@Override
public boolean run() {
for (Gene g : config.getSnpEffectPredictor().getGenome().getGenes()) {
// System.out.println(g.getGeneName());
for (Transcript tr : g) {
if (!tr.isProteinCoding()) continue;
if (tr.introns().size() < 2) continue;
// System.out.println("\t" + tr.getId());
for (Intron i : tr.introns()) {
int pos = i.getStart() + (int) (Math.random() * (i.size() - 2)) + 1;
String line =
i.getChromosomeName()
+ "\t"
+ pos
+ "\t.\tA\tT\t.\t.\tAC=1;GENE="
+ g.getGeneName()
+ ";TR="
+ tr.getId()
+ ";INTRON="
+ i.getRank();
System.out.println(line);
sb.append(line + "\n");
}
}
}
Gpr.toFile(Gpr.HOME + "/introns_test.vcf", sb);
return true;
}
/** Show a genome in a 'standard' way */
String show(Genome genome) {
StringBuilder sb = new StringBuilder();
// Genome
sb.append(genome.getVersion() + "\n");
// Chromosomes
for (Chromosome chr : genome) sb.append(chr + "\n");
// Genes
ArrayList<Gene> genes = new ArrayList<Gene>();
// Sort genes
for (Gene gene : genome.getGenes()) genes.add(gene);
Collections.sort(genes);
// Show genes
for (Gene gene : genes) {
sb.append(gene);
for (Transcript tr : gene.sortedStrand())
sb.append("\t\tCDS '" + tr.getId() + "': " + tr.cds() + "\n");
}
return sb.toString();
}
/**
* Add into to a hash
*
* @param hits
* @param marker
* @param hit2add
* @param showGeneDetails
* @param compareTemplate
*/
void regionsAddHit(
HashSet<String> hits,
Marker hit2add,
Marker marker,
boolean showGeneDetails,
boolean compareTemplate) {
String hitStr = hit2add.getClass().getSimpleName();
if (compareTemplate) {
Gene gene = (Gene) hit2add.findParent(Gene.class);
if (gene != null)
hitStr +=
(hit2add.isStrandPlus() == marker.isStrandPlus())
? "_TEMPLATE_STRAND"
: "_NON_TEMPLATE_STRAND";
}
if (showGeneDetails && (hit2add instanceof Gene)) {
Gene gene = (Gene) hit2add;
hitStr +=
"["
+ gene.getBioType()
+ ", "
+ gene.getGeneName()
+ ", "
+ (gene.isProteinCoding() ? "protein" : "not-protein")
+ "]";
}
hits.add(hitStr); // Add marker name to the list
}
/**
* Return a collection of intervals that intersect 'marker' Query resulting genes, transcripts and
* exons to get ALL types of intervals possible
*
* @return
*/
public Markers queryDeep(Marker marker) {
if (Config.get().isErrorOnMissingChromo() && isChromosomeMissing(marker))
throw new RuntimeEOFException("Chromosome missing for marker: " + marker);
boolean hitChromo = false;
Markers hits = new Markers();
Markers intersects = query(marker);
if (intersects.size() > 0) {
for (Marker m : intersects) {
hits.add(m);
if (m instanceof Chromosome) {
hitChromo = true; // OK (we have to hit a chromosome, otherwise it's an error
} else if (m instanceof Gene) {
// Analyze Genes
Gene gene = (Gene) m;
hits.addAll(gene.query(marker));
}
}
}
if (!hitChromo && Config.get().isErrorChromoHit())
throw new RuntimeException("ERROR: Out of chromosome range. " + marker);
return hits;
}
/** Get biotype */
public String getBiotype() {
Gene gene = getGene();
if (gene == null) return "";
Transcript tr = getTranscript();
if (tr != null) return tr.getBioType();
else if (gene.getGenome().hasCodingInfo())
return (gene.isProteinCoding() ? "coding" : "non-coding");
return "";
}
/** Count bases covered for each marker type */
public void countBases() {
// ---
// Add all markers
// ---
Markers markers = new Markers();
markers.add(snpEffectPredictor.getMarkers());
for (Gene gene : snpEffectPredictor.getGenome().getGenes()) {
markers.add(gene);
markers.add(gene.markers());
}
for (Chromosome chr : snpEffectPredictor.getGenome()) markers.add(chr);
// ---
// Calculate raw counts
// ---
for (Marker m : markers) {
String mtype = markerTypes.getType(m);
String msubtype = markerTypes.getSubType(m);
rawCountMarkers.inc(mtype);
rawCountBases.inc(mtype, m.size());
// Count sub-types (if any)
if (msubtype != null) {
rawCountMarkers.inc(msubtype);
rawCountBases.inc(msubtype, m.size());
}
}
// ---
// Count number of bases for each marker type (overlap and join)
