Examples with GoTree com.github.lindenb.jvarkit.util.go.GoTree used on opensource projects

Example 1 with GoTree

use of com.github.lindenb.jvarkit.util.go.GoTree in project jvarkit by lindenb.

the class VcfBurdenGoEnrichment method doWork.

@Override
public int doWork(final List<String> args) {
    if (StringUtil.isBlank(this.readingGo.goUri)) {
        LOG.error("Undefined GOs uri.");
        return -1;
    }
    if (this.geneFile == null || !this.geneFile.exists()) {
        LOG.error("Undefined gene file option.");
        return -1;
    }
    try {
        final GoTree gotree = this.readingGo.createParser().setIgnoreDbXRef(true).parse(this.readingGo.goUri);
        List<GoTree.Term> terms = new ArrayList<>(gotree.getTerms());
        final Map<GoTree.Term, Node> term2node = new HashMap<>();
        // build the node TREE
        while (!terms.isEmpty()) {
            int i = 0;
            while (i < terms.size()) {
                final GoTree.Term t = terms.get(i);
                if (!t.hasRelations()) {
                    term2node.put(t, new Node(t));
                    terms.remove(i);
                } else if (t.getRelations().stream().allMatch(L -> term2node.containsKey(L.getTo()))) {
                    final Node n = new Node(t);
                    n.parents.addAll(t.getRelations().stream().map(L -> term2node.get(L.getTo())).collect(Collectors.toSet()));
                    term2node.put(t, n);
                    terms.remove(i);
                } else {
                    i++;
                }
            }
        }
        terms = null;
        final Set<String> unknownAcn = new HashSet<>();
        final Map<String, Set<Node>> gene2node = new HashMap<>();
        final BufferedReader r = IOUtils.openFileForBufferedReading(this.geneFile);
        String line;
        while ((line = r.readLine()) != null) {
            if (line.isEmpty() || line.startsWith("#"))
                continue;
            final int t = line.indexOf('\t');
            if (t == -1) {
                r.close();
                LOG.error("tab missing in " + line + " of " + this.geneFile);
                return -1;
            }
            final String gene = line.substring(0, t).trim();
            if (StringUtil.isBlank(gene)) {
                r.close();
                LOG.error("Emtpy gene in " + line);
                return -1;
            }
            // using getTermByName because found sysnonym in GOA
            final String termAcn = line.substring(t + 1).trim();
            if (unknownAcn.contains(termAcn))
                continue;
            final GoTree.Term term = gotree.getTermByName(termAcn);
            if (term == null && !unknownAcn.contains(termAcn)) {
                unknownAcn.add(termAcn);
                LOG.warning("Don't know this GO term in " + line + " of " + this.geneFile + ". Could be obsolete, synonym, go specific division. Skipping.");
                continue;
            }
            final Node node = term2node.get(term);
            if (node == null) {
                r.close();
                LOG.error("Don't know this node in " + line + " of " + this.geneFile);
                return -1;
            }
            Set<Node> nodes = gene2node.get(gene);
            if (nodes == null) {
                nodes = new HashSet<>();
                gene2node.put(gene, nodes);
            }
            node.numGenes++;
            nodes.add(node);
        }
        ;
        // clean up
        unknownAcn.clear();
        r.close();
        final VcfIterator iter = openVcfIterator(oneFileOrNull(args));
        final VCFHeader header = iter.getHeader();
        final VepPredictionParser vepParser = new VepPredictionParserFactory(header).get();
        final AnnPredictionParser annParser = new AnnPredictionParserFactory(header).get();
        final Set<Pedigree.Person> persons;
        if (this.pedFile != null) {
            final Pedigree pedigree = Pedigree.newParser().parse(this.pedFile);
            persons = new Pedigree.CaseControlExtractor().extract(header, pedigree);
        } else {
            persons = new Pedigree.CaseControlExtractor().extract(header);
        }
        final Set<Pedigree.Person> affected = persons.stream().filter(P -> P.isAffected()).collect(Collectors.toSet());
        final Set<Pedigree.Person> unaffected = persons.stream().filter(P -> P.isUnaffected()).collect(Collectors.toSet());
        if (affected.isEmpty()) {
            LOG.error("No Affected individual");
