Examples with RiskAllele uk.ac.ebi.spot.goci.model.RiskAllele used on opensource projects

Example 6 with RiskAllele

use of uk.ac.ebi.spot.goci.model.RiskAllele in project goci by EBISPOT.

the class DefaultGWASOWLConverter method convertAssociation.

protected void convertAssociation(Association association, OWLOntology ontology, Set<String> issuedWarnings) {
    // get the trait association class
    OWLClass taClass = getDataFactory().getOWLClass(IRI.create(OntologyConstants.TRAIT_ASSOCIATION_CLASS_IRI));
    IRI taIndIRI = getMinter().mint(OntologyConstants.GWAS_ONTOLOGY_BASE_IRI, association);
    // create a new trait association instance
    OWLNamedIndividual taIndiv = getDataFactory().getOWLNamedIndividual(taIndIRI);
    // assert class membership
    OWLClassAssertionAxiom classAssertion = getDataFactory().getOWLClassAssertionAxiom(taClass, taIndiv);
    getManager().addAxiom(ontology, classAssertion);
    // get datatype relations
    OWLDataProperty has_p_value = getDataFactory().getOWLDataProperty(IRI.create(OntologyConstants.HAS_P_VALUE_PROPERTY_IRI));
    // get annotation relations
    OWLAnnotationProperty rdfsLabel = getDataFactory().getOWLAnnotationProperty(OWLRDFVocabulary.RDFS_LABEL.getIRI());
    //pvalue but says it was less then 10-6. So if we have no pvalue we just don't add it.
    if (association.getPvalueMantissa() != null && association.getPvalueExponent() != null) {
        double pval = association.getPvalueMantissa() * Math.pow(10, association.getPvalueExponent());
        OWLLiteral pValue = getDataFactory().getOWLLiteral(pval);
        //            OWLLiteral pValue = getDataFactory().getOWLLiteral(association.getPvalueMantissa()+"e"+association.getPvalueExponent());
        OWLDataPropertyAssertionAxiom p_value_relation = getDataFactory().getOWLDataPropertyAssertionAxiom(has_p_value, taIndiv, pValue);
        AddAxiom add_p_value = new AddAxiom(ontology, p_value_relation);
        getManager().applyChange(add_p_value);
    }
    // get the snp instance for this association
    OWLNamedIndividual snpIndiv;
    String rsId = null;
    for (Locus locus : association.getLoci()) {
        for (RiskAllele riskAllele : locus.getStrongestRiskAlleles()) {
            SingleNucleotidePolymorphism snp = riskAllele.getSnp();
            rsId = snp.getRsId();
            snpIndiv = getDataFactory().getOWLNamedIndividual(getMinter().mint(OntologyConstants.GWAS_ONTOLOGY_BASE_IRI, snp));
            if (snpIndiv == null) {
                String warning = "A new SNP with the given RSID only will be created";
                if (!issuedWarnings.contains(warning)) {
                    getLog().warn(warning);
                    issuedWarnings.add(warning);
                }
                snpIndiv = getDataFactory().getOWLNamedIndividual(getMinter().mint(OntologyConstants.GWAS_ONTOLOGY_BASE_IRI, "SingleNucleotidePolymorphism", snp.getRsId(), true));
                // assert class membership
                OWLClass snpClass = getDataFactory().getOWLClass(IRI.create(OntologyConstants.SNP_CLASS_IRI));
                OWLClassAssertionAxiom snpClassAssertion = getDataFactory().getOWLClassAssertionAxiom(snpClass, snpIndiv);
                getManager().addAxiom(ontology, snpClassAssertion);
                // assert rsid relation
                OWLDataProperty has_snp_rsid = getDataFactory().getOWLDataProperty(IRI.create(OntologyConstants.HAS_SNP_REFERENCE_ID_PROPERTY_IRI));
                OWLLiteral rsid = getDataFactory().getOWLLiteral(snp.getRsId());
                OWLDataPropertyAssertionAxiom rsid_relation = getDataFactory().getOWLDataPropertyAssertionAxiom(has_snp_rsid, snpIndiv, rsid);
                AddAxiom add_rsid = new AddAxiom(ontology, rsid_relation);
                getManager().applyChange(add_rsid);
                // assert label
                OWLAnnotationAssertionAxiom snp_label_annotation = getDataFactory().getOWLAnnotationAssertionAxiom(rdfsLabel, snpIndiv.getIRI(), rsid);
                AddAxiom add_snp_label = new AddAxiom(ontology, snp_label_annotation);
                getManager().applyChange(add_snp_label);
            }
            // get object properties
            OWLObjectProperty has_subject = getDataFactory().getOWLObjectProperty(IRI.create(OntologyConstants.HAS_SUBJECT_IRI));
            OWLObjectProperty is_subject_of = getDataFactory().getOWLObjectProperty(IRI.create(OntologyConstants.IS_SUBJECT_OF_IRI));
