All Downloads are FREE. Search and download functionalities are using the official Maven repository.

org.opencb.cellbase.lib.builders.PharmGKBBuilder Maven / Gradle / Ivy

The newest version!
/*
 * Copyright 2015-2020 OpenCB
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package org.opencb.cellbase.lib.builders;

import com.fasterxml.jackson.databind.ObjectMapper;
import com.fasterxml.jackson.databind.ObjectReader;
import org.apache.commons.collections4.CollectionUtils;
import org.apache.commons.lang3.StringUtils;
import org.opencb.biodata.models.core.Xref;
import org.opencb.biodata.models.pharma.*;
import org.opencb.biodata.models.pharma.guideline.BasicObject;
import org.opencb.cellbase.core.serializer.CellBaseFileSerializer;
import org.opencb.commons.utils.FileUtils;

import java.io.BufferedReader;
import java.io.File;
import java.io.IOException;
import java.nio.file.Path;
import java.util.*;
import java.util.stream.Collectors;

import static org.opencb.cellbase.lib.EtlCommons.*;

public class PharmGKBBuilder extends CellBaseBuilder {

    private final Path inputDir;
    private final Path pharmGKBDir;

    private static final String CHEMICALS_BASENAME = "chemicals";
    private static final String CHEMICALS_TSV_FILENAME = "chemicals.tsv";

    private static final String VARIANTS_BASENAME = "variants";
    private static final String VARIANTS_TSV_FILENAME = "variants.tsv";

    private static final String GENES_BASENAME = "genes";
    private static final String GENES_TSV_FILENAME = "genes.tsv";

    private static final String CLINICAL_ANNOTATIONS_BASENAME = "clinicalAnnotations";
    private static final String CLINICAL_ANNOTATIONS_TSV_FILENAME = "clinical_annotations.tsv";
    private static final String CLINICAL_ANN_ALLELES_TSV_FILENAME = "clinical_ann_alleles.tsv";
    private static final String CLINICAL_ANN_EVIDENCE_TSV_FILENAME = "clinical_ann_evidence.tsv";

    private static final String VARIANT_ANNOTATIONS_BASENAME = "variantAnnotations";
    private static final String VARIANT_ANNOTATIONS_TSV_FILENAME = "var_drug_ann.tsv";
    private static final String PHENOTYPE_ANNOTATIONS_TSV_FILENAME = "var_pheno_ann.tsv";
    private static final String FUNCTIONAL_ANNOTATIONS_TSV_FILENAME = "var_fa_ann.tsv";
    private static final String STUDY_PARAMETERS_TSV_FILENAME = "study_parameters.tsv";

    private static final String GUIDELINE_ANNOTATIONS_BASENAME = "guidelineAnnotations";

    private static final String DRUG_LABELS_BASENAME = "drugLabels";
    private static final String DRUG_LABELS_TSV_FILENAME = "drugLabels.tsv";

    private static final String RELATIONSHIPS_BASENAME = "relationships";
    private static final String RELATIONSHIPS_TSV_FILENAME = "relationships.tsv";

    private static final String GUIDELINE_ANNOTATION_EVIDENCE_TYPE = "Guideline Annotation";
    private static final String DRUG_LABEL_ANNOTATION_EVIDENCE_TYPE = "Label Annotation";
    private static final String VARIANT_ANNOTATION_EVIDENCE_TYPE = "Variant Drug Annotation";
    private static final String FUNCTIONAL_ANNOTATION_EVIDENCE_TYPE = "Variant Functional Assay Annotation";
    private static final String PHENOTYPE_ANNOTATION_EVIDENCE_TYPE = "Variant Phenotype Annotation";

    private static final String LOCATION_KEY = "location";
    private static final String CHROMOSOME_KEY = "chrom";
    private static final String POSITION_KEY = "pos";

    private static final String GENE_ENTITY = "Gene";
    private static final String CHEMICAL_ENTITY = "Chemical";

    private static final String PHARMGKB_ID_KEY = "PHARMGKB_ID";
    private static final String PHARMGKB_ASSOCIATION_TYPE_KEY = "PHARMGKB_ASSOCIATION_TYPE";
    private static final String PHARMGKB_LEVEL_OVERRIDE_KEY = "PHARMGKB_LEVEL_OVERRIDE";
    private static final String PHARMGKB_LEVEL_MODIFIERS_KEY = "PHARMGKB_LEVEL_MODIFIERS";
    private static final String PHARMGKB_LAST_UPDATE_DATE_KEY = "PHARMGKB_LAST_UPDATE_DATE";
    private static final String PHARMGKB_IS_VIP_KEY = "PHARMGKB_IS_VIP";

    public PharmGKBBuilder(Path inputDir, CellBaseFileSerializer serializer) {
        super(serializer);

        this.inputDir = inputDir;
        this.pharmGKBDir = inputDir.resolve(PHARMGKB_DATA);
    }

    @Override
    public void parse() throws Exception {
        // Check input folder
        FileUtils.checkDirectory(inputDir);

        // PharmGKB
        FileUtils.checkDirectory(pharmGKBDir);
        logger.info("Parsing {} files and building the data models...", PHARMGKB_NAME);

        // Parse chemical file
        Map chemicalsMap = parseChemicalFile();

        // Parse clinical annotation files
        parseClinicalAnnotationFiles(chemicalsMap);

        // Parse gene file
        parseGeneFile(chemicalsMap);

        logger.info("Parsing {} files finished.", PHARMGKB_NAME);

