/**
* Is this single change a LOF?
*
* <p>Criteria: 1) Core splice sites acceptors or donors (only CORE ones) 2) Stop gained (if this
* happens at the last part of the protein, we assume it has no effect) 3) Frame shifts
*
* @param changeEffect
* @return
*/
protected boolean isNmd(ChangeEffect changeEffect) {
Transcript tr = changeEffect.getTranscript();
if (tr == null)
throw new RuntimeException("Transcript not found for change:\n\t" + changeEffect);
// Only one exon? Nothing to do (there is no exon-exon junction)
if (tr.numChilds() <= 1) return false;
// Find last valid NMD position
int lastNmdPos = lastNmdPos(tr);
if (lastNmdPos < 0) return false; // No valid 'lastNmdPos'? => There is no NMD event.
// Does this change affect the region 'before' this last NMD position? => It is assumed to be
// NMD
Variant seqChange = changeEffect.getSeqChange();
boolean nmd;
if (tr.isStrandPlus()) nmd = seqChange.getStart() <= lastNmdPos;
else nmd = lastNmdPos <= seqChange.getEnd();
// Update sets and counters
if (nmd) {
transcriptsNmd.add(
changeEffect
.getTranscript()); // Unique transcripts affected (WARNING: null will be added)
genesNmd.add(changeEffect.getGene()); // Unique genes affected (WARNING: null will be added)
nmdCount++;
}
return nmd;
}
/**
* Predict the effect of a seqChange
*
* @param seqChange : Sequence change
* @param seqChangeRef : Before analyzing results, we have to change markers using seqChangerRef
* to create a new reference 'on the fly'
*/
public ChangeEffects seqChangeEffect(Variant seqChange, Variant seqChangeRef) {
ChangeEffects changeEffects = new ChangeEffects(seqChange, seqChangeRef);
// ---
// Chromosome missing?
// ---
if (Config.get().isErrorOnMissingChromo() && isChromosomeMissing(seqChange)) {
changeEffects.addErrorWarning(ErrorWarningType.ERROR_CHROMOSOME_NOT_FOUND);
return changeEffects;
}
// ---
// Check that this is not a huge deletion.
// Huge deletions would crash the rest of the algorithm, so we need to stop them here.
// ---
if (seqChange.isDel() && (seqChange.size() > HUGE_DELETION_SIZE_THRESHOLD)) {
// Get chromosome
String chromoName = seqChange.getChromosomeName();
Chromosome chr = genome.getChromosome(chromoName);
if (chr.size() > 0) {
double ratio = seqChange.size() / ((double) chr.size());
if (ratio > HUGE_DELETION_RATIO_THRESHOLD) {
changeEffects.add(chr, EffectType.CHROMOSOME_LARGE_DELETION, "");
return changeEffects;
}
}
}
// ---
// Query interval tree: Which intervals does seqChange intersect?
// ---
Markers intersects = query(seqChange);
// Show all results
boolean hitChromo = false, hitSomething = false;
if (intersects.size() > 0) {
for (Marker marker : intersects) {
if (marker instanceof Chromosome) hitChromo = true; // Do we hit any chromosome?
else { // Analyze all markers
marker.seqChangeEffect(seqChange, changeEffects, seqChangeRef);
hitSomething = true;
}
}
}
// Any errors or intergenic (i.e. did not hit any gene)
if (!hitChromo) {
if (Config.get().isErrorChromoHit())
changeEffects.addErrorWarning(ErrorWarningType.ERROR_OUT_OF_CHROMOSOME_RANGE);
} else if (!hitSomething) {
if (Config.get().isOnlyRegulation()) changeEffects.setEffectType(EffectType.NONE);
else changeEffects.setEffectType(EffectType.INTERGENIC);
}
return changeEffects;
}
/** Net AA change (only for InDels) */
public String getAaNetChange() {
String aaLong = "", aaShort = "";
if (variant.isIns()) {
aaShort = getAaRef().toUpperCase();
aaLong = getAaAlt().toUpperCase();
} else if (variant.isDel()) {
aaLong = getAaRef().toUpperCase();
aaShort = getAaAlt().toUpperCase();
}
if (aaLong.startsWith(aaShort)) return aaLong.substring(aaShort.length());
if (aaLong.endsWith(aaLong)) return aaLong.substring(0, aaLong.length() - aaShort.length());
return aaLong;
}
@Override
public int compareTo(VariantEffect varEffOther) {
// Sort by impact
int comp = getEffectImpact().compareTo(varEffOther.getEffectImpact());
if (comp != 0) return comp;
// Sort by effect
comp = getEffectType().compareTo(varEffOther.getEffectType());
if (comp != 0) return comp;
// Sort by: Is canonical transcript?
