/** Create a list of variants from this VcfEntry */
public List<Variant> variants() {
if (variants != null) return variants;
// Create list of variants
variants = new LinkedList<>();
// Create one Variant for each ALT
Chromosome chr = (Chromosome) parent;
if (!isVariant()) {
// Not a variant?
List<Variant> vars = variants(chr, start, ref, null, id);
String alt = ".";
// Add original 'ALT' field as genotype
for (Variant variant : vars) variant.setGenotype(alt);
variants.addAll(vars);
} else {
// At least one variant
for (String alt : alts) {
if (!isVariant(alt)) alt = null;
List<Variant> vars = variants(chr, start, ref, alt, id);
variants.addAll(vars);
}
}
return variants;
}
/** Annotate a VCF entry */
public boolean annotate(VcfEntry vcfEntry) throws IOException {
boolean annotated = false;
Set<String> idSet = new HashSet<>();
Map<String, String> infos = new HashMap<>();
boolean exists = false;
// ---
// Find all matching database entries
// Note that QueryResult.variantVcfEntry can be 'null'
// ---
List<QueryResult> queryResults = new LinkedList<>();
Set<VcfEntry> uniqueVcfEntries = new HashSet<>();
for (Variant var : vcfEntry.variants()) {
// Skip huge structural variants
if (var.isStructuralHuge()) continue;
// Query database
Collection<VariantVcfEntry> results = query(var);
// Make sure we add all found VcfEntries
for (VariantVcfEntry dbEntry : results) uniqueVcfEntries.add(dbEntry.getVcfEntry());
// Add query and result
QueryResult qr = new QueryResult(var, results);
queryResults.add(qr);
if (debug) Gpr.debug("Adding QueryResult: " + qr);
}
// Try to find INFO fields that we might have not seen before
if (useAllInfoFields) {
for (VcfEntry ve : uniqueVcfEntries) discoverInfoFields(ve);
}
// Add INFO fields using 'REF' data
findDbInfoRef(infos, uniqueVcfEntries);
// ---
// Annotate all fields
// ---
for (QueryResult qr : queryResults) {
if (debug) Gpr.debug("Processing QueryResult: " + qr);
if (useId) findDbId(idSet, qr);
if (existsInfoField != null) exists |= findDbExists(qr);
if (useInfoFields) findDbInfo(infos, qr);
}
// Annotate input vcfEntry
annotated |= annotateIds(vcfEntry, idSet);
annotated |= annotateInfo(vcfEntry, infos);
if (exists) annotateExists(vcfEntry);
return annotated;
}
/** Does database entry 'dbVcfEntry' match 'variant'? */
protected boolean match(Variant var, VariantVcfEntry dbEntry) {
// Do coordinates match?
if (var.getChromosomeName().equals(dbEntry.getChromosomeName()) //
&& var.getStart() == dbEntry.getStart() //
&& var.getEnd() == dbEntry.getEnd() //
) {
if (useRefAlt) {
// Compare Ref & Alt
if (var.getReference().equalsIgnoreCase(dbEntry.getReference()) //
&& var.getAlt().equalsIgnoreCase(dbEntry.getAlt()) //
) return true;
} else {
// No need to use Ref & Alt, it's a match
return true;
}
}
return false;
}
/** Query VCF entries intersecting 'marker' at node 'idx' */
protected void queryIntersects(Interval queryMarker, int idx, Markers results) {
if (intersectFilePosStart[idx] == null) return;
if (debug) Gpr.debug("queryIntersects\tidx: " + idx);
// Read entries from disk
List<VcfEntry> vcfEntries = readEntries(idx);
// Find matching entries
for (VcfEntry ve : vcfEntries) {
// If any variant within the vcfEntry intersects the query
// marker, we store this VCF entry as a result
for (Variant var : ve.variants()) {
if (var.intersects(queryMarker)) {
if (debug) Gpr.debug("\tAdding matching result: " + ve);
results.add(ve);
break; // Store this entry only once
}
}
// Past query's end coordinate? We don't need to look any further
if (queryMarker.getEnd() < ve.getStart()) return;
}
}
/** Create a variant */
List<Variant> variants(Chromosome chromo, int start, String reference, String alt, String id) {
List<Variant> list = null;
if (alt != null) alt = alt.toUpperCase();
if (alt == null || alt.isEmpty() || alt.equals(reference)) {
// Non-variant
list = Variant.factory(chromo, start, reference, null, id, false);
} else if (alt.charAt(0) == '<') {
// Structural variants
if (alt.startsWith("<DEL")) {