        // Generation the pharmacogenomics JSON file
        logger.info("Writing {} JSON file to {} ...", PHARMACOGENOMICS_DATA, serializer.getOutdir());
        int counter = 0;
        for (Map.Entry entry : chemicalsMap.entrySet()) {
            ((CellBaseFileSerializer) serializer).serialize(entry.getValue(), PHARMACOGENOMICS_DATA);
            if (++counter % 1000 == 0) {
                logger.info("\t\t {} chemicals/drugs written.", counter);
            }
        }
        serializer.close();
        logger.info("Writing {} JSON file done!", PHARMACOGENOMICS_DATA);
    }

    private Map parseChemicalFile() throws IOException {
        Path chemicalsFile = pharmGKBDir.resolve(CHEMICALS_BASENAME).resolve(CHEMICALS_TSV_FILENAME);
        Map chemicalsMap = new HashMap<>();
        try (BufferedReader br = FileUtils.newBufferedReader(chemicalsFile)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                // 0                      1     2              3            4              5    6                7      8
                // PharmGKB Accession ID  Name  Generic Names  Trade Names  Brand Mixtures Type Cross-references SMILES InChI
                // 9                10                  11                        12                       13            14
                // Dosing Guideline External Vocabulary Clinical Annotation Count Variant Annotation Count Pathway Count VIP Count
                // 15                        16                             17                           18
                // Dosing Guideline Sources  Top Clinical Annotation Level  Top FDA Label Testing Level  Top Any Drug Label Testing Level
                // 19                     20                 21                  22               23
                // Label Has Dosing Info  Has Rx Annotation  RxNorm Identifiers  ATC Identifiers  PubChem Compound Identifiers
                PharmaChemical pharmaChemical = new PharmaChemical()
                        .setId(fields[0])
                        .setSource(PHARMGKB_NAME)
                        .setName(fields[1])
                        .setSmiles(fields[7])
                        .setInChI(fields[8]);

                // Generic Names
                if (StringUtils.isNotEmpty(fields[2])) {
                    pharmaChemical.setGenericNames(stringFieldToList(fields[2]));
                }

                // Trade Names
                if (StringUtils.isNotEmpty(fields[3])) {
                    pharmaChemical.setTradeNames(stringFieldToList(fields[3]));
                }

                // Brand Mixtures
                if (StringUtils.isNotEmpty(fields[4])) {
                    pharmaChemical.setTradeMixtures(stringFieldToList(fields[4]));
                }

                // Types
                if (StringUtils.isNotEmpty(fields[5])) {
                    pharmaChemical.setTypes(Arrays.stream(fields[5].split(",")).map(String::trim).collect(Collectors.toList()));
                }

                // We need to keep the name not the ID to map by drug name in the clinical annotation method
                chemicalsMap.put(pharmaChemical.getName(), pharmaChemical);
            }
        }
        logger.info("Number of Chemical items read {}", chemicalsMap.size());

        return chemicalsMap;
    }

    /**
     * This method parses clinical_annotations.tsv, then it parses alleles and evidences to add them to the first one.
     * @param chemicalsMap
     * @throws IOException
     */
    private void parseClinicalAnnotationFiles(Map chemicalsMap) throws IOException {
        Map variantAnnotationMap = new HashMap<>();
        Map> drugToVariantAnnotationIdMap = new HashMap<>();

        Map> variantMap = parseVariantFile();

        // clinical_annotations.tsv
        try (BufferedReader br = FileUtils.newBufferedReader(pharmGKBDir.resolve(CLINICAL_ANNOTATIONS_BASENAME)
                .resolve(CLINICAL_ANNOTATIONS_TSV_FILENAME))) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);

                // Sanity check
                if (StringUtils.isEmpty(fields[0])) {
                    logger.warn("Clinical annotation ID is missing in clinical annotations line: {}", line);
                    continue;
                }

                // 0                       1                   2     3                  4               5                6
                // Clinical Annotation ID  Variant/Haplotypes  Gene  Level of Evidence  Level Override  Level Modifiers  Score
                // 7                   8           9               10       11            12                                13
                // Phenotype Category  PMID Count  Evidence Count  Drug(s)  Phenotype(s)  Latest History Date (YYYY-MM-DD)  URL
                // 14
                // Specialty Population
                PharmaVariantAnnotation pharmaVariantAnnotation = new PharmaVariantAnnotation()
                        .setConfidence(fields[3])
                        .setScore(fields[6])
                        .setUrl(fields[13])
                        .setPopulation(fields[14]);

                // Variant or haplotypes
                if (StringUtils.isNotEmpty(fields[1])) {
                    if (isHaplotype(fields[1])) {
                        // Haplotype
                        pharmaVariantAnnotation.setHaplotypes(getHaplotypeList(fields[1]));
                    } else {
                        // Variant
                        pharmaVariantAnnotation.setVariantId(fields[1]);
                    }
                }

                // Genes
                if (StringUtils.isNotEmpty(fields[2])) {
                    pharmaVariantAnnotation.setGeneNames(Arrays.asList(fields[2].split(";")));
                }

                if (StringUtils.isNotEmpty(fields[7])) {
                    pharmaVariantAnnotation.setPhenotypeTypes(Arrays.asList(fields[7].split(";")));
                }

                if (StringUtils.isNotEmpty(fields[11])) {
                    pharmaVariantAnnotation.setPhenotypes(Arrays.asList(fields[11].split(";")));
                }

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_ID_KEY, fields[0]);
                attributes.put(PHARMGKB_LEVEL_OVERRIDE_KEY, fields[4]);
                attributes.put(PHARMGKB_LEVEL_MODIFIERS_KEY, fields[5]);
                attributes.put(PHARMGKB_LAST_UPDATE_DATE_KEY, fields[12]);
                pharmaVariantAnnotation.setAttributes(attributes);