Transcript trThis = getTranscript();
Transcript trOther = varEffOther.getTranscript();
if (trThis != null && trOther != null) {
comp = (trOther.isCanonical() ? 1 : 0) - (trThis.isCanonical() ? 1 : 0);
}
if (comp != 0) return comp;
// Sort by genomic coordinate of affected 'marker'
if ((trThis != null) && (trOther != null)) comp = trThis.compareToPos(trOther);
if (comp != 0) return comp;
// Compare IDs
if ((trThis != null) && (trOther != null)) comp = trThis.getId().compareTo(trOther.getId());
if (comp != 0) return comp;
// Compare by marker
if ((getMarker() != null) && (varEffOther.getMarker() != null))
comp = getMarker().compareToPos(varEffOther.getMarker());
if (comp != 0) return comp;
// Sort by variant (most of the time this is equal)
return variant.compareTo(varEffOther.getVariant());
}
public static void main(String[] args) {
Timer.showStdErr("Start");
// Create an input file iterator
String inputFile = Gpr.HOME + "/snpEff/ins.txt";
VariantFileIterator seqChangeFileIterator = new SeqChangeTxtFileIterator(inputFile);
// ---
// Iterate over input file
// ---
for (Variant seqChange : seqChangeFileIterator) {
System.out.println(seqChange + "\t" + seqChange.getChangeType());
}
// Zzz zzz = new Zzz(null);
// zzz.load("testHg3771Chr1");
// zzz.run();
// Timer.showStdErr("End");
}
/** Return functional class of this effect (i.e. NONSENSE, MISSENSE, SILENT or NONE) */
public FunctionalClass getFunctionalClass() {
if (variant.isSnp()) {
if (!aaAlt.equals(aaRef)) {
CodonTable codonTable = marker.codonTable();
if (codonTable.isStop(codonsAlt)) return FunctionalClass.NONSENSE;
return FunctionalClass.MISSENSE;
}
if (!codonsAlt.equals(codonsRef)) return FunctionalClass.SILENT;
}
return FunctionalClass.NONE;
}
/** Return impact of this effect */
public EffectImpact getEffectImpact() {
if (effectImpact == null) {
if ((variant != null) && (!variant.isVariant())) {
// Not a change? => Modifier
effectImpact = EffectImpact.MODIFIER;
} else {
// Get efefct's type highest impact
effectImpact = EffectImpact.MODIFIER;
for (EffectImpact eimp : effectImpacts)
if (eimp.compareTo(effectImpact) < 0) effectImpact = eimp;
}
}
return effectImpact;
}
/**
* Is this deletion a LOF?
