Examples with Program nl.hartwigmedicalfoundation.bachelor.Program used on opensource projects

Example 1 with Program

use of nl.hartwigmedicalfoundation.bachelor.Program in project hmftools by hartwigmedical.

the class BachelorEligibility method processCopyNumbers.

@NotNull
Collection<EligibilityReport> processCopyNumbers(final String patient, final List<GeneCopyNumber> copyNumbers) {
    final List<EligibilityReport> results = Lists.newArrayList();
    for (final GeneCopyNumber copyNumber : copyNumbers) {
        // TODO: verify the germline check
        final boolean isGermline = copyNumber.germlineHet2HomRegions() + copyNumber.germlineHomRegions() > 0;
        final List<String> matchingPrograms = programs.stream().filter(program -> program.copyNumberProcessor().test(copyNumber)).map(BachelorProgram::name).collect(Collectors.toList());
        final List<EligibilityReport> interimResults = matchingPrograms.stream().map(p -> ImmutableEligibilityReport.builder().patient(patient).source(isGermline ? GERMLINE_DELETION : SOMATIC_DELETION).program(p).id("").genes(copyNumber.gene()).chrom(copyNumber.chromosome()).pos(copyNumber.start()).ref("").alts("").effects("").build()).collect(Collectors.toList());
        results.addAll(interimResults);
    }
    return results;
}

Example 2 with Program

use of nl.hartwigmedicalfoundation.bachelor.Program in project hmftools by hartwigmedical.

the class BachelorEligibility method fromMap.

static BachelorEligibility fromMap(final Map<String, Program> input) {
    final BachelorEligibility result = new BachelorEligibility();
    for (final Program program : input.values()) {
        final Multimap<String, String> geneToEnsemblMap = HashMultimap.create();
        program.getPanel().stream().map(ProgramPanel::getGene).flatMap(Collection::stream).forEach(g -> geneToEnsemblMap.put(g.getName(), g.getEnsembl()));
        // NOTE: copy number and SVs are untested/unverified for now, but leave in support for them
        // process copy number sections
        final List<Predicate<GeneCopyNumber>> cnvPredicates = Lists.newArrayList();
        for (final ProgramPanel panel : program.getPanel()) {
            final List<GeneIdentifier> genes = panel.getGene();
            if (panel.getEffect().contains(OtherEffect.HOMOZYGOUS_DELETION)) {
                final Predicate<GeneCopyNumber> geneCopyNumberPredicate = cnv -> genes.stream().anyMatch(g -> g.getEnsembl().equals(cnv.transcriptID()));
                // TODO: we are matching on transcript ID here but we only have canonical transcripts in our panel file
                cnvPredicates.add(geneCopyNumberPredicate);
            }
        }
        // process structural variant disruptions
        final List<Predicate<HmfGenomeRegion>> disruptionPredicates = Lists.newArrayList();
        for (final ProgramPanel panel : program.getPanel()) {
            final List<GeneIdentifier> genes = panel.getGene();
            if (panel.getEffect().contains(OtherEffect.GENE_DISRUPTION)) {
                final Predicate<HmfGenomeRegion> disruptionPredicate = sv -> genes.stream().anyMatch(g -> g.getEnsembl().equals(sv.transcriptID()));