                // Add some fields from the variant map
                if (variantMap.containsKey(pharmaVariantAnnotation.getVariantId())) {
                    pharmaVariantAnnotation.setLocation((String) variantMap.get(pharmaVariantAnnotation.getVariantId()).get(LOCATION_KEY));
                    pharmaVariantAnnotation
                            .setChromosome((String) variantMap.get(pharmaVariantAnnotation.getVariantId()).get(CHROMOSOME_KEY));
                    pharmaVariantAnnotation.setPosition((int) variantMap.get(pharmaVariantAnnotation.getVariantId()).get(POSITION_KEY));
                } else {
                    logger.warn("Variant {} from clinical annotation not found in the variant map, so chromosome and position are not set",
                            pharmaVariantAnnotation.getVariantId());
                }

                // Add the annotation to the annotationMap by annotation ID
                variantAnnotationMap.put(fields[0], pharmaVariantAnnotation);

                // Process the drug names to update the drugToClinicalAnnotationId map
                // This will be used at the end of the method to update the chemical map
                if (StringUtils.isNotEmpty(fields[10])) {
                    // Drugs are separated by semicolon
                    String[] drugs = fields[10].split(";");
                    for (String drug : drugs) {
                        if (!drugToVariantAnnotationIdMap.containsKey(drug)) {
                            // Add the drug to the map
                            drugToVariantAnnotationIdMap.put(drug, new ArrayList<>());
                        }
                        // Add the clinical annotation ID to that drug
                        drugToVariantAnnotationIdMap.get(drug).add(fields[0]);
                    }
                }
            }
        }

        // Update the clinical annotation map by parsing the clinical annotation evidences
        parseClinicalAnnotationEvidenceFile(variantAnnotationMap);

        // Update the clinical annotation map by parsing the clinical annotation alleles
        parseClinicalAnnotationAlleleFile(variantAnnotationMap);

        // Update chemicals map by adding the clinical annotation
        for (Map.Entry> entry : drugToVariantAnnotationIdMap.entrySet()) {
            if (chemicalsMap.containsKey(entry.getKey())) {
                for (String variantAnnotationId : entry.getValue()) {
                    chemicalsMap.get(entry.getKey()).getVariants().add(variantAnnotationMap.get(variantAnnotationId));
                }
            } else {
                logger.warn("Drug '{}' not found in the chemicals map", entry.getKey());
            }
        }
    }

    private Map> parseVariantFile() throws IOException {
        Map> variantMap = new HashMap<>();
        // Parse the variant file (i.e., variants.tsv)
        Path varPath = pharmGKBDir.resolve(VARIANTS_BASENAME).resolve(VARIANTS_TSV_FILENAME);
        try (BufferedReader br = FileUtils.newBufferedReader(varPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                String variantName = fields[1];

                // Sanity check
                if (StringUtils.isEmpty(variantName)) {
                    logger.warn("Variant name is missing in variant line: {}", line);
                    continue;
                }

                if (variantMap.containsKey(variantName)) {
                    logger.warn("Variant name is duplicated in variant line: {}", line);
                    continue;
                }

                // 0           1             2         3             4         5                         6
                // Variant ID  Variant Name  Gene IDs  Gene Symbols  Location  Variant Annotation count  Clinical Annotation count
                // 7                                    8                           9                      10
                // Level 1/2 Clinical Annotation count  Guideline Annotation count  Label Annotation count Synonyms
                String location = fields[4];
                if (StringUtils.isEmpty(location)) {
                    logger.warn("Location is missing for Variant name {}", variantName);
                    continue;
                }
                if (!location.startsWith("NC_")) {
                    logger.warn("Unknown location {}, it has to be a RefSeq ID", location);
                    continue;
                }
                Map attrMap = new HashMap<>();
                String[] splits = location.split("[_\\.:]");
                try {
                    int chrom = Integer.parseInt(splits[1]);
                    if (chrom >= 1 && chrom <= 22) {
                        attrMap.put(CHROMOSOME_KEY, String.valueOf(chrom));
                    } else if (chrom == 23) {
                        attrMap.put(CHROMOSOME_KEY, "X");
                    } else if (chrom == 24) {
                        attrMap.put(CHROMOSOME_KEY, "Y");
                    } else if (chrom == 12920) {
                        attrMap.put(CHROMOSOME_KEY, "MT");
                    } else {
                        logger.warn("Unknown chromosome {}", chrom);
                        continue;
                    }
                } catch (NumberFormatException e) {
                    logger.warn("Error computing chromosome from location {}: {}", location, e.getMessage());
                    continue;
                }
                try {
                    int position = Integer.parseInt(splits[3]);
                    attrMap.put(POSITION_KEY, position);
                } catch (NumberFormatException e) {
                    logger.warn("Error computing chromosome position from location {}: {}", location, e.getMessage());
                    continue;
                }
                attrMap.put(LOCATION_KEY, attrMap.get(CHROMOSOME_KEY) + ":" + attrMap.get(POSITION_KEY));

                // Add it to the variant map
                variantMap.put(variantName, attrMap);
            }
        }
        logger.info("Number of variants = {}", variantMap.size());

        return variantMap;
    }

    private void parseClinicalAnnotationEvidenceFile(Map variantAnnotationMap) throws IOException {
        // For CellBase, variant annotation correponds to the PharmGKB clinical annotation
        // Processing clinical annotation evidences implies to process the variant annotation, guideline annotations,
        // drug label annotations, phenotype annotations and functional analysis annotations
        Map variantAssociationMap = new HashMap<>();
        Map guidelineAnnotationsMap = parseGuidelineAnnotationFiles();
        Map drugLabelAnnotationsMap = parseDrugLabelAnnotationFile();

        // Parse study parameters and update the variant, phenotype and functional annotations with the parsed study parameters
        parseVariantAnnotationFile(variantAssociationMap);
        parsePhenotypeAnnotationFile(variantAssociationMap);
        parseFunctionalAnnotationFile(variantAssociationMap);
        parseStudyParameterFile(variantAssociationMap);