*
* <p>Criteria: 1) First (coding) exon deleted 2) More than 50% of coding sequence deleted
*
* @param changeEffect
* @return
*/
protected boolean isLofDeletion(ChangeEffect changeEffect) {
Transcript tr = changeEffect.getTranscript();
if (tr == null)
throw new RuntimeException("Transcript not found for change:\n\t" + changeEffect);
// ---
// Criteria:
// 1) First (coding) exon deleted
// ---
if (changeEffect.getEffectType() == EffectType.EXON_DELETED) {
Variant seqChange = changeEffect.getSeqChange();
if (seqChange == null)
throw new RuntimeException("Cannot retrieve 'seqChange' from EXON_DELETED effect!");
if (seqChange.includes(tr.getFirstCodingExon())) return true;
}
// ---
// Criteria:
// 2) More than 50% of coding sequence deleted
// ---
// Find coding part of the transcript (i.e. no UTRs)
Variant seqChange = changeEffect.getSeqChange();
int cdsStart = tr.isStrandPlus() ? tr.getCdsStart() : tr.getCdsEnd();
int cdsEnd = tr.isStrandPlus() ? tr.getCdsEnd() : tr.getCdsStart();
Marker coding = new Marker(seqChange.getChromosome(), cdsStart, cdsEnd, 1, "");
// Create an interval intersecting the CDS and the deletion
int start = Math.max(cdsStart, seqChange.getStart());
int end = Math.min(cdsEnd, seqChange.getEnd());
if (start >= end) return false; // No intersections with coding part of the exon? => not LOF
Marker codingDeleted = new Marker(seqChange.getChromosome(), start, end, 1, "");
// Count:
// - number of coding bases deleted
// - number of coding bases
int codingBasesDeleted = 0, codingBases = 0;
for (Exon exon : tr) {
codingBasesDeleted += codingDeleted.intersectSize(exon);
codingBases += coding.intersectSize(exon);
}
// More than a threshold? => It is a LOF
double percDeleted = codingBasesDeleted / ((double) codingBases);
return (percDeleted > deleteProteinCodingBases);
}
/** Annotate a VCF entry */
public boolean annotate(Variant variant, Map<String, String> info) {
if (verbose) Gpr.showMark(++countVariants, SHOW_EVERY);
// Find in database
Collection<DbNsfpEntry> dbEntries = dbNsfp.query(variant);
if (dbEntries == null || dbEntries.isEmpty()) return false;
// Add all INFO fields that refer to this allele
boolean annotated = false;
for (String fieldKey : fieldsToAdd.keySet()) {
// Are there any values to annotate?
String infoValue = getVcfInfo(dbEntries, fieldKey);
// Missing or empty?
if (annotateEmpty) {
if (infoValue.isEmpty()) infoValue = ".";
} else if (isDbNsfpValueEmpty(infoValue)) {
infoValue = null;
}
// Add annotations
if (infoValue != null) {
String oldInfo = info.get(fieldKey);
if (oldInfo == null) oldInfo = "";
info.put(fieldKey, oldInfo + (oldInfo.isEmpty() ? "" : ",") + infoValue);
annotated = true;
}
}
// Show progress
if (annotated) {
countAnnotated++;
if (debug) Gpr.debug("Annotated: " + variant.toStr());
}
return annotated;
}
public String toString(boolean useSeqOntology, boolean useHgvs) {
// Get data to show
String geneId = "", geneName = "", bioType = "", transcriptId = "", exonId = "", customId = "";
int exonRank = -1;
if (marker != null) {
// Gene Id, name and biotype
Gene gene = getGene();
Transcript tr = getTranscript();
// CDS size info
if (gene != null) {
geneId = gene.getId();
geneName = gene.getGeneName();
bioType = getBiotype();
}
// Update trId
if (tr != null) transcriptId = tr.getId();
// Exon rank information
Exon exon = getExon();
if (exon != null) {
exonId = exon.getId();
exonRank = exon.getRank();
}
// Regulation
if (isRegulation()) bioType = ((Regulation) marker).getCellType();
