/**
* 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;
}
/**
* 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;
}
/**
* Count number of bases, for a given chromosome and marker type
*
* @param mtype
* @param chr
* @param markers
* @return
*/
void countBases(String mtype, Chromosome chr, Markers markers) {
String chrName = chr.getChromosomeName();
if (verbose) System.err.print(" " + chrName);
// Initialize
byte busy[] = new byte[chr.size()];
for (int i = 0; i < busy.length; i++) busy[i] = 0;
for (Marker m : markers) {
// Same marker type & same chromo? Count bases
if (m.getChromosomeName().equals(chrName) && markerTypes.isType(m, mtype)) {
for (int i = m.getStart(); i <= m.getEnd(); i++) busy[i] = 1;
}
}
int latest = 0;
for (int i = 0; i < busy.length; i++) {
// Transition? Count another marker
if ((i > 0) && (busy[i] != 0) && (busy[i - 1] == 0)) {
if ((i - latest) <= readLength)
countBases.inc(mtype, i - latest); // Intervals are less than one read away? Unify them
else countMarkers.inc(mtype);
}
// Base busy? Count another base
if (busy[i] != 0) {
countBases.inc(mtype);
latest = i;
}
}
}
/** 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);
}
}
/**
* Is the chromosome missing in this marker?
*
* @param marker
* @return
*/
boolean isChromosomeMissing(Marker marker) {
// Missing chromosome in marker?
if (marker.getChromosome() == null) return true;
// Missing chromosome in genome?
String chrName = marker.getChromosomeName();
Chromosome chr = genome.getChromosome(chrName);
if (chr == null) return true;
// Chromosome length is 1 or less?
if (chr.size() < 1) return true;
// Tree not found in interval forest?
if (!intervalForest.hasTree(chrName)) return true;
// OK, we have the chromosome
return false;
}