                // TODO: we are matching on transcript ID here but we only have canonical transcripts in our panel file
                disruptionPredicates.add(disruptionPredicate);
            }
        }
        // process variants from vcf
        final List<Predicate<VariantModel>> panelPredicates = Lists.newArrayList();
        List<String> requiredEffects = Lists.newArrayList();
        List<String> panelTranscripts = Lists.newArrayList();
        for (final ProgramPanel panel : program.getPanel()) {
            final List<GeneIdentifier> genes = panel.getGene();
            // take up a collection of the effects to search for
            requiredEffects = panel.getSnpEffect().stream().map(SnpEffect::value).collect(Collectors.toList());
            panelTranscripts = genes.stream().map(GeneIdentifier::getEnsembl).collect(Collectors.toList());
            final List<String> effects = requiredEffects;
            final Predicate<VariantModel> panelPredicate = v -> genes.stream().anyMatch(p -> v.sampleAnnotations().stream().anyMatch(a -> a.featureID().equals(p.getEnsembl()) && effects.stream().anyMatch(x -> a.effects().contains(x))));
            panelPredicates.add(panelPredicate);
            // update query targets
            for (final GeneIdentifier g : genes) {
                final HmfGenomeRegion region = allTranscriptsMap.get(g.getEnsembl());
                if (region == null) {
                    final HmfGenomeRegion namedRegion = allGenesMap.get(g.getName());
                    if (namedRegion == null) {
                        LOGGER.warn("Program {} gene {} non-canonical transcript {} couldn't find region, transcript will be skipped", program.getName(), g.getName(), g.getEnsembl());
                    // just skip this gene for now
                    } else {
                        result.variantLocationsToQuery.add(namedRegion);
                    }
                } else {
                    result.variantLocationsToQuery.add(region);
                }
            }
        }
        final Predicate<VariantModel> inPanel = v -> panelPredicates.stream().anyMatch(p -> p.test(v));
        final Predicate<VariantModel> inBlacklist = new BlacklistPredicate(geneToEnsemblMap.values(), program.getBlacklist());
        final Predicate<VariantModel> inWhitelist = new WhitelistPredicate(geneToEnsemblMap, program.getWhitelist());
        final Predicate<VariantModel> snvPredicate = v -> inPanel.test(v) ? !inBlacklist.test(v) : inWhitelist.test(v);
        final Predicate<GeneCopyNumber> copyNumberPredicate = cnv -> cnvPredicates.stream().anyMatch(p -> p.test(cnv)) && cnv.minCopyNumber() < MAX_COPY_NUMBER_FOR_LOSS;
        final Predicate<HmfGenomeRegion> disruptionPredicate = disruption -> disruptionPredicates.stream().anyMatch(p -> p.test(disruption));
        BachelorProgram bachelorProgram = new BachelorProgram(program.getName(), snvPredicate, copyNumberPredicate, disruptionPredicate, requiredEffects, panelTranscripts);
        result.programs.add(bachelorProgram);
    }
    return result;
}