}
// Add seqChage's ID
if (!variant.getId().isEmpty()) customId += variant.getId();
// Add custom markers
if ((marker != null) && (marker instanceof Custom))
customId += (customId.isEmpty() ? "" : ";") + marker.getId();
// CDS length
int cdsSize = getCdsLength();
String errWarn = error + (error.isEmpty() ? "" : "|") + warning;
String aaChange = "";
if (useHgvs) aaChange = getHgvs();
else aaChange = ((aaRef.length() + aaAlt.length()) > 0 ? aaRef + "/" + aaAlt : "");
return errWarn //
+ "\t"
+ geneId //
+ "\t"
+ geneName //
+ "\t"
+ bioType //
+ "\t"
+ transcriptId //
+ "\t"
+ exonId //
+ "\t"
+ (exonRank >= 0 ? exonRank : "") //
+ "\t"
+ effect(false, false, false, useSeqOntology) //
+ "\t"
+ aaChange //
+ "\t"
+ ((codonsRef.length() + codonsAlt.length()) > 0 ? codonsRef + "/" + codonsAlt : "") //
+ "\t"
+ (codonNum >= 0 ? (codonNum + 1) : "") //
+ "\t"
+ (codonDegeneracy >= 0 ? codonDegeneracy + "" : "") //
+ "\t"
+ (cdsSize >= 0 ? cdsSize : "") //
+ "\t"
+ (codonsAroundOld.length() > 0 ? codonsAroundOld + " / " + codonsAroundNew : "") //
+ "\t"
+ (aasAroundOld.length() > 0 ? aasAroundOld + " / " + aasAroundNew : "") //
+ "\t"
+ customId //
;
}
/** Get genotype string */
public String getGenotype() {
if (variant == null) return "";
return variant.getGenotype();
}
/** Codon change string (if it's not too long) */
public String getCodonChangeMax() {
if (variant.size() > MAX_CODON_SEQUENCE_LEN)
return ""; // Cap length in order not to make VCF files grow too much
if (codonsRef.isEmpty() && codonsAlt.isEmpty()) return "";
return codonsRef + "/" + codonsAlt;
}
void analyze(int i, int pos, String ref, String mnp) {
String codons = codons();
Variant seqChange = new Variant(chromosome, pos, ref + "", mnp + "", "");
// ---
// Calculate effects
// ---
VariantEffects effects = snpEffectPredictor.variantEffect(seqChange);
// Show
VariantEffect effect = null;
if (effects.size() > 1) { // Usually there is only one effect
for (VariantEffect ce : effects) {
if ((ce.getEffectType() != EffectType.SPLICE_SITE_ACCEPTOR) //
&& (ce.getEffectType() != EffectType.SPLICE_SITE_DONOR) //
&& (ce.getEffectType() != EffectType.INTRON) //
&& (ce.getEffectType() != EffectType.INTERGENIC) //
) //
effect = ce;
}
} else effect = effects.get();
if (effect != null) {
String effStr = effect.effect(true, true, true, false);
if (codons.length() > 1) {
String codonsExp[] = codons.split("/");
boolean error =
(!codonsExp[0].toUpperCase().equals(effect.getCodonsRef().toUpperCase()) //
|| !codonsExp[1].toUpperCase().equals(effect.getCodonsAlt().toUpperCase()));
if (error || debug) {
Gpr.debug(
"Fatal error:" //
+ "\n\tPos : "
+ pos //
+ "\n\tSeqChange : "
+ seqChange
+ (seqChange.isStrandPlus() ? "+" : "-") //
+ "\n\tCodon (exp) : "
+ codons //
+ "\n\tCodon (pred) : "
+ effect.getCodonsRef().toUpperCase()
+ "/"
+ effect.getCodonsAlt().toUpperCase() //
+ "\n\tEffect (pred) : "
+ effStr //
+ "\n\tEffect (pred) : "
+ effect //
+ "\n\tGene : "
+ gene //
+ "\n\tChromo : "
+ chromoSequence //
);
}
/** Error? Dump so we can debug... */
if (error) {
System.err.println("Error. Dumping data");
Save save = new Save();
save.snpEffectPredictor = snpEffectPredictor;
save.chromoSequence = chromoSequence;
save.chromoNewSequence = chromoNewSequence;
save.ref = ref;
save.pos = pos;
save.mnp = mnp;
String outFile = "/tmp/sep_" + i + "_" + pos + ".bin";
Gpr.toFileSerialize(outFile, save);
throw new RuntimeException("Codons do not match!\n\tData dumped: '" + outFile + "'");
}
}
}
}