            // assert relations
            OWLObjectPropertyAssertionAxiom has_subject_snp_relation = getDataFactory().getOWLObjectPropertyAssertionAxiom(has_subject, taIndiv, snpIndiv);
            AddAxiom add_has_subject_snp = new AddAxiom(ontology, has_subject_snp_relation);
            getManager().applyChange(add_has_subject_snp);
            OWLObjectPropertyAssertionAxiom is_subject_of_snp_relation = getDataFactory().getOWLObjectPropertyAssertionAxiom(is_subject_of, snpIndiv, taIndiv);
            AddAxiom add_is_subject_of_snp = new AddAxiom(ontology, is_subject_of_snp_relation);
            getManager().applyChange(add_is_subject_of_snp);
        }
        // get the EFO class for the trait
        for (EfoTrait efoTrait : association.getEfoTraits()) {
            OWLClass traitClass;
            traitClass = getDataFactory().getOWLClass(IRI.create(efoTrait.getUri()));
            if (traitClass == null) {
                String warning = "This trait will be mapped to Experimental Factor";
                if (!issuedWarnings.contains(warning)) {
                    getLog().warn(warning);
                    issuedWarnings.add(warning);
                }
                traitClass = getDataFactory().getOWLClass(IRI.create(OntologyConstants.EXPERIMENTAL_FACTOR_CLASS_IRI));
            }
            // create a new trait instance (puns the class)
            IRI traitIRI = traitClass.getIRI();
            OWLNamedIndividual traitIndiv = getDataFactory().getOWLNamedIndividual(traitIRI);
            if (ontology.containsIndividualInSignature(traitIRI)) {
                getLog().trace("Trait individual '" + traitIRI.toString() + "' (type: " + traitClass + ") already exists");
            } else {
                getLog().trace("Creating trait individual '" + traitIRI.toString() + "' (type: " + traitClass + ")");
            }
            // and also add the gwas label to the individual so we don't lose curated data
            OWLDataProperty has_gwas_trait_name = getDataFactory().getOWLDataProperty(IRI.create(OntologyConstants.HAS_GWAS_TRAIT_NAME_PROPERTY_IRI));
            OWLLiteral gwasTrait = getDataFactory().getOWLLiteral(association.getStudy().getDiseaseTrait().getTrait());
            OWLDataPropertyAssertionAxiom gwas_trait_relation = getDataFactory().getOWLDataPropertyAssertionAxiom(has_gwas_trait_name, taIndiv, gwasTrait);
            AddAxiom add_gwas_trait_name = new AddAxiom(ontology, gwas_trait_relation);
            getManager().applyChange(add_gwas_trait_name);
            // assert class membership
            OWLClassAssertionAxiom traitClassAssertion = getDataFactory().getOWLClassAssertionAxiom(traitClass, traitIndiv);
            getManager().addAxiom(ontology, traitClassAssertion);
            // get object properties
            OWLObjectProperty has_object = getDataFactory().getOWLObjectProperty(IRI.create(OntologyConstants.HAS_OBJECT_IRI));
            OWLObjectProperty is_object_of = getDataFactory().getOWLObjectProperty(IRI.create(OntologyConstants.IS_OBJECT_OF_IRI));
            // assert relations
            OWLObjectPropertyAssertionAxiom has_object_trait_relation = getDataFactory().getOWLObjectPropertyAssertionAxiom(has_object, taIndiv, traitIndiv);
            AddAxiom add_has_object_trait = new AddAxiom(ontology, has_object_trait_relation);
            getManager().applyChange(add_has_object_trait);
            OWLObjectPropertyAssertionAxiom is_object_of_trait_relation = getDataFactory().getOWLObjectPropertyAssertionAxiom(is_object_of, traitIndiv, taIndiv);
            AddAxiom add_is_object_of_trait = new AddAxiom(ontology, is_object_of_trait_relation);
            getManager().applyChange(add_is_object_of_trait);
        }
        // finally, assert label for this association
        OWLLiteral label = getDataFactory().getOWLLiteral("Association between " + rsId + " and " + association.getStudy().getDiseaseTrait().getTrait());
        OWLAnnotationAssertionAxiom label_annotation = getDataFactory().getOWLAnnotationAssertionAxiom(rdfsLabel, taIndiv.getIRI(), label);
        AddAxiom add_band_label = new AddAxiom(ontology, label_annotation);
        getManager().applyChange(add_band_label);
    }
}

Example 7 with RiskAllele

use of uk.ac.ebi.spot.goci.model.RiskAllele in project goci by EBISPOT.

the class DefaultGWASOWLPublisher method validateGWASData.