        // Parse the clinical annotation alleles file (i.e., clinical_ann_alleles.tsv)
        Path evidencesPath = pharmGKBDir.resolve(CLINICAL_ANNOTATIONS_BASENAME).resolve(CLINICAL_ANN_EVIDENCE_TSV_FILENAME);
        try (BufferedReader br = FileUtils.newBufferedReader(evidencesPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                String clinicalAnnotationId = fields[0];

                // Sanity check
                if (StringUtils.isEmpty(clinicalAnnotationId)) {
                    logger.warn("Clinical annotation ID is missing in clinical annotation evidence line: {}", line);
                    continue;
                }

                // 0                       1            2              3             4     5        6
                // Clinical Annotation ID  Evidence ID  Evidence Type  Evidence URL  PMID  Summary  Score
                String evidenceId = fields[1];
                String evidenceType = fields[2];
                PharmaClinicalEvidence evidence = new PharmaClinicalEvidence()
                        .setType(evidenceType)
                        .setUrl(fields[3])
                        .setPubmed(fields[4])
                        .setSummary(fields[5])
                        .setScore(fields[6]);

                switch (evidenceType) {
                    case VARIANT_ANNOTATION_EVIDENCE_TYPE:
                    case PHENOTYPE_ANNOTATION_EVIDENCE_TYPE:
                    case FUNCTIONAL_ANNOTATION_EVIDENCE_TYPE: {
                        if (variantAssociationMap.containsKey(evidenceId)) {
                            evidence.getVariantAssociations().add(variantAssociationMap.get(evidenceId));
                        } else {
                            logger.warn("Evidence ID '{}' of type '{}' not found in the variant association map", evidenceId, evidenceType);
                        }
                        break;
                    }
                    case GUIDELINE_ANNOTATION_EVIDENCE_TYPE: {
                        if (guidelineAnnotationsMap.containsKey(evidenceId)) {
                            evidence.getGuidelineAnnotations().add(guidelineAnnotationsMap.get(evidenceId));
                        } else {
                            logger.warn("Evidence ID '{}' of type '{}' not found in the variant annotations map",
                                    evidenceId, evidenceType);
                        }
                        break;
                    }
                    case DRUG_LABEL_ANNOTATION_EVIDENCE_TYPE: {
                        if (drugLabelAnnotationsMap.containsKey(evidenceId)) {
                            evidence.getDrugLabelAnnotations().add(drugLabelAnnotationsMap.get(evidenceId));
                        } else {
                            logger.warn("Evidence ID '{}' of type '{}' not found in the drug label annotations map",
                                    evidenceId, evidenceType);
                        }
                        break;
                    }
                    default: {
                        logger.warn("Unknown evidence type '{}': this evidence is skipped. Valid evidence types are: {}",
                                evidenceType,
                                StringUtils.join(
                                        Arrays.asList(VARIANT_ANNOTATION_EVIDENCE_TYPE, GUIDELINE_ANNOTATION_EVIDENCE_TYPE,
                                                DRUG_LABEL_ANNOTATION_EVIDENCE_TYPE, FUNCTIONAL_ANNOTATION_EVIDENCE_TYPE,
                                                PHENOTYPE_ANNOTATION_EVIDENCE_TYPE), ","));
                        break;
                    }
                }

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_ID_KEY, fields[0]);
                evidence.setAttributes(attributes);

                // Add evidence to clinical annotation
                if (variantAnnotationMap.containsKey(clinicalAnnotationId)) {
                    variantAnnotationMap.get(clinicalAnnotationId).getEvidences().add(evidence);
                } else {
                    logger.warn("Clinical annotation ID {} from clinical annotation evidence not found in clinical annotations",
                            clinicalAnnotationId);
                }
            }
        }
    }

    private void parseClinicalAnnotationAlleleFile(Map variantAnnotationMap) throws IOException {
        // Parse the clinical annotation alleles file (i.e., clinical_ann_alleles.tsv)
        Path allelesPath = pharmGKBDir.resolve(CLINICAL_ANNOTATIONS_BASENAME).resolve(CLINICAL_ANN_ALLELES_TSV_FILENAME);
        try (BufferedReader br = FileUtils.newBufferedReader(allelesPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                // For CellBase, variant annotation is equivalent to PharmGKB clinical annotation
                String variantAnnotationId = fields[0];

                // Sanity check
                if (StringUtils.isEmpty(variantAnnotationId)) {
                    logger.warn("Clinical annotation ID is missing in clinical annotation alleles line: {}", line);
                    continue;
                }

                // 0                       1                2                3
                // Clinical Annotation ID  Genotype/Allele  Annotation Text  Allele Function
                PharmaClinicalAllele clinicalAllele = new PharmaClinicalAllele()
                        .setAllele(fields[1])
                        .setAnnotation(fields[2])
                        .setDescription(fields[3]);

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_ID_KEY, variantAnnotationId);
                clinicalAllele.setAttributes(attributes);

                // Add allele to clinical annotation
                if (variantAnnotationMap.containsKey(variantAnnotationId)) {
                    variantAnnotationMap.get(variantAnnotationId).getAlleles().add(clinicalAllele);
                } else {
                    logger.warn("Clinical annotation ID {} from clinical annotation alleles file not found in the clinical annotations map",
                            variantAnnotationId);
                }
            }
        }
    }

    private void parseVariantAnnotationFile(Map variantAssociationMap) throws IOException {
        // For CellBase, variant association corresponds to PharmGKB variant annotation
        // Parse the variant annotation file (i.e., var_drug_ann.tsv)
        Path varDrugPath = pharmGKBDir.resolve(VARIANT_ANNOTATIONS_BASENAME).resolve(VARIANT_ANNOTATIONS_TSV_FILENAME);
        int counter = 0;
        try (BufferedReader br = FileUtils.newBufferedReader(varDrugPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                String variantAnnotationId = fields[0];

