private VCFMap getVcfMap(SequenceI seq)
{
/*
* simplest case: sequence has id and length matching a VCF contig
String seqRef = seqCoords.getAssemblyId();
MapList map = seqCoords.getMap();
- if (!vcfSpeciesMatchesSequence(vcfAssembly, species))
+ // note this requires the configured species to match that
+ // returned with the Ensembl sequence; todo: support aliases?
+ if (!vcfSpecies.equalsIgnoreCase(species))
{
+ Cache.log.warn("No VCF loaded to " + seq.getName()
+ + " as species not matched");
return null;
}
- if (vcfAssemblyMatchesSequence(vcfAssembly, seqRef))
+ if (seqRef.equalsIgnoreCase(vcfAssembly))
{
return new VCFMap(chromosome, map);
}
- if (!"GRCh38".equalsIgnoreCase(seqRef) // Ensembl
- || !vcfAssembly.contains("Homo_sapiens_assembly19")) // gnomAD
- {
- return null;
- }
-
/*
- * map chromosomal coordinates from sequence to VCF if the VCF
- * data has a different reference assembly to the sequence
+ * VCF data has a different reference assembly to the sequence:
+ * query Ensembl to map chromosomal coordinates from sequence to VCF
*/
- // TODO generalise for cases other than GRCh38 -> GRCh37 !
- // - or get the user to choose in a dialog
-
List<int[]> toVcfRanges = new ArrayList<>();
List<int[]> fromSequenceRanges = new ArrayList<>();
- String toRef = "GRCh37";
for (int[] range : map.getToRanges())
{
}
int[] newRange = mapReferenceRange(range, chromosome, "human", seqRef,
- toRef);
+ vcfAssembly);
if (newRange == null)
{
Cache.log.error(
String.format("Failed to map %s:%s:%s:%d:%d to %s", species,
- chromosome, seqRef, range[0], range[1], toRef));
+ chromosome, seqRef, range[0], range[1],
+ vcfAssembly));
continue;
}
else
}
/**
- * 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;
- }
-
- /**
- * 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)
+ protected int addVcfVariants(SequenceI seq, VCFMap map)
{
boolean forwardStrand = map.map.isToForwardStrand();
}
/**
- * A convenience method to get the AF value for the given alternate allele
- * index
- *
- * @param variant
- * @param alleleIndex
- * @return
- */
- protected float getAlleleFrequency(VariantContext variant, int alleleIndex)
- {
- float score = 0f;
- String attributeValue = getAttributeValue(variant,
- ALLELE_FREQUENCY_KEY, alleleIndex);
- if (attributeValue != null)
- {
- try
- {
- score = Float.parseFloat(attributeValue);
- } catch (NumberFormatException e)
- {
- // leave as 0
- }
- }
-
- return score;
- }
-
- /**
* A convenience method to get an attribute value for an alternate allele
*
* @param variant
String type = SequenceOntologyI.SEQUENCE_VARIANT;
if (consequence != null)
{
- type = getOntologyTerm(seq, variant, altAlleleIndex,
- consequence);
+ type = getOntologyTerm(consequence);
}
- float score = getAlleleFrequency(variant, altAlleleIndex);
-
SequenceFeature sf = new SequenceFeature(type, alleles, featureStart,
- featureEnd, score, FEATURE_GROUP_VCF);
+ featureEnd, FEATURE_GROUP_VCF);
sf.setSource(sourceId);
sf.setValue(Gff3Helper.ALLELES, alleles);
- addAlleleProperties(variant, seq, sf, altAlleleIndex, consequence);
+ addAlleleProperties(variant, sf, altAlleleIndex, consequence);
seq.addSequenceFeature(sf);
* <li>sequence id can be matched to VEP Feature (or SnpEff Feature_ID)</li>
* </ul>
*
- * @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 getOntologyTerm(String consequence)
{
String type = SequenceOntologyI.SEQUENCE_VARIANT;
+ /*
+ * could we associate Consequence data with this allele and feature (transcript)?
+ * if so, prefer the consequence term from that data
+ */
if (csqAlleleFieldIndex == -1) // && snpEffAlleleFieldIndex == -1
{
/*
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);
* Add any allele-specific VCF key-value data to the sequence feature
*
* @param variant
- * @param seq
* @param sf
* @param altAlelleIndex
* (0, 1..)
* if not null, the consequence specific to this sequence (transcript
* feature) and allele
*/
- protected void addAlleleProperties(VariantContext variant, SequenceI seq,
+ protected void addAlleleProperties(VariantContext variant,
SequenceFeature sf, final int altAlelleIndex, String consequence)
{
Map<String, Object> atts = variant.getAttributes();
*/
if (CSQ_FIELD.equals(key))
{
- addConsequences(variant, seq, sf, consequence);
+ addConsequences(variant, sf, consequence);
+ continue;
+ }
+
+ /*
+ * filter out fields we don't want to capture
+ */
+ if (!vcfFieldsOfInterest.contains(key))
+ {
continue;
}
* transcript (sequence) being processed)
*
* @param variant
- * @param seq
* @param sf
* @param myConsequence
*/
- protected void addConsequences(VariantContext variant, SequenceI seq,
- SequenceFeature sf, String myConsequence)
+ protected void addConsequences(VariantContext variant, SequenceFeature sf,
+ String myConsequence)
{
Object value = variant.getAttribute(CSQ_FIELD);
- // TODO if CSQ not present, try ANN (for SnpEff consequence data)?
if (value == null || !(value instanceof List<?>))
{
git://source.jalview.org / jalview.git / blobdiff