/**
     * Validates the data obtained from the GWAS catalog (prior to converting to OWL)
     *
     * @param studies the set of studies to validate
     */
protected void validateGWASData(Collection<Study> studies) {
    // now check a random assortment of 5 studies for trait associations, abandoning broken ones
    int count = 0;
    int noAssocCount = 0;
    int termMismatches = 0;
    for (Study study : studies) {
        //            try {
        Collection<Association> associations = study.getAssociations();
        getLog().debug("Study (PubMed ID '" + study.getPubmedId() + "') had " + associations.size() + " associations");
        if (associations.size() > 0) {
            for (Association association : associations) {
                String efoTraitsDashSepList = "";
                for (EfoTrait efoTrait : association.getEfoTraits()) {
                    if ("".equals(efoTraitsDashSepList)) {
                        efoTraitsDashSepList.concat(efoTrait.getTrait());
                    } else {
                        efoTraitsDashSepList.concat(", " + efoTrait.getTrait());
                    }
                }
                for (Locus locus : association.getLoci()) {
                    for (RiskAllele riskAllele : locus.getStrongestRiskAlleles()) {
                        getLog().debug(//                                "    Association: SNP '" + association.getAssociatedSNP().getRSID() +
                        "    Association: SNP '" + riskAllele.getSnp().getRsId() + "' <-> Trait '" + efoTraitsDashSepList.toString() + "'");
                    }
                }
            }
            count++;
        } else {
            noAssocCount++;
        }
    }
    int eligCount = studies.size() - noAssocCount;
    int correctCount = count + termMismatches;
    getLog().info("\n\nREPORT:\n" + eligCount + "/" + studies.size() + " declared associations and therefore could usefully be mapped.\n" + (eligCount - count - termMismatches) + "/" + eligCount + " failed due to data integrity concerns.\n" + count + "/" + correctCount + " studies could be completely mapped after passing all checks.\n" + termMismatches + "/" + correctCount + " failed due to missing or duplicated terms in EFO");
}

Example 8 with RiskAllele

use of uk.ac.ebi.spot.goci.model.RiskAllele in project goci by EBISPOT.

the class DefaultGWASOWLPublisher method filterAndPublishGWASData.

private OWLOntology filterAndPublishGWASData(OWLOntology conversion, Collection<Study> studies) throws OWLConversionException {
    //TODO : check with tony : Discard studies which are not yet associated with a trait.
    //Discard studies which are not associated with a disease trait and those which haven't been published yet
    //by the GWAS catalog.
    //        Iterator<Study> iterator = studies.iterator();
    //        while(iterator.hasNext()){
    //            Study study = iterator.next();
    //            if(study.getDiseaseTrait() == null) {
    //                iterator.remove();
    //            }
    //            else if( study.getHousekeeping().getCatalogPublishDate() == null) {
    //                iterator.remove();
    //            }
    //            else if(study.getHousekeeping().getCatalogUnpublishDate() != null){
    //                iterator.remove();
    //            }
    //        }
    Collection<Study> filteredStudies = new ArrayList<Study>();
    Collection<Association> filteredTraitAssociations = new ArrayList<Association>();
    Collection<SingleNucleotidePolymorphism> filteredSNPs = new ArrayList<SingleNucleotidePolymorphism>();
    int count = 0;
    int studyLimit = getStudiesLimit() == -1 ? Integer.MAX_VALUE : getStudiesLimit();
    Iterator<Study> studyIterator = studies.iterator();
    while (count < studyLimit && studyIterator.hasNext()) {
        Study nextStudy = studyIterator.next();
        //only process a study if no date filter has been provided or if the study's publication date is smaller than the filter date
        if (FilterProperties.getDateFilter() == null || nextStudy.getPublicationDate().before(FilterProperties.getDateFilter())) {
            System.out.println("Qualifying study");
            for (Association nextTA : nextStudy.getAssociations()) {
                float filter = 0;
                float pval = 0;
                if (FilterProperties.getPvalueFilter() != null) {
                    filter = (float) (FilterProperties.getPvalueMant() * Math.pow(10, FilterProperties.getPvalueExp()));
                    pval = (float) (nextTA.getPvalueMantissa() * Math.pow(10, nextTA.getPvalueExponent()));
                    System.out.println("Your comparators are " + filter + " and " + pval);
                }
                if (FilterProperties.getPvalueFilter() == null || pval < filter) {
                    System.out.println("Qualifying association");
                    filteredTraitAssociations.add(nextTA);
                    for (Locus locus : nextTA.getLoci()) {
                        for (RiskAllele riskAllele : locus.getStrongestRiskAlleles()) {
                            filteredSNPs.add(riskAllele.getSnp());
                        }
                    }
                }
            }
            filteredStudies.add(nextStudy);
            count++;
        }
    }
    // convert this data, starting with SNPs (no dependencies) and working up to studies
    getLog().debug("Starting conversion to OWL...");
    getLog().debug("Converting " + filteredSNPs.size() + " filtered SNPs...");
    getConverter().addSNPsToOntology(filteredSNPs, conversion);
    getLog().debug("Converting " + filteredTraitAssociations.size() + " filtered Trait Associations...");
    getConverter().addAssociationsToOntology(filteredTraitAssociations, conversion);
    getLog().debug("Converting " + filteredStudies.size() + " filtered Studies...");
    getConverter().addStudiesToOntology(filteredStudies, conversion);
    getLog().debug("All conversion done!");
    return conversion;
}