// Case: Deletion
// 2 321682 . T <DEL> 6 PASS
// IMPRECISE;SVTYPE=DEL;END=321887;SVLEN=-105;CIPOS=-56,20;CIEND=-10,62
String ch = ref;
int startNew = start;
if (end > start) {
startNew = start + reference.length();
int size = end - startNew + 1;
char change[] = new char[size];
for (int i = 0; i < change.length; i++)
change[i] = reference.length() > i ? reference.charAt(i) : 'N';
ch = new String(change);
}
list = Variant.factory(chromo, startNew, ch, "", id, false);
} else if (alt.startsWith("<INV")) {
// Inversion
int startNew = start + reference.length();
Variant var = new Variant(chromo, startNew, end, id);
var.setVariantType(VariantType.INV);
list = new LinkedList<>();
list.add(var);
} else if (alt.startsWith("<DUP")) {
// Duplication
int startNew = start + reference.length();
Variant var = new Variant(chromo, startNew, end, id);
var.setVariantType(VariantType.DUP);
list = new LinkedList<>();
list.add(var);
}
} else if ((alt.indexOf('[') >= 0) || (alt.indexOf(']') >= 0)) {
// Translocations
// Parse ALT string
boolean left = alt.indexOf(']') >= 0;
String sep = (left ? "\\]" : "\\[");
String tpos[] = alt.split(sep);
String pos = tpos[1];
boolean before = (alt.indexOf(']') == 0) || (alt.indexOf('[') == 0);
String altBases = (before ? tpos[2] : tpos[0]);
// Parse 'chr:start'
String posSplit[] = pos.split(":");
String trChrName = posSplit[0];
Chromosome trChr = chromo.getGenome().getOrCreateChromosome(trChrName);
int trStart = Gpr.parseIntSafe(posSplit[1]) - 1;
VariantBnd var = new VariantBnd(chromo, start, ref, altBases, trChr, trStart, left, before);
list = new LinkedList<>();
list.add(var);
} else if (reference.length() == alt.length()) {
// Case: SNP, MNP
if (reference.length() == 1) {
// SNPs
// 20 3 . C G . PASS DP=100
list = Variant.factory(chromo, start, reference, alt, id, true);
} else {
// MNPs
// 20 3 . TC AT . PASS DP=100
// Sometimes the first bases are the same and we can trim them
int startDiff = Integer.MAX_VALUE;
for (int i = 0; i < reference.length(); i++)
if (reference.charAt(i) != alt.charAt(i)) startDiff = Math.min(startDiff, i);
// MNPs
// Sometimes the last bases are the same and we can trim them
int endDiff = 0;
for (int i = reference.length() - 1; i >= 0; i--)
if (reference.charAt(i) != alt.charAt(i)) endDiff = Math.max(endDiff, i);
String newRef = reference.substring(startDiff, endDiff + 1);
String newAlt = alt.substring(startDiff, endDiff + 1);
list = Variant.factory(chromo, start + startDiff, newRef, newAlt, id, true);
}
} else {
// Short Insertions, Deletions or Mixed Variants (substitutions)
VcfRefAltAlign align = new VcfRefAltAlign(alt, reference);
align.align();
int startDiff = align.getOffset();
switch (align.getVariantType()) {
case DEL:
// Case: Deletion
// 20 2 . TC T . PASS DP=100
// 20 2 . AGAC AAC . PASS DP=100
String ref = "";
String ch = align.getAlignment();
if (!ch.startsWith("-"))
throw new RuntimeException(
"Deletion '" + ch + "' does not start with '-'. This should never happen!");
list = Variant.factory(chromo, start + startDiff, ref, ch, id, true);
break;
case INS:
// Case: Insertion of A { tC ; tCA } tC is the reference allele
// 20 2 . TC TCA . PASS DP=100
ch = align.getAlignment();
ref = "";
if (!ch.startsWith("+"))
throw new RuntimeException(
"Insertion '" + ch + "' does not start with '+'. This should never happen!");
list = Variant.factory(chromo, start + startDiff, ref, ch, id, true);
break;
case MIXED:
// Case: Mixed variant (substitution)
reference = reference.substring(startDiff);
alt = alt.substring(startDiff);
list = Variant.factory(chromo, start + startDiff, reference, alt, id, true);
break;
default:
// Other change type?
throw new RuntimeException(
"Unsupported VCF change type '"
+ align.getVariantType()
+ "'\n\tRef: "
+ reference
+ "'\n\tAlt: '"
+ alt
+ "'\n\tVcfEntry: "
+ this);
}
}
// ---
// Add original 'ALT' field as genotype
// ---
if (list == null) list = new LinkedList<>();
for (Variant variant : list) variant.setGenotype(alt);
return list;
}