Examples with PhysicalLocation ubic.gemma.model.genome.PhysicalLocation used on opensource projects

Example 16 with PhysicalLocation

use of ubic.gemma.model.genome.PhysicalLocation in project Gemma by PavlidisLab.

the class NcbiGeneConverter method getPhysicalLocation.

private PhysicalLocation getPhysicalLocation(NCBIGene2Accession acc, Gene gene) {
    PhysicalLocation pl = PhysicalLocation.Factory.newInstance();
    pl.setChromosome(gene.getPhysicalLocation().getChromosome());
    if (acc.getOrientation() != null) {
        pl.setStrand(acc.getOrientation());
    }
    if (acc.getStartPosition() != null) {
        pl.setNucleotide(acc.getStartPosition());
        pl.setNucleotideLength((int) Math.abs(acc.getEndPosition() - acc.getStartPosition()));
        pl.setBin(SequenceBinUtils.binFromRange(acc.getStartPosition().intValue(), acc.getEndPosition().intValue()));
    }
    return pl;
}

Example 17 with PhysicalLocation

use of ubic.gemma.model.genome.PhysicalLocation in project Gemma by PavlidisLab.

the class GoldenPathSequenceAnalysis method findClosestGene.

/**
 * Given a location, find the nearest gene on the same strand, including only "known", "refseq" or "ensembl"
 * transcripts.
 *
 * @param chromosome chromosome
 * @param queryStart start
 * @param queryEnd end
 * @param strand Either '+' or '-'
 * @param maxWindow the number of bases on each side to look, at most, in addition to looking inside the given
 *        region.
 * @return the Gene closest to the given location. This is a transient instance, not from Gemma's database.
 */
public Gene findClosestGene(String chromosome, Long queryStart, Long queryEnd, String strand, int maxWindow) {
    if (queryEnd < queryStart)
        throw new IllegalArgumentException("End must not be less than start");
    long round = 0L;
    int numRounds = 5;
    int increment = (int) (maxWindow / (double) numRounds);
    while (round < numRounds) {
        long left = queryStart + round * increment;
        long right = queryEnd + round * increment;
        Collection<GeneProduct> geneProducts = this.findRefGenesByLocation(chromosome, left, right, strand);
        geneProducts.addAll(this.findKnownGenesByLocation(chromosome, left, right, strand));
        Gene nearest = null;
        int closestSoFar = Integer.MAX_VALUE;
        for (GeneProduct geneProduct : geneProducts) {
            PhysicalLocation gpl = geneProduct.getPhysicalLocation();
            Long start = gpl.getNucleotide();
            Long end = start + gpl.getNucleotideLength();
            int gap = (int) Math.min(left - end, start - right);
            if (gap < closestSoFar) {
                closestSoFar = gap;
                nearest = geneProduct.getGene();
            }
        }
        if (nearest != null)
            return nearest;
        round++;
    }
    return null;
}

Example 18 with PhysicalLocation

use of ubic.gemma.model.genome.PhysicalLocation in project Gemma by PavlidisLab.

the class GoldenPathSequenceAnalysis method getExons.

/**
 * Fill in the exon information for a gene, given the raw blobs from the GoldenPath database.
 * Be sure to pass the right Blob arguments!
 *
 * @param exonStarts starts
 * @param exonEnds ends
 * @throws SQLException sql problem
 */
private Collection<PhysicalLocation> getExons(Chromosome chrom, Blob exonStarts, Blob exonEnds) throws SQLException {
    Collection<PhysicalLocation> exons = new HashSet<>();
    if (exonStarts == null || exonEnds == null) {
        return exons;
    }
    String exonStartLocations = SQLUtils.blobToString(exonStarts);
    String exonEndLocations = SQLUtils.blobToString(exonEnds);
    int[] exonStartsInts = SequenceManipulation.blatLocationsToIntArray(exonStartLocations);
    int[] exonEndsInts = SequenceManipulation.blatLocationsToIntArray(exonEndLocations);
    assert exonStartsInts.length == exonEndsInts.length;
    for (int i = 0; i < exonEndsInts.length; i++) {
        int exonStart = exonStartsInts[i];
        int exonEnd = exonEndsInts[i];
        PhysicalLocation exon = PhysicalLocation.Factory.newInstance();
        exon.setChromosome(chrom);
        assert chrom.getTaxon() != null;
        exon.setNucleotide((long) exonStart);
        exon.setNucleotideLength(exonEnd - exonStart);
        exon.setBin(SequenceBinUtils.binFromRange(exonStart, exonEnd));
        exons.add(exon);
    }
    return exons;
}

Example 19 with PhysicalLocation

use of ubic.gemma.model.genome.PhysicalLocation in project Gemma by PavlidisLab.

the class GoldenPathSequenceAnalysis method computeLocationInGene.