Example 9 with RiskAllele

use of uk.ac.ebi.spot.goci.model.RiskAllele in project goci by EBISPOT.

the class AssociationService method loadAssociatedData.

public void loadAssociatedData(Association association) {
    int traitCount = association.getEfoTraits().size();
    Study study = studyService.fetchOne(association.getStudy());
    AtomicInteger reportedGeneCount = new AtomicInteger();
    Collection<SingleNucleotidePolymorphism> snps = new HashSet<>();
    Collection<Region> regions = new HashSet<>();
    Collection<Gene> mappedGenes = new HashSet<>();
    Map<String, Set<String>> mappedGeneEntrezIds = new HashMap<>();
    Map<String, Set<String>> mappedGeneEnsemblIds = new HashMap<>();
    association.getLoci().forEach(locus -> {
        locus.getStrongestRiskAlleles().stream().map(RiskAllele::getSnp).forEach(snp -> {
            Collection<Location> snpLocations = snp.getLocations();
            for (Location location : snpLocations) {
                regions.add(location.getRegion());
            }
            snp.getGenomicContexts().forEach(context -> {
                mappedGenes.add(context.getGene());
                String geneName = context.getGene().getGeneName();
                Collection<EntrezGene> geneEntrezGeneIds = context.getGene().getEntrezGeneIds();
                Collection<EnsemblGene> geneEnsemblGeneIds = context.getGene().getEnsemblGeneIds();
                if (mappedGeneEntrezIds.containsKey(geneName)) {
                    for (EntrezGene entrezGene : geneEntrezGeneIds) {
                        mappedGeneEntrezIds.get(geneName).add(entrezGene.getEntrezGeneId());
                    }
                } else {
                    Set<String> entrezIds = new HashSet<>();
                    for (EntrezGene entrezGene : geneEntrezGeneIds) {
                        entrezIds.add(entrezGene.getEntrezGeneId());
                    }
                    mappedGeneEntrezIds.put(geneName, entrezIds);
                }
                if (mappedGeneEnsemblIds.containsKey(geneName)) {
                    for (EnsemblGene ensemblGene : geneEnsemblGeneIds) {
                        mappedGeneEnsemblIds.get(geneName).add(ensemblGene.getEnsemblGeneId());
                    }
                } else {
                    Set<String> ensemblIds = new HashSet<>();
                    for (EnsemblGene ensemblGene : geneEnsemblGeneIds) {
                        ensemblIds.add(ensemblGene.getEnsemblGeneId());
                    }
                    mappedGeneEntrezIds.put(geneName, ensemblIds);
                }
            });
            snps.add(snp);
        });
        snps.addAll(locus.getStrongestRiskAlleles().stream().map(RiskAllele::getSnp).collect(Collectors.toList()));
        reportedGeneCount.addAndGet(locus.getAuthorReportedGenes().size());
        locus.getAuthorReportedGenes().forEach(authorReportedGene -> {
            authorReportedGene.getEnsemblGeneIds().size();
            authorReportedGene.getEntrezGeneIds().size();
        });
    });
    getLog().trace("Association '" + association.getId() + "' is mapped to " + "" + traitCount + " EFO traits where study id = " + study.getId() + " " + "(author reported " + reportedGeneCount + " gene(s)); " + "this reports on " + snps.size() + " SNPs in " + regions.size() + " regions, " + "mapped to " + mappedGenes.size() + " genes.");
}

Example 10 with RiskAllele

use of uk.ac.ebi.spot.goci.model.RiskAllele in project goci by EBISPOT.

the class SingleSnpMultiSnpAssociationService method createForm.