            return -1;
        }
        if (unaffected.isEmpty()) {
            LOG.error("No unaffected individual");
            return -1;
        }
        final List<String> lookColumns = Arrays.asList("CCDS", "Feature", "ENSP", "Gene", "HGNC", "HGNC_ID", "SYMBOL", "RefSeq");
        final Predicate<Genotype> isWildGenotype = G -> {
            if (G == null)
                return false;
            return G.isHomRef();
        };
        final Predicate<Genotype> isAltGenotype = G -> {
            if (G == null)
                return false;
            return G.isCalled() && !G.isHomRef();
        };
        final SAMSequenceDictionaryProgress progress = new SAMSequenceDictionaryProgress(header).logger(LOG);
        while (iter.hasNext()) {
            final VariantContext ctx = progress.watch(iter.next());
            if (!this.variantFilter.test(ctx))
                continue;
            final Set<String> genes = new HashSet<>();
            for (final String predStr : ctx.getAttributeAsList(vepParser.getTag()).stream().map(O -> String.class.cast(O)).collect(Collectors.toList())) {
                final VepPredictionParser.VepPrediction pred = vepParser.parseOnePrediction(ctx, predStr);
                for (final String col : lookColumns) {
                    final String token = pred.getByCol(col);
                    if (!StringUtil.isBlank(token)) {
                        genes.add(token);
                    }
                }
            }
            for (final String predStr : ctx.getAttributeAsList(annParser.getTag()).stream().map(O -> String.class.cast(O)).collect(Collectors.toList())) {
                final AnnPredictionParser.AnnPrediction pred = annParser.parseOnePrediction(predStr);
                final String token = pred.getGeneName();
                if (!StringUtil.isBlank(token)) {
                    genes.add(token);
                }
            }
            if (genes.isEmpty())
                continue;
            final Set<Node> nodes = genes.stream().filter(G -> gene2node.containsKey(G)).flatMap(G -> gene2node.get(G).stream()).collect(Collectors.toSet());
            if (nodes.isEmpty())
                continue;
            final long unaffected_alt = unaffected.stream().map(P -> ctx.getGenotype(P.getId())).filter(G -> this.genotypeFilter.test(ctx, G)).filter(isAltGenotype).count();
            final long affected_alt = affected.stream().map(P -> ctx.getGenotype(P.getId())).filter(G -> this.genotypeFilter.test(ctx, G)).filter(isAltGenotype).count();
            /* no informative */
            if (unaffected_alt + affected_alt == 0L) {
                continue;
            }
            final long affected_ref = affected.stream().map(P -> ctx.getGenotype(P.getId())).filter(G -> this.genotypeFilter.test(ctx, G)).filter(isWildGenotype).count();
            final long unaffected_ref = unaffected.stream().map(P -> ctx.getGenotype(P.getId())).filter(G -> this.genotypeFilter.test(ctx, G)).filter(isWildGenotype).count();
            nodes.stream().forEach(N -> N.resetVisitedFlag());
            nodes.stream().forEach(N -> N.visit(unaffected_ref, unaffected_alt, affected_ref, affected_alt));
        }
        iter.close();
        progress.finish();
        LOG.info("Calculating Fisher and dumping.. please wait");
        final PrintWriter pw = super.openFileOrStdoutAsPrintWriter(this.outputFile);
        pw.println("#go_term\tfisher\tname\tgo_term_depth\tcount_genes_in_this_node" + "\tunaffected_ref_gt" + "\tunaffected_alt_gt" + "\taffected_ref_gt" + "\taffected_alt_gt");
        term2node.values().stream().filter(N -> this.show_never_seeen_term || N.sum() > 0L).sorted((n1, n2) -> Double.compare(n1.fisher(), n2.fisher())).forEach(N -> {
            pw.print(N.goTerm.getAcn());
            pw.print('\t');
            pw.print(N.fisher());
            pw.print("\t");
            pw.print(N.goTerm.getName().replaceAll("[ \',\\-]+", "_"));
            pw.print("\t");
            pw.print(N.goTerm.getMinDepth());
            pw.print('\t');
            pw.print(N.numGenes);
            pw.print('\t');
            pw.print(N.unaffected_ref);
            pw.print('\t');
            pw.print(N.unaffected_alt);
            pw.print('\t');
            pw.print(N.affected_ref);
            pw.print('\t');
            pw.print(N.affected_alt);
            pw.println();
        });
        pw.flush();
        pw.close();
        return 0;
    } catch (final Exception err) {
        LOG.error(err);
        return -1;
    }
}