                // Sanity check
                if (StringUtils.isEmpty(variantAnnotationId)) {
                    logger.warn("Variant annotation ID is missing in variant annotations line: {}", line);
                    continue;
                }

                // 0                        1                       2       3       4           5
                // Variant Annotation ID   Variant/Haplotypes      Gene    Drug(s) PMID    Phenotype Category
                // 6           7       8               9           10
                // Significance    Notes   Sentence        Alleles Specialty Population
                PharmaVariantAssociation variantAssociation = new PharmaVariantAssociation()
                        .setGeneName(fields[2])
                        .setPubmed(fields[4])
                        .setPhenotypeType(fields[5])
                        .setSignificance(fields[6])
                        .setDiscussion(fields[7])
                        .setDescription(fields[8])
                        .setAlleles(fields[9])
                        .setPopulation(fields[10]);

                // Variant or haplotypes
                if (StringUtils.isNotEmpty(fields[1])) {
                    if (isHaplotype(fields[1])) {
                        // Haplotype
                        variantAssociation.setHaplotypes(getHaplotypeList(fields[1]));
                    } else {
                        // Variant
                        variantAssociation.setVariantId(fields[1]);
                    }
                }

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_ID_KEY, fields[0]);
                attributes.put(PHARMGKB_ASSOCIATION_TYPE_KEY, VARIANT_ANNOTATION_EVIDENCE_TYPE);
                variantAssociation.setAttributes(attributes);

                if (StringUtils.isNotEmpty(fields[3])) {
                    variantAssociation.setDrugs(stringFieldToList(fields[3]));
                }

                // Add the annotation to the variantAnnotationMap by variant and gene
                variantAssociationMap.put(variantAnnotationId, variantAssociation);
                counter++;
            }
        }
        logger.info("Number of variant annotations = {}", counter);
    }

    private Map parseGuidelineAnnotationFiles() throws IOException {
        Map guidelineAnnotationMap = new HashMap<>();

        ObjectMapper mapper = new ObjectMapper();
        ObjectReader objectReader = mapper.readerFor(PharmaGuidelineAnnotation.class);

        // Parse the guideline annotations JSON files
        Path guidelinesPath = pharmGKBDir.resolve(GUIDELINE_ANNOTATIONS_BASENAME);
        FileUtils.checkDirectory(guidelinesPath);
        for (File file : Objects.requireNonNull(guidelinesPath.toFile().listFiles())) {
            if (file.getName().endsWith("json")) {
                PharmaGuidelineAnnotation guidelineAnnotation = objectReader.readValue(file);
                if (guidelineAnnotation.getGuideline() != null
                        && StringUtils.isEmpty(guidelineAnnotation.getGuideline().getId())) {
                    logger.warn("Guideline ID is missing for guideline filename: {}", file.getName());
                    continue;
                }
                // Add the guideline annotation to the map by guideline ID (= Evidence ID)
                guidelineAnnotationMap.put(guidelineAnnotation.getGuideline().getId(), guidelineAnnotation);
            }
        }
        logger.info("Number of guideline annotations = {}", guidelineAnnotationMap.size());

        return guidelineAnnotationMap;
    }

    private Map parseDrugLabelAnnotationFile() throws IOException {
        Map drugLabelAnnotationMap = new HashMap<>();
        // Parse the drug labels annotations file (i.e., drugLabels.tsv)
        Path drugLabelPath = pharmGKBDir.resolve(DRUG_LABELS_BASENAME).resolve(DRUG_LABELS_TSV_FILENAME);
        try (BufferedReader br = FileUtils.newBufferedReader(drugLabelPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                String drugLabelId = fields[0];

                // Sanity check
                if (StringUtils.isEmpty(drugLabelId)) {
                    logger.warn("PharmGKB ID is missing in drug label line: {}", line);
                    continue;
                }

                // 0            1     2       3               4              5                     6                7
                // PharmGKB ID  Name  Source  Biomarker Flag  Testing Level  Has Prescribing Info  Has Dosing Info  Has Alternate Drug
                // 8              9            10         11     12                   13
                // Cancer Genome  Prescribing  Chemicals  Genes  Variants/Haplotypes  Latest History Date (YYYY-MM-DD)
                PharmaDrugLabelAnnotation labelAnnotation = new PharmaDrugLabelAnnotation()
                        .setName(fields[1])
                        .setSource(fields[2])
                        .setBiomarkerFlag(fields[3])
                        .setTestingLevel(fields[4])
                        .setPrescribingInfo(fields[5])
                        .setDosingInfo(fields[6])
                        .setAlternateDrug(fields[7])
                        .setCancerGenome(fields[8]);

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_ID_KEY, drugLabelId);
                labelAnnotation.setAttributes(attributes);

                // Add the drug label annotation to the map by ParhmGKB (= Evidence ID)
                drugLabelAnnotationMap.put(drugLabelId, labelAnnotation);
            }
        }
        logger.info("Number of drug label annotations = {}", drugLabelAnnotationMap.size());

        return drugLabelAnnotationMap;
    }

    private void parsePhenotypeAnnotationFile(Map variantAssociationMap) throws IOException {
        // Parse the variant annotation file (i.e., var_pheno_ann.tsv)
        Path varDrugPath = pharmGKBDir.resolve(VARIANT_ANNOTATIONS_BASENAME).resolve(PHENOTYPE_ANNOTATIONS_TSV_FILENAME);
        int counter = 0;
        try (BufferedReader br = FileUtils.newBufferedReader(varDrugPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                String variantAnnotationId = fields[0];

                // Sanity check
                if (StringUtils.isEmpty(variantAnnotationId)) {
                    logger.warn("Variant annotation ID is missing in phenotype annotations line: {}", line);
                    continue;
                }