// Creates form which we can then return to view for editing etc.
@Override
public SnpAssociationForm createForm(Association association) {
    SnpAssociationStandardMultiForm form = new SnpAssociationStandardMultiForm();
    // Set association ID
    form.setAssociationId(association.getId());
    // Set simple string and float association attributes
    form.setRiskFrequency(association.getRiskFrequency());
    form.setPvalueDescription(association.getPvalueDescription());
    form.setSnpType(association.getSnpType());
    form.setMultiSnpHaplotype(association.getMultiSnpHaplotype());
    form.setSnpApproved(association.getSnpApproved());
    form.setPvalueMantissa(association.getPvalueMantissa());
    form.setPvalueExponent(association.getPvalueExponent());
    form.setStandardError(association.getStandardError());
    form.setRange(association.getRange());
    form.setDescription(association.getDescription());
    // Set OR/Beta values
    form.setOrPerCopyNum(association.getOrPerCopyNum());
    form.setOrPerCopyRecip(association.getOrPerCopyRecip());
    form.setOrPerCopyRecipRange(association.getOrPerCopyRecipRange());
    form.setBetaNum(association.getBetaNum());
    form.setBetaUnit(association.getBetaUnit());
    form.setBetaDirection(association.getBetaDirection());
    // Add collection of Efo traits
    form.setEfoTraits(association.getEfoTraits());
    // For each locus get genes and risk alleles
    Collection<Locus> loci = association.getLoci();
    Collection<Gene> locusGenes = new ArrayList<>();
    Collection<RiskAllele> locusRiskAlleles = new ArrayList<RiskAllele>();
    // For multi-snp and standard snps we assume their is only one locus
    for (Locus locus : loci) {
        locusGenes.addAll(locus.getAuthorReportedGenes());
        locusRiskAlleles.addAll(locus.getStrongestRiskAlleles().stream().sorted((v1, v2) -> Long.compare(v1.getId(), v2.getId())).collect(Collectors.toList()));
        // There should only be one locus thus should be safe to set these here
        form.setMultiSnpHaplotypeNum(locus.getHaplotypeSnpCount());
        form.setMultiSnpHaplotypeDescr(locus.getDescription());
    }
    // Get name of gene and add to form
    Collection<String> authorReportedGenes = new ArrayList<>();
    for (Gene locusGene : locusGenes) {
        authorReportedGenes.add(locusGene.getGeneName());
    }
    form.setAuthorReportedGenes(authorReportedGenes);
    // Handle snp rows
    Collection<GenomicContext> snpGenomicContexts = new ArrayList<GenomicContext>();
    Collection<SingleNucleotidePolymorphism> snps = new ArrayList<>();
    List<SnpFormRow> snpFormRows = new ArrayList<SnpFormRow>();
    List<SnpMappingForm> snpMappingForms = new ArrayList<SnpMappingForm>();
    for (RiskAllele riskAllele : locusRiskAlleles) {
        SnpFormRow snpFormRow = new SnpFormRow();
        snpFormRow.setStrongestRiskAllele(riskAllele.getRiskAlleleName());
        SingleNucleotidePolymorphism snp = riskAllele.getSnp();
        snps.add(snp);
        String rsID = snp.getRsId();
        snpFormRow.setSnp(rsID);
        Collection<Location> locations = snp.getLocations();
        for (Location location : locations) {
            SnpMappingForm snpMappingForm = new SnpMappingForm(rsID, location);
            snpMappingForms.add(snpMappingForm);
        }
        // Set proxy if one is present
        Collection<String> proxySnps = new ArrayList<>();
        if (riskAllele.getProxySnps() != null) {
            for (SingleNucleotidePolymorphism riskAlleleProxySnp : riskAllele.getProxySnps()) {
                proxySnps.add(riskAlleleProxySnp.getRsId());
            }
        }
        snpFormRow.setProxySnps(proxySnps);
        snpGenomicContexts.addAll(genomicContextRepository.findBySnpId(snp.getId()));
        snpFormRows.add(snpFormRow);
    }
    form.setSnpMappingForms(snpMappingForms);
    form.setGenomicContexts(snpGenomicContexts);
    form.setSnps(snps);
    form.setSnpFormRows(snpFormRows);
    return form;
}