/**
 * Given a location and a gene product, compute the distance from the 3' end of the gene product as well as the
 * amount of overlap. If the location has low overlaps with known exons (threshold set by
 * RECHECK_OVERLAP_THRESHOLD), we optionally search for mRNAs in the region. If there are overlapping mRNAs, we use
 * the best overlap value. If the overlap is still not high enough we optionally check ESTs.
 *
 * @param chromosome chromosome
 * @param queryStart start
 * @param queryEnd end
 * @param starts Start locations of alignments of the query (target coordinates)
 * @param sizes Sizes of alignments of the query.
 * @param geneProduct GeneProduct with which the overlap and distance is to be computed.
 * @param method method
 * @param config The useEsts and useRNA options are relevant
 * @return a ThreePrimeData object containing the results.
 */
private BlatAssociation computeLocationInGene(String chromosome, Long queryStart, Long queryEnd, String starts, String sizes, GeneProduct geneProduct, ThreePrimeDistanceMethod method, ProbeMapperConfig config) {
    assert geneProduct != null : "GeneProduct is null";
    BlatAssociation blatAssociation = BlatAssociation.Factory.newInstance();
    blatAssociation.setGeneProduct(geneProduct);
    blatAssociation.setThreePrimeDistanceMeasurementMethod(method);
    PhysicalLocation geneLoc = geneProduct.getPhysicalLocation();
    assert geneLoc != null : "PhysicalLocation for GeneProduct " + geneProduct + " is null";
    assert geneLoc.getNucleotide() != null;
    int geneStart = geneLoc.getNucleotide().intValue();
    int geneEnd = geneLoc.getNucleotide().intValue() + geneLoc.getNucleotideLength();
    int exonOverlap = 0;
    if (starts != null & sizes != null) {
        exonOverlap = SequenceManipulation.getGeneProductExonOverlap(starts, sizes, geneLoc.getStrand(), geneProduct);
        int totalSize = SequenceManipulation.totalSize(sizes);
        if (config.isUseMrnas() && exonOverlap / (double) (totalSize) < GoldenPathSequenceAnalysis.RECHECK_OVERLAP_THRESHOLD) {
            int newOverlap = this.checkRNAs(chromosome, queryStart, queryEnd, starts, sizes, exonOverlap, geneLoc.getStrand(), geneProduct.getGene());
            if (newOverlap > exonOverlap) {
                GoldenPath.log.debug("mRNA overlap was higher than primary transcript");
                exonOverlap = newOverlap;
            }
        }
        if (config.isUseEsts() && exonOverlap / (double) (totalSize) < GoldenPathSequenceAnalysis.RECHECK_OVERLAP_THRESHOLD) {
            int newOverlap = this.checkESTs(chromosome, queryStart, queryEnd, starts, sizes, exonOverlap, geneLoc.getStrand());
            if (newOverlap > exonOverlap) {
                GoldenPath.log.debug("Exon overlap was higher than mrna or  primary transcript");
                exonOverlap = newOverlap;
            }
        }
        assert exonOverlap <= totalSize;
    }
    blatAssociation.setOverlap(exonOverlap);
    if (method == ThreePrimeDistanceMethod.MIDDLE) {
        int center = SequenceManipulation.findCenter(starts, sizes);
        if (geneLoc.getStrand().equals("+")) {
            // then the 3' end is at the 'end'. : >>>>>>>>>>>>>>>>>>>>>*>>>>> (* is where we might be)
            blatAssociation.setThreePrimeDistance((long) Math.max(0, geneEnd - center));
        } else if (geneProduct.getPhysicalLocation().getStrand().equals("-")) {
            // then the 3' end is at the 'start'. : <<<*<<<<<<<<<<<<<<<<<<<<<<<
            blatAssociation.setThreePrimeDistance((long) Math.max(0, center - geneStart));
        } else {
            throw new IllegalArgumentException("Strand wasn't '+' or '-'");
        }
    } else if (method == ThreePrimeDistanceMethod.RIGHT) {
        if (geneLoc.getStrand().equals("+")) {
            // then the 3' end is at the 'end'. : >>>>>>>>>>>>>>>>>>>>>*>>>>> (* is where we might be)
            blatAssociation.setThreePrimeDistance(Math.max(0, geneEnd - queryEnd));
        } else if (geneProduct.getPhysicalLocation().getStrand().equals("-")) {
            // then the 3' end is at the 'start'. : <<<*<<<<<<<<<<<<<<<<<<<<<<<
            blatAssociation.setThreePrimeDistance(Math.max(0, queryStart - geneStart));
        } else {
            throw new IllegalArgumentException("Strand wasn't '+' or '-'");
        }
    } else if (method == ThreePrimeDistanceMethod.LEFT) {
        throw new UnsupportedOperationException("Left edge measure not supported");
    } else {
        throw new IllegalArgumentException("Unknown method");
    }
    return blatAssociation;
}

Example 20 with PhysicalLocation

use of ubic.gemma.model.genome.PhysicalLocation in project Gemma by PavlidisLab.

the class GeneServiceImpl method getPhysicalLocationsValueObjects.

@Override
@Transactional(readOnly = true)
public Collection<PhysicalLocationValueObject> getPhysicalLocationsValueObjects(Gene gene) {
    if (gene == null) {
        return Collections.emptyList();
    }
    gene = this.thaw(gene);
    Collection<GeneProduct> gpCollection = gene.getProducts();
    Collection<PhysicalLocationValueObject> locations = new LinkedList<>();
    if (gpCollection == null)
        return null;
    for (GeneProduct gp : gpCollection) {
        PhysicalLocation physicalLocation = gp.getPhysicalLocation();
        if (physicalLocation == null) {
            AbstractService.log.warn(gene.getOfficialSymbol() + " product " + gp.getName() + " (id:" + gp.getId() + ") has no location.");
            continue;
        }
        // Only add if the physical location of the product is different from any we already know.
        PhysicalLocationValueObject vo = new PhysicalLocationValueObject(physicalLocation);
        if (!locations.contains(vo)) {
            locations.add(vo);
        }
    }
    return locations;
}