                // 0                      1                   2     3        4     5                   6             7      8
                // Variant Annotation ID  Variant/Haplotypes  Gene  Drug(s)  PMID  Phenotype Category  Significance  Notes  Sentence
                // 9        10                   .....
                // Alleles  Specialty Population .....
                PharmaVariantAssociation variantAssociation = new PharmaVariantAssociation()
                        .setGeneName(fields[2])
                        .setPubmed(fields[4])
                        .setPhenotypeType(fields[5])
                        .setSignificance(fields[6])
                        .setDiscussion(fields[7])
                        .setDescription(fields[8])
                        .setAlleles(fields[9])
                        .setPopulation(fields[10]);

                // Variant or haplotypes
                if (StringUtils.isNotEmpty(fields[1])) {
                    if (isHaplotype(fields[1])) {
                        // Haplotype
                        variantAssociation.setHaplotypes(getHaplotypeList(fields[1]));
                    } else {
                        // Variant
                        variantAssociation.setVariantId(fields[1]);
                    }
                }

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_ID_KEY, variantAnnotationId);
                attributes.put(PHARMGKB_ASSOCIATION_TYPE_KEY, PHENOTYPE_ANNOTATION_EVIDENCE_TYPE);
                variantAssociation.setAttributes(attributes);

                if (StringUtils.isNotEmpty(fields[3])) {
                    variantAssociation.setDrugs(stringFieldToList(fields[3]));
                }

                // Add the annotation to the variantAnnotationMap by variant and gene
                variantAssociationMap.put(variantAnnotationId, variantAssociation);
                counter++;
            }
        }
        logger.info("Number of phenotype annotations = {}", counter);
    }

    private void parseFunctionalAnnotationFile(Map variantAssociationMap) throws IOException {
        // Parse the variant annotation file (i.e., var_fa_ann.tsv)
        Path varDrugPath = pharmGKBDir.resolve(VARIANT_ANNOTATIONS_BASENAME).resolve(FUNCTIONAL_ANNOTATIONS_TSV_FILENAME);
        int counter = 0;
        try (BufferedReader br = FileUtils.newBufferedReader(varDrugPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                String variantAnnotationId = fields[0];

                // Sanity check
                if (StringUtils.isEmpty(variantAnnotationId)) {
                    logger.warn("Variant annotation ID is missing in variant annotations line: {}", line);
                    continue;
                }

                // 0                        1                       2       3       4           5
                // Variant Annotation ID   Variant/Haplotypes      Gene    Drug(s) PMID    Phenotype Category
                // 6           7       8               9           10                    11          .....
                // Significance    Notes   Sentence        Alleles Specialty Population  Assay type  .....
                PharmaVariantAssociation variantAssociation = new PharmaVariantAssociation()
                        .setGeneName(fields[2])
                        .setPubmed(fields[4])
                        .setPhenotypeType(fields[5])
                        .setSignificance(fields[6])
                        .setDiscussion(fields[7])
                        .setDescription(fields[8])
                        .setAlleles(fields[9])
                        .setPopulation(fields[10])
                        .setAssayType(fields[11]);

                // Variant or haplotypes
                if (StringUtils.isNotEmpty(fields[1])) {
                    if (isHaplotype(fields[1])) {
                        // Haplotype
                        variantAssociation.setHaplotypes(getHaplotypeList(fields[1]));
                    } else {
                        // Variant
                        variantAssociation.setVariantId(fields[1]);
                    }
                }

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_ID_KEY, variantAnnotationId);
                attributes.put(PHARMGKB_ASSOCIATION_TYPE_KEY, FUNCTIONAL_ANNOTATION_EVIDENCE_TYPE);
                variantAssociation.setAttributes(attributes);

                if (StringUtils.isNotEmpty(fields[3])) {
                    variantAssociation.setDrugs(stringFieldToList(fields[3]));
                }

                // Add the annotation to the variantAnnotationMap by variant and gene
                variantAssociationMap.put(variantAnnotationId, variantAssociation);
                counter++;
            }
        }
        logger.info("Number of variant annotations = {}", counter);
    }

    private void parseStudyParameterFile(Map variantAssociationMap) throws IOException {
        Map> studyParametersMap = new HashMap<>();
        // Parse the study parameters file (i.e., study_parameters.tsv)
        Path studyParamsPath = pharmGKBDir.resolve(VARIANT_ANNOTATIONS_BASENAME).resolve(STUDY_PARAMETERS_TSV_FILENAME);
        try (BufferedReader br = FileUtils.newBufferedReader(studyParamsPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                String variantAnnotationId = fields[1];

                // Sanity check
                if (StringUtils.isEmpty(variantAnnotationId)) {
                    logger.warn("Variant annotation ID is missing in study parameters line: {}", line);
                    continue;
                }

                // 0                    1                      2           3            4               5
                // Study Parameters ID  Variant Annotation ID  Study Type  Study Cases  Study Controls  Characteristics
                // 6                     7                   8                             9
                // Characteristics Type  Frequency In Cases  Allele Of Frequency In Cases  Frequency In Controls
                // 10                               11       12               13          14                        15
                // Allele Of Frequency In Controls  P Value  Ratio Stat Type  Ratio Stat  Confidence Interval Start Confidence Interval Stop
                // 16
                // Biogeographical Groups
                PharmaStudyParameters studyParams = new PharmaStudyParameters()
                        .setStudyType(fields[2])
                        .setStudyCases(fields[3])
                        .setStudyControls(fields[4])
                        .setCharacteristics(fields[5])
                        .setCharacteristicsType(fields[6])
                        .setFrequencyInCases(fields[7])
                        .setAlleleOfFrequencyInCases(fields[8])
                        .setFrequencyInControls(fields[9])
                        .setAlleleOfFrequencyInControls(fields[10])
                        .setpValue(fields[11])
                        .setRatioStatType(fields[12])
                        .setRatioStat(fields[13])
                        .setConfidenceIntervalStart(fields[14])
                        .setConfidenceIntervalStop(fields[15])
                        .setBiogeographicalGroups(fields[16]);