Example 3 with Program

use of nl.hartwigmedicalfoundation.bachelor.Program in project hmftools by hartwigmedical.

the class BachelorHelper method loadXML.

@NotNull
public static Map<String, Program> loadXML(final Path path) throws IOException, SAXException {
    final BachelorSchema schema = BachelorSchema.make();
    final List<Program> programs = Files.walk(path).filter(p -> p.toString().endsWith(".xml")).map(schema::processXML).filter(Objects::nonNull).collect(Collectors.toList());
    final Map<String, Program> result = Maps.newHashMap();
    for (final Program p : programs) {
        if (result.containsKey(p.getName())) {
            LOGGER.error("duplicate programs detected: {}", p.getName());
            System.exit(1);
        } else {
            result.put(p.getName(), p);
        }
    }
    return result;
}

Example 4 with Program

use of nl.hartwigmedicalfoundation.bachelor.Program in project hmftools by hartwigmedical.

the class BachelorApplication method main.

public static void main(final String... args) {
    final Options options = createOptions();
    try {
        final CommandLine cmd = createCommandLine(options, args);
        // load configs
        final Map<String, Program> map;
        if (cmd.hasOption(CONFIG_DIRECTORY)) {
            map = BachelorHelper.loadXML(Paths.get(cmd.getOptionValue(CONFIG_DIRECTORY)));
        } else if (cmd.hasOption(CONFIG_XML)) {
            map = BachelorHelper.loadXML(Paths.get(cmd.getOptionValue(CONFIG_XML)));
        } else {
            LOGGER.error("config directory or xml required!");
            System.exit(1);
            return;
        }
        if (cmd.hasOption(VALIDATE)) {
            System.exit(0);
            return;
        }
        if (map.isEmpty()) {
            LOGGER.error("no programs loaded, exiting");
            System.exit(1);
            return;
        }
        final BachelorEligibility eligibility = BachelorEligibility.fromMap(map);
        LOGGER.info("beginning processing");
        final boolean germline = cmd.hasOption(GERMLINE);
        final boolean somatic = cmd.hasOption(SOMATIC);
        final boolean copyNumber = cmd.hasOption(COPYNUMBER);
        final boolean structuralVariants = cmd.hasOption(SV);
        final boolean doAll = !(germline || somatic || copyNumber || structuralVariants);
        final List<File> filesToMerge;
        if (cmd.hasOption(BATCH_DIRECTORY)) {
            final Path root = Paths.get(cmd.getOptionValue(BATCH_DIRECTORY));
            try (final Stream<Path> stream = Files.walk(root, 1, FileVisitOption.FOLLOW_LINKS).parallel()) {
                filesToMerge = stream.filter(p -> p.toFile().isDirectory()).filter(p -> !p.equals(root)).map(RunDirectory::new).map(run -> process(eligibility, run, germline || doAll, somatic || doAll, copyNumber || doAll, structuralVariants || doAll)).collect(Collectors.toList());
            }
        } else if (cmd.hasOption(RUN_DIRECTORY)) {
            final Path path = Paths.get(cmd.getOptionValue(RUN_DIRECTORY));
            if (!Files.exists(path)) {
                LOGGER.error("-runDirectory path does not exist");
                System.exit(1);
                return;
            }
            filesToMerge = Collections.singletonList(process(eligibility, new RunDirectory(path), germline || doAll, somatic || doAll, copyNumber || doAll, structuralVariants || doAll));
        } else {
            LOGGER.error("requires either a batch or single run directory");
            System.exit(1);
            return;
        }
        LOGGER.info("processing complete");
        LOGGER.info("merging to CSV {}", cmd.getOptionValue(OUTPUT));
        try (final BufferedWriter writer = Files.newBufferedWriter(Paths.get(cmd.getOptionValue(OUTPUT)))) {
            // header
            writer.write(fileHeader());
            writer.newLine();
            for (final File file : filesToMerge) {
                final List<String> lines = Files.readAllLines(file.toPath());
                for (final String line : lines) {
                    writer.write(line);
                    writer.newLine();
                }
            }
        }
        LOGGER.info("output written");
        LOGGER.info("bachelor done");
    } catch (final ParseException e) {
        printHelpAndExit(options);
    } catch (Exception e) {
        e.printStackTrace();
    }
}

Example 5 with Program

use of nl.hartwigmedicalfoundation.bachelor.Program in project hmftools by hartwigmedical.

the class BachelorEligibility method processStructuralVariant.

private Collection<EligibilityReport> processStructuralVariant(final String patient, final GenomePosition position, final GenomePosition other, final StructuralVariantType svType) {
    final List<EligibilityReport> results = Lists.newArrayList();
    // TODO: can we do better than this performance wise? new map?
    for (final HmfGenomeRegion region : allGenesByChromosomeMap.get(position.chromosome())) {
        if (!region.contains(position)) {
            continue;
        }
        // skip non-inversion intronic variants
        if (region.contains(other) && svType != StructuralVariantType.INV) {
            final int intronStart = intron(region.exome(), position);
            final int intronEnd = intron(region.exome(), other);
            // the variant is intronic in a gene -- we will filter it
            if (intronStart >= 0 && intronStart == intronEnd) {
                continue;
            }
        }
        programs.stream().filter(p -> p.disruptionProcessor().test(region)).map(p -> ImmutableEligibilityReport.builder().patient(patient).source(SOMATIC_DISRUPTION).program(p.name()).id("").genes(region.gene()).chrom(region.chromosome()).pos(position.position()).ref("").alts("").effects("").build()).forEach(results::add);
    }
    return results;
}