// ---
for (String mtype : rawCountMarkers.keysSorted()) {
if (mtype.equals(Chromosome.class.getSimpleName()))
continue; // We calculate chromosomes later (it's faster)
if (verbose) System.err.print(mtype + ":");
if (countMarkers.get(mtype) == 0) {
for (Chromosome chr : snpEffectPredictor.getGenome()) countBases(mtype, chr, markers);
}
if (verbose) System.err.println("");
}
// Show chromosomes length
String mtype = Chromosome.class.getSimpleName();
for (Chromosome chr : snpEffectPredictor.getGenome()) {
countBases.inc(mtype, chr.size());
countMarkers.inc(mtype);
}
}
public void testCase_05_PaeruPA14muccA() {
String genome = "paeru.PA14";
String gff3File = "tests/paeru.PA14.muccA.gff";
String resultFile = "tests/paeru.PA14.muccA.txt";
SnpEffectPredictor sep = buildAndCompare(genome, gff3File, resultFile, true, false);
// Make sure no splice site is added
Gene gene = sep.getGenome().getGenes().iterator().next();
Transcript tr = gene.iterator().next();
List<SpliceSite> spliceSites = tr.createSpliceSites(SpliceSite.CORE_SPLICE_SITE_SIZE, 0, 0, 0);
Assert.assertEquals(0, spliceSites.size());
}
/**
* Find closest gene to this marker
*
* <p>In case more than one 'closest' gene is found (e.g. two or more genes at the same distance).
* The following rules apply:
*
* <p>i) If many genes have the same 'closest distance', coding genes are preferred.
*
* <p>ii) If more than one coding gene has the same 'closet distance', a random gene is returned.
*
* @param inputInterval
*/
public Gene queryClosestGene(Marker inputInterval) {
int initialExtension = 1000;
String chrName = inputInterval.getChromosomeName();
Chromosome chr = genome.getChromosome(chrName);
if (chr == null) return null;
if (chr.size() > 0) {
// Extend interval to capture 'close' genes
for (int extend = initialExtension; extend < chr.size(); extend *= 2) {
int start = Math.max(inputInterval.getStart() - extend, 0);
int end = inputInterval.getEnd() + extend;
Marker extended = new Marker(chr, start, end, 1, "");
// Find all genes that intersect with the interval
Markers markers = query(extended);
Markers genes = new Markers();
int minDist = Integer.MAX_VALUE;
for (Marker m : markers) {
if (m instanceof Gene) {
int dist = m.distance(inputInterval);
if (dist < minDist) {
genes.add(m);
minDist = dist;
}
}
}
// Found something?
if (genes.size() > 0) {
// Find a gene having distance 'minDist'. Prefer coding genes
Gene minDistGene = null;
for (Marker m : genes) {
int dist = m.distance(inputInterval);
if (dist == minDist) {
Gene gene = (Gene) m;
if (minDistGene == null) minDistGene = gene;
else if (!minDistGene.isProteinCoding() && gene.isProteinCoding()) minDistGene = gene;
}
}
return minDistGene;
}
}
}
// Nothing found
return null;
}
/**
* What percentile of the transcripts in this gene are affected?
*
* @param gene
* @return
*/
double percentOfTranscriptsAffected(Gene gene, HashSet<Transcript> transcripts) {
if (gene == null) return 0;
// Count how many transcript are affected in each gene
int countAffected = 0;
for (Transcript tr : gene) if (transcripts.contains(tr)) countAffected++;
return countAffected / ((double) gene.numChilds());
}
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 //
;
}
/**
* 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 isLof(ChangeEffect changeEffect) {
// Not a sequence change? => Not LOF
if ((changeEffect.getSeqChange() != null) && (!changeEffect.getSeqChange().isVariant()))
return false;
// Is this change affecting a protein coding gene?
Gene gene = changeEffect.getGene();
Transcript tr = changeEffect.getTranscript();
if ((gene == null) // No gene affected?
|| (tr == null) // No transcript affected?
|| (!gene.isProteinCoding()
&& !config.isTreatAllAsProteinCoding()) // Not a protein coding gene?
|| (!tr.isProteinCoding()
&& !config.isTreatAllAsProteinCoding()) // Not a protein coding transcript?
) return false;
// ---
// Is this variant a LOF?
