/**
* 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;
}
/**
* 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
}
/** 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 "";
}
/**
* 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;
}