2 * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3 * Copyright (C) $$Year-Rel$$ The Jalview Authors
5 * This file is part of Jalview.
7 * Jalview is free software: you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation, either version 3
10 * of the License, or (at your option) any later version.
12 * Jalview is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty
14 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15 * PURPOSE. See the GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
19 * The Jalview Authors are detailed in the 'AUTHORS' file.
23 import java.awt.Color;
24 import java.io.IOException;
25 import java.util.ArrayList;
26 import java.util.Arrays;
27 import java.util.Collections;
28 import java.util.HashMap;
29 import java.util.LinkedHashMap;
30 import java.util.List;
32 import java.util.Map.Entry;
33 import java.util.TreeMap;
35 import jalview.analysis.AlignmentUtils;
36 import jalview.analysis.SequenceIdMatcher;
37 import jalview.api.AlignViewportI;
38 import jalview.api.FeatureColourI;
39 import jalview.api.FeatureRenderer;
40 import jalview.api.FeaturesSourceI;
41 import jalview.datamodel.AlignedCodonFrame;
42 import jalview.datamodel.Alignment;
43 import jalview.datamodel.AlignmentI;
44 import jalview.datamodel.MappedFeatures;
45 import jalview.datamodel.SequenceDummy;
46 import jalview.datamodel.SequenceFeature;
47 import jalview.datamodel.SequenceI;
48 import jalview.datamodel.features.FeatureMatcherSet;
49 import jalview.datamodel.features.FeatureMatcherSetI;
50 import jalview.gui.Desktop;
51 import jalview.io.gff.GffHelperFactory;
52 import jalview.io.gff.GffHelperI;
53 import jalview.schemes.FeatureColour;
54 import jalview.util.ColorUtils;
55 import jalview.util.MapList;
56 import jalview.util.ParseHtmlBodyAndLinks;
57 import jalview.util.StringUtils;
60 * Parses and writes features files, which may be in Jalview, GFF2 or GFF3
61 * format. These are tab-delimited formats but with differences in the use of
64 * A Jalview feature file may define feature colours and then declare that the
65 * remainder of the file is in GFF format with the line 'GFF'.
67 * GFF3 files may include alignment mappings for features, which Jalview will
68 * attempt to model, and may include sequence data following a ##FASTA line.
75 public class FeaturesFile extends AlignFile implements FeaturesSourceI
77 private static final String EQUALS = "=";
79 private static final String TAB_REGEX = "\\t";
81 private static final String STARTGROUP = "STARTGROUP";
83 private static final String ENDGROUP = "ENDGROUP";
85 private static final String STARTFILTERS = "STARTFILTERS";
87 private static final String ENDFILTERS = "ENDFILTERS";
89 private static final String ID_NOT_SPECIFIED = "ID_NOT_SPECIFIED";
91 protected static final String GFF_VERSION = "##gff-version";
93 private AlignmentI lastmatchedAl = null;
95 private SequenceIdMatcher matcher = null;
97 protected AlignmentI dataset;
99 protected int gffVersion;
102 * Creates a new FeaturesFile object.