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_ID_KEY, variantAnnotationId);
                studyParams.setAttributes(attributes);

                // Add the study parameters map
                if (!studyParametersMap.containsKey(variantAnnotationId)) {
                    studyParametersMap.put(variantAnnotationId, new ArrayList<>());
                }
                studyParametersMap.get(variantAnnotationId).add(studyParams);
            }
        }
        logger.info("Number of study parameters lines = {}", studyParametersMap.size());

        for (Map.Entry> entry : studyParametersMap.entrySet()) {
            if (variantAssociationMap.containsKey(entry.getKey())) {
                variantAssociationMap.get(entry.getKey()).setStudyParameters(entry.getValue());
            } else {
                logger.warn("Study parameters with variant annotation ID {} not found in variant association map", entry.getKey());
            }
        }
    }

    private void parseGeneFile(Map chemicalsMap) throws IOException {
        // To relate genes with chemicals we will take the relationships from:
        //    1. From guidelines (from the members 'relatedGenes' and 'relatedChemicals')
        //    2. From the file relationships.tsv (from the relationship Gene - Chemical)

        // Create the PharmGKB gene ID map by chemical name
        Map> pgkbGeneIdMapByChemicalName = new HashMap<>();

        // Create and populate guideline annotations map by PharmGKB gene ID
        List guidelineAnnotations = new ArrayList<>(parseGuidelineAnnotationFiles().values());
        Map> guidelineAnnotationMapByPgkbGeneId = new HashMap<>();
        for (PharmaGuidelineAnnotation guidelineAnnotation : guidelineAnnotations) {
            if (guidelineAnnotation.getGuideline() != null
                    && CollectionUtils.isNotEmpty(guidelineAnnotation.getGuideline().getRelatedGenes())) {
                for (BasicObject relatedGene : guidelineAnnotation.getGuideline().getRelatedGenes()) {
                    if (StringUtils.isNotEmpty(relatedGene.getId())) {
                        String pgkbGeneId = relatedGene.getId();
                        if (StringUtils.isNotEmpty(pgkbGeneId)) {
                            // Populate the guideline annotation map by PharmGKB gene ID
                            if (!guidelineAnnotationMapByPgkbGeneId.containsKey(pgkbGeneId)) {
                                guidelineAnnotationMapByPgkbGeneId.put(pgkbGeneId, new ArrayList<>());
                            }
                            guidelineAnnotationMapByPgkbGeneId.get(pgkbGeneId).add(guidelineAnnotation);

                            // Populate the PharmGKB gene ID map by chemical names
                            if (CollectionUtils.isNotEmpty(guidelineAnnotation.getGuideline().getRelatedChemicals())) {
                                for (BasicObject relatedChemical : guidelineAnnotation.getGuideline().getRelatedChemicals()) {
                                    String chemicalName = relatedChemical.getName();
                                    if (StringUtils.isNotEmpty(chemicalName)) {
                                        if (!pgkbGeneIdMapByChemicalName.containsKey(chemicalName)) {
                                            pgkbGeneIdMapByChemicalName.put(chemicalName, new HashSet<>());
                                        }
                                        pgkbGeneIdMapByChemicalName.get(chemicalName).add(pgkbGeneId);
                                    }
                                }
                            }
                        }
                    }
                }
            }
        }

        // Parse the genes file (i.e., genes.tsv)
        Map geneAnnotationMapByPgkbGeneId = new HashMap<>();
        Path genesPath = pharmGKBDir.resolve(GENES_BASENAME).resolve(GENES_TSV_FILENAME);
        try (BufferedReader br = FileUtils.newBufferedReader(genesPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);
                String pgkbGeneId = fields[0];

                // Sanity check
                if (StringUtils.isEmpty(pgkbGeneId)) {
                    logger.warn("PharmGKB accession ID is missing in genes file line: {}", line);
                    continue;
                }
                // 0                      1             2        3           4     5       6                7                  8
                // PharmGKB Accession Id  NCBI Gene ID  HGNC ID  Ensembl Id  Name  Symbol  Alternate Names  Alternate Symbols  Is VIP
                // 9                       10                11                         12          13
                // Has Variant Annotation  Cross-references  Has CPIC Dosing Guideline  Chromosome  Chromosomal Start - GRCh37
                // 14                         15                          16
                // Chromosomal Stop - GRCh37  Chromosomal Start - GRCh38  Chromosomal Stop - GRCh38
                PharmaGeneAnnotation geneAnnotation = new PharmaGeneAnnotation()
                        .setId(pgkbGeneId)
                        .setName(fields[4]);

                List xrefs = new ArrayList<>();
                if (StringUtils.isNotEmpty(fields[1])) {
                    xrefs.add(new Xref(fields[1], "NCBI", "NCBI gene ID"));
                }
                if (StringUtils.isNotEmpty(fields[2])) {
                    xrefs.add(new Xref(fields[2], "HGNC", "HGNC gene ID"));
                }
                if (StringUtils.isNotEmpty(fields[3])) {
                    xrefs.add(new Xref(fields[3], "Ensembl", "Ensembl gene ID"));
                }
                if (StringUtils.isNotEmpty(fields[5])) {
                    xrefs.add(new Xref(fields[5], "HGNC", "HGNC gene symbol"));
                }
                if (CollectionUtils.isNotEmpty(xrefs)) {
                    geneAnnotation.setXrefs(xrefs);
                }

                if (StringUtils.isNotEmpty(fields[9])) {
                    geneAnnotation.setHasVariantAnnotation(fields[9].toLowerCase(Locale.ROOT).equals("yes"));
                }