// ---
boolean lof = false;
// Frame shifts
if (changeEffect.getEffectType() == EffectType.FRAME_SHIFT) {
// It is assumed that even with a protein coding change at the last 5% of the protein, the
// protein could still be functional.
double perc = percentCds(changeEffect);
lof |= (ignoreProteinCodingBefore <= perc) && (perc <= ignoreProteinCodingAfter);
}
// Deletion? Is another method to check
if (changeEffect.getSeqChange().isDel()) lof |= isLofDeletion(changeEffect);
// The following effect types can be considered LOF
switch (changeEffect.getEffectType()) {
case SPLICE_SITE_ACCEPTOR:
case SPLICE_SITE_DONOR:
// Core splice sites are considered LOF
if ((changeEffect.getMarker() != null)
&& (changeEffect.getMarker() instanceof SpliceSite)) {
// Get splice site marker and check if it is 'core'
SpliceSite spliceSite = (SpliceSite) changeEffect.getMarker();
if (spliceSite.intersectsCoreSpliceSite(changeEffect.getSeqChange()))
lof = true; // Does it intersect the CORE splice site?
}
break;
case STOP_GAINED:
lof |= isNmd(changeEffect);
break;
case RARE_AMINO_ACID:
case START_LOST:
// This one is not in the referenced papers, but we assume that RARE AA and START_LOSS
// changes are damaging.
lof = true;
break;
default: // All others are not considered LOF
break;
}
// Update sets
if (lof) {
transcriptsLof.add(
changeEffect
.getTranscript()); // Unique transcripts affected (WARNING: null will be added)
genesLof.add(changeEffect.getGene()); // Unique genes affected (WARNING: null will be added)
}
return lof;
}
/**
* Remove all transcripts that are NOT in the list
*
* @return : Number of transcripts removed
*/
public int retainAllTranscripts(Set<String> trIds) {
int total = 0;
for (Gene g : genome.getGenes()) total += g.keepTranscripts(trIds);
return total;
}
/** Remove all non-canonical transcripts */
public void removeNonCanonical() {
for (Gene g : genome.getGenes()) g.removeNonCanonical();
}
/**
* Name of the regions hit by a marker
*
* @param marker
* @param showGeneDetails
* @param compareTemplate
* @param id : Only use genes or transcripts matching this ID
* @return
*/
public Set<String> regions(
Marker marker, boolean showGeneDetails, boolean compareTemplate, String id) {
if (Config.get().isErrorOnMissingChromo() && isChromosomeMissing(marker))
throw new RuntimeEOFException("Chromosome missing for marker: " + marker);
boolean hitChromo = false;
HashSet<String> hits = new HashSet<String>();
Markers intersects = query(marker);
if (intersects.size() > 0) {
for (Marker markerInt : intersects) {
if (markerInt instanceof Chromosome) {
hitChromo = true; // OK (we have to hit a chromosome, otherwise it's an error
hits.add(markerInt.getClass().getSimpleName()); // Add marker name to the list
} else if (markerInt instanceof Gene) {
// Analyze Genes
Gene gene = (Gene) markerInt;
regionsAddHit(hits, gene, marker, showGeneDetails, compareTemplate);
// For all transcripts...
for (Transcript tr : gene) {
if ((id == null)
|| gene.getId().equals(id)
|| tr.getId().equals(id)) { // Mathes ID? (...or no ID to match)
// Does it intersect this transcript?
if (tr.intersects(marker)) {
regionsAddHit(hits, tr, marker, showGeneDetails, compareTemplate);
// Does it intersect a UTR?
for (Utr utr : tr.getUtrs())
if (utr.intersects(marker))
regionsAddHit(hits, utr, marker, showGeneDetails, compareTemplate);
// Does it intersect an exon?
for (Exon ex : tr)
if (ex.intersects(marker))
regionsAddHit(hits, ex, marker, showGeneDetails, compareTemplate);
// Does it intersect an intron?
for (Intron intron : tr.introns())
if (intron.intersects(marker))
regionsAddHit(hits, intron, marker, showGeneDetails, compareTemplate);
}
}
}
} else {
// No ID to match?
if (id == null) regionsAddHit(hits, markerInt, marker, showGeneDetails, compareTemplate);
else {
// Is ID from transcript?
Transcript tr = (Transcript) markerInt.findParent(Transcript.class);
if ((tr != null) && (tr.getId().equals(id))) {
regionsAddHit(
hits,
markerInt,
marker,
showGeneDetails,
compareTemplate); // Transcript ID matches => count
} else {
// Is ID from gene?
Gene gene = (Gene) markerInt.findParent(Gene.class);
if ((gene != null) && (gene.getId().equals(id)))
regionsAddHit(
hits,
markerInt,
marker,
showGeneDetails,
compareTemplate); // Gene ID matches => count
}
}
}
}
}
if (!hitChromo) throw new RuntimeException("ERROR: Out of chromosome range. " + marker);
return hits;
}