104 public FeaturesFile()
109 * Constructor which does not parse the file immediately
111 * @param file File or String filename
113 * @throws IOException
115 public FeaturesFile(Object file, DataSourceType paste)
118 super(false, file, paste);
123 * @throws IOException
125 public FeaturesFile(FileParse source) throws IOException
131 * Constructor that optionally parses the file immediately
133 * @param parseImmediately
136 * @throws IOException
138 public FeaturesFile(boolean parseImmediately, Object file,
139 DataSourceType type) throws IOException
141 super(parseImmediately, file, type);
145 * Parse GFF or sequence features file using case-independent matching,
149 * - alignment/dataset containing sequences that are to be annotated
151 * - hashtable to store feature colour definitions
153 * - process html strings into plain text
154 * @return true if features were added
156 public boolean parse(AlignmentI align,
157 Map<String, FeatureColourI> colours, boolean removeHTML)
159 return parse(align, colours, removeHTML, false);
163 * Extends the default addProperties by also adding peptide-to-cDNA mappings
164 * (if any) derived while parsing a GFF file
167 public void addProperties(AlignmentI al)
169 super.addProperties(al);
170 if (dataset != null && dataset.getCodonFrames() != null)
172 AlignmentI ds = (al.getDataset() == null) ? al : al.getDataset();
173 for (AlignedCodonFrame codons : dataset.getCodonFrames())
175 ds.addCodonFrame(codons);
181 * Parse GFF or Jalview format sequence features file
184 * - alignment/dataset containing sequences that are to be annotated
186 * - map to store feature colour definitions
188 * - process html strings into plain text
189 * @param relaxedIdmatching
190 * - when true, ID matches to compound sequence IDs are allowed
191 * @return true if features were added
193 public boolean parse(AlignmentI align,
194 Map<String, FeatureColourI> colours, boolean removeHTML,
195 boolean relaxedIdmatching)
197 return parse(align, colours, null, removeHTML, relaxedIdmatching);
201 * Parse GFF or Jalview format sequence features file
204 * - alignment/dataset containing sequences that are to be annotated
206 * - map to store feature colour definitions
208 * - map to store feature filter definitions
210 * - process html strings into plain text
211 * @param relaxedIdmatching
212 * - when true, ID matches to compound sequence IDs are allowed
213 * @return true if features were added
215 public boolean parse(AlignmentI align,
216 Map<String, FeatureColourI> colours,
217 Map<String, FeatureMatcherSetI> filters, boolean removeHTML,
218 boolean relaxedIdmatching)
220 Map<String, String> gffProps = new HashMap<>();
222 * keep track of any sequences we try to create from the data
224 List<SequenceI> newseqs = new ArrayList<>();
230 String featureGroup = null;
232 while ((line = nextLine()) != null)
234 // skip comments/process pragmas
235 if (line.length() == 0 || line.startsWith("#"))
237 if (line.toLowerCase().startsWith("##"))
239 processGffPragma(line, gffProps, align, newseqs);
244 gffColumns = line.split(TAB_REGEX);
245 if (gffColumns.length == 1)
247 if (line.trim().equalsIgnoreCase("GFF"))
250 * Jalview features file with appended GFF
251 * assume GFF2 (though it may declare ##gff-version 3)
258 if (gffColumns.length > 0 && gffColumns.length < 4)
261 * if 2 or 3 tokens, we anticipate either 'startgroup', 'endgroup' or
262 * a feature type colour specification
264 String ft = gffColumns[0];
265 if (ft.equalsIgnoreCase(STARTFILTERS))
267 parseFilters(filters);
270 if (ft.equalsIgnoreCase(STARTGROUP))
272 featureGroup = gffColumns[1];
274 else if (ft.equalsIgnoreCase(ENDGROUP))
276 // We should check whether this is the current group,
277 // but at present there's no way of showing more than 1 group
282 String colscheme = gffColumns[1];
283 FeatureColourI colour = FeatureColour
284 .parseJalviewFeatureColour(colscheme);
287 colours.put(ft, colour);
294 * if not a comment, GFF pragma, startgroup, endgroup or feature
295 * colour specification, that just leaves a feature details line
296 * in either Jalview or GFF format
300 parseJalviewFeature(line, gffColumns, align, colours, removeHTML,
301 relaxedIdmatching, featureGroup);
305 parseGff(gffColumns, align, relaxedIdmatching, newseqs);
309 } catch (Exception ex)
311 // should report somewhere useful for UI if necessary
312 warningMessage = ((warningMessage == null) ? "" : warningMessage)
313 + "Parsing error at\n" + line;
314 System.out.println("Error parsing feature file: " + ex + "\n" + line);
315 ex.printStackTrace(System.err);
321 * experimental - add any dummy sequences with features to the alignment
322 * - we need them for Ensembl feature extraction - though maybe not otherwise
324 for (SequenceI newseq : newseqs)
326 if (newseq.getFeatures().hasFeatures())
328 align.addSequence(newseq);
335 * Reads input lines from STARTFILTERS to ENDFILTERS and adds a feature type
336 * filter to the map for each line parsed. After exit from this method,
337 * nextLine() should return the line after ENDFILTERS (or we are already at
338 * end of file if ENDFILTERS was missing).