                // Set guidelines by getting them from the guideline annotations map
                if (guidelineAnnotationMapByPgkbGeneId.containsKey(pgkbGeneId)) {
                    geneAnnotation.setGuidelineAnnotations(guidelineAnnotationMapByPgkbGeneId.get(pgkbGeneId));
                }

                Map attributes = new HashMap<>();
                attributes.put(PHARMGKB_IS_VIP_KEY, fields[8]);
                geneAnnotation.setAttributes(attributes);

                // Add to the map
                if (geneAnnotationMapByPgkbGeneId.containsKey(pgkbGeneId)) {
                    logger.warn("PharmGKB gene ID {} is duplicated in the PharmGKB file {}", pgkbGeneId, GENES_TSV_FILENAME);
                } else {
                    geneAnnotationMapByPgkbGeneId.put(pgkbGeneId, geneAnnotation);
                }
            }
        }

        // Parse the chemical-gene relationships and update the PharmGKB gene ID map byh chemical name
        // In addtion, updata the gene annotation map with additional fields (e.g., evidences, pubmeds...)
        parseChemicalGeneRelationships(pgkbGeneIdMapByChemicalName, geneAnnotationMapByPgkbGeneId);

        // Finally, update the chemical map with the gene annotation
        for (Map.Entry entry : chemicalsMap.entrySet()) {
            String chemicalName = entry.getKey();
            if (pgkbGeneIdMapByChemicalName.containsKey(chemicalName)) {
                for (String pgkbGeneId : pgkbGeneIdMapByChemicalName.get(chemicalName)) {
                    if (geneAnnotationMapByPgkbGeneId.containsKey(pgkbGeneId)) {
                        entry.getValue().getGenes().add(geneAnnotationMapByPgkbGeneId.get(pgkbGeneId));
                    }
                }
            }
        }

        logger.info("Number of parsed genes = {}", geneAnnotationMapByPgkbGeneId.size());
    }

    private void parseChemicalGeneRelationships(Map> pgkbGeneIdMapByChemicalName,
                                                Map geneAnnotationMapByPgkbGeneId) throws IOException {
        int counter = 0;
        // Parse the genes file (i.e., relationships.tsv)
        Path relationshipsPath = pharmGKBDir.resolve(RELATIONSHIPS_BASENAME).resolve(RELATIONSHIPS_TSV_FILENAME);
        try (BufferedReader br = FileUtils.newBufferedReader(relationshipsPath)) {
            // Skip first line, i.e. the header line
            String line = br.readLine();
            while ((line = br.readLine()) != null) {
                String[] fields = line.split("\t", -1);

                // 0           1             2             3           4             5             6         7            8   0   10
                // Entity1_id  Entity1_name  Entity1_type  Entity2_id  Entity2_name  Entity2_type  Evidence  Association  PK  PD  PMIDs
                String pgkbGeneId = fields[0];
                String entity1Type = fields[2];
                String chemicalName = fields[4];
                String entity2Type = fields[5];
                if (StringUtils.isNotEmpty(pgkbGeneId) && StringUtils.isNotEmpty(entity1Type) && StringUtils.isNotEmpty(chemicalName)
                        && StringUtils.isNotEmpty(entity2Type) && entity1Type.equals(GENE_ENTITY) && entity2Type.equals(CHEMICAL_ENTITY)) {
                    if (!pgkbGeneIdMapByChemicalName.containsKey(chemicalName)) {
                        pgkbGeneIdMapByChemicalName.put(chemicalName, new HashSet<>());
                    }
                    pgkbGeneIdMapByChemicalName.get(chemicalName).add(pgkbGeneId);

                    // Update gene annotation map
                    if (geneAnnotationMapByPgkbGeneId.containsKey(pgkbGeneId)) {
                        if (StringUtils.isNotEmpty(fields[6])) {
                            geneAnnotationMapByPgkbGeneId.get(pgkbGeneId).setEvidences(Arrays.asList(fields[6].split(",", -1)));
                        }
                        if (StringUtils.isNotEmpty(fields[7])) {
                            geneAnnotationMapByPgkbGeneId.get(pgkbGeneId).setConfidence(fields[7]);
                        }
                        if (StringUtils.isNotEmpty(fields[10])) {
                            geneAnnotationMapByPgkbGeneId.get(pgkbGeneId).setPubmed(Arrays.asList(fields[10].split(";", -1)));
                        }
                    } else {
                        logger.warn("PhamGKB gene ID {} found in the file {} but not in the file {}", pgkbGeneId,
                                RELATIONSHIPS_TSV_FILENAME, GENES_TSV_FILENAME);
                    }
                    counter++;
                }
            }
        }
        logger.info("Number of parsed {}-{} relationships = {}", GENE_ENTITY, CHEMICAL_ENTITY, counter);
    }

    private List stringFieldToList(String field) {
        if (field.startsWith("\"")) {
            return Arrays
                    .stream(field.replace("\"\"\"", "\"").replace("\"\"", "\"").replace("\", \"", "\",\"").split("\",\""))
                    .map(s -> s.replace("\"", "").trim())
                    .collect(Collectors.toList());
        } else {
            if (field.contains(", ")) {
                return Arrays
                        .stream(field.replace(", ", ",").split(","))
                        .map(String::trim)
                        .collect(Collectors.toList());
            } else {
                return Collections.singletonList(field);
            }
        }
    }

    private boolean isHaplotype(String value) {
        return (!value.startsWith("rs") && value.contains("*"));
    }

    private List getHaplotypeList(String value) {
        return Arrays.stream(value.split(",")).map(s -> s.trim()).collect(Collectors.toList());
    }
}




© 2015 - 2024 Weber Informatics LLC | Privacy Policy