341 * @throws IOException
343 protected void parseFilters(Map<String, FeatureMatcherSetI> filters)
347 while ((line = nextLine()) != null)
349 if (line.toUpperCase().startsWith(ENDFILTERS))
353 String[] tokens = line.split(TAB_REGEX);
354 if (tokens.length != 2)
356 System.err.println(String.format("Invalid token count %d for %d",
357 tokens.length, line));
361 String featureType = tokens[0];
362 FeatureMatcherSetI fm = FeatureMatcherSet.fromString(tokens[1]);
363 if (fm != null && filters != null)
365 filters.put(featureType, fm);
372 * Try to parse a Jalview format feature specification and add it as a
373 * sequence feature to any matching sequences in the alignment. Returns true
374 * if successful (a feature was added), or false if not.
379 * @param featureColours
381 * @param relaxedIdmatching
382 * @param featureGroup
384 protected boolean parseJalviewFeature(String line, String[] gffColumns,
385 AlignmentI alignment, Map<String, FeatureColourI> featureColours,
386 boolean removeHTML, boolean relaxedIdMatching,
390 * tokens: description seqid seqIndex start end type [score]
392 if (gffColumns.length < 6)
394 System.err.println("Ignoring feature line '" + line
395 + "' with too few columns (" + gffColumns.length + ")");
398 String desc = gffColumns[0];
399 String seqId = gffColumns[1];
400 SequenceI seq = findSequence(seqId, alignment, null, relaxedIdMatching);
402 if (!ID_NOT_SPECIFIED.equals(seqId))
404 seq = findSequence(seqId, alignment, null, relaxedIdMatching);
410 String seqIndex = gffColumns[2];
413 int idx = Integer.parseInt(seqIndex);
414 seq = alignment.getSequenceAt(idx);
415 } catch (NumberFormatException ex)
417 System.err.println("Invalid sequence index: " + seqIndex);
423 System.out.println("Sequence not found: " + line);
427 int startPos = Integer.parseInt(gffColumns[3]);
428 int endPos = Integer.parseInt(gffColumns[4]);
430 String ft = gffColumns[5];
432 if (!featureColours.containsKey(ft))
435 * Perhaps an old style groups file with no colours -
436 * synthesize a colour from the feature type
438 Color colour = ColorUtils.createColourFromName(ft);
439 featureColours.put(ft, new FeatureColour(colour));
441 SequenceFeature sf = null;
442 if (gffColumns.length > 6)
444 float score = Float.NaN;
447 score = Float.valueOf(gffColumns[6]).floatValue();
448 } catch (NumberFormatException ex)
450 sf = new SequenceFeature(ft, desc, startPos, endPos, featureGroup);
452 sf = new SequenceFeature(ft, desc, startPos, endPos, score,
457 sf = new SequenceFeature(ft, desc, startPos, endPos, featureGroup);
460 parseDescriptionHTML(sf, removeHTML);
462 seq.addSequenceFeature(sf);
465 && (seq = alignment.findName(seq, seqId, false)) != null)
467 seq.addSequenceFeature(new SequenceFeature(sf));
473 * clear any temporary handles used to speed up ID matching
475 protected void resetMatcher()
477 lastmatchedAl = null;
482 * Returns a sequence matching the given id, as follows
484 * <li>strict matching is on exact sequence name</li>
485 * <li>relaxed matching allows matching on a token within the sequence name,
487 * <li>first tries to find a match in the alignment sequences</li>
488 * <li>else tries to find a match in the new sequences already generated while
489 * parsing the features file</li>
490 * <li>else creates a new placeholder sequence, adds it to the new sequences
491 * list, and returns it</li>
497 * @param relaxedIdMatching
501 protected SequenceI findSequence(String seqId, AlignmentI align,
502 List<SequenceI> newseqs, boolean relaxedIdMatching)
504 // TODO encapsulate in SequenceIdMatcher, share the matcher
505 // with the GffHelper (removing code duplication)
506 SequenceI match = null;
507 if (relaxedIdMatching)
509 if (lastmatchedAl != align)
511 lastmatchedAl = align;
512 matcher = new SequenceIdMatcher(align.getSequencesArray());
515 matcher.addAll(newseqs);
518 match = matcher.findIdMatch(seqId);
522 match = align.findName(seqId, true);
523 if (match == null && newseqs != null)
525 for (SequenceI m : newseqs)
527 if (seqId.equals(m.getName()))
535 if (match == null && newseqs != null)
537 match = new SequenceDummy(seqId);
538 if (relaxedIdMatching)
540 matcher.addAll(Arrays.asList(new SequenceI[] { match }));
542 // add dummy sequence to the newseqs list
548 public void parseDescriptionHTML(SequenceFeature sf, boolean removeHTML)
550 if (sf.getDescription() == null)
554 ParseHtmlBodyAndLinks parsed = new ParseHtmlBodyAndLinks(
555 sf.getDescription(), removeHTML, newline);
559 sf.setDescription(parsed.getNonHtmlContent());
562 for (String link : parsed.getLinks())
569 * Returns contents of a Jalview format features file, for visible features, as
570 * filtered by type and group. Features with a null group are displayed if their
571 * feature type is visible. Non-positional features may optionally be included
572 * (with no check on type or group).
576 * @param includeNonPositional
577 * if true, include non-positional features
578 * (regardless of group or type)
579 * @param includeComplement
580 * if true, include visible complementary
581 * (CDS/protein) positional features, with
582 * locations converted to local sequence
586 public String printJalviewFormat(SequenceI[] sequences,
587 FeatureRenderer fr, boolean includeNonPositional,
588 boolean includeComplement)
590 Map<String, FeatureColourI> visibleColours = fr
591 .getDisplayedFeatureCols();
592 Map<String, FeatureMatcherSetI> featureFilters = fr.getFeatureFilters();
595 * write out feature colours (if we know them)
597 // TODO: decide if feature links should also be written here ?
598 StringBuilder out = new StringBuilder(256);
599 if (visibleColours != null)
601 for (Entry<String, FeatureColourI> featureColour : visibleColours
604 FeatureColourI colour = featureColour.getValue();
605 out.append(colour.toJalviewFormat(featureColour.getKey())).append(
610 String[] types = visibleColours == null ? new String[0]
611 : visibleColours.keySet()
612 .toArray(new String[visibleColours.keySet().size()]);
615 * feature filters if any
617 outputFeatureFilters(out, visibleColours, featureFilters);
620 * output features within groups
622 int count = outputFeaturesByGroup(out, fr, types, sequences,
623 includeNonPositional);
625 if (includeComplement)
627 count += outputComplementFeatures(out, fr, sequences);
630 return count > 0 ? out.toString() : "No Features Visible";
634 * Outputs any visible complementary (CDS/peptide) positional features as
635 * Jalview format, within feature group. The coordinates of the linked features
636 * are converted to the corresponding positions of the local sequences.
643 private int outputComplementFeatures(StringBuilder out,
644 FeatureRenderer fr, SequenceI[] sequences)
646 AlignViewportI comp = fr.getViewport().getCodingComplement();
647 FeatureRenderer fr2 = Desktop.getAlignFrameFor(comp)
648 .getFeatureRenderer();
651 * bin features by feature group and sequence
653 Map<String, Map<String, List<SequenceFeature>>> map = new TreeMap<>(
654 String.CASE_INSENSITIVE_ORDER);
657 for (SequenceI seq : sequences)
660 * find complementary features
662 List<SequenceFeature> complementary = findComplementaryFeatures(seq,
664 String seqName = seq.getName();
666 for (SequenceFeature sf : complementary)
668 String group = sf.getFeatureGroup();
669 if (!map.containsKey(group))
671 map.put(group, new LinkedHashMap<>()); // preserves sequence order
673 Map<String, List<SequenceFeature>> groupFeatures = map.get(group);
674 if (!groupFeatures.containsKey(seqName))
676 groupFeatures.put(seqName, new ArrayList<>());
678 List<SequenceFeature> foundFeatures = groupFeatures.get(seqName);
679 foundFeatures.add(sf);
685 * output features by group
687 for (Entry<String, Map<String, List<SequenceFeature>>> groupFeatures : map.entrySet())
690 String group = groupFeatures.getKey();
691 if (!"".equals(group))
693 out.append(STARTGROUP).append(TAB).append(group).append(newline);
695 Map<String, List<SequenceFeature>> seqFeaturesMap = groupFeatures
697 for (Entry<String, List<SequenceFeature>> seqFeatures : seqFeaturesMap
700 String sequenceName = seqFeatures.getKey();
701 for (SequenceFeature sf : seqFeatures.getValue())
703 formatJalviewFeature(out, sequenceName, sf);
706 if (!"".equals(group))
708 out.append(ENDGROUP).append(TAB).append(group).append(newline);
716 * Answers a list of mapped features visible in the (CDS/protein) complement,
717 * with feature positions translated to local sequence coordinates
723 protected List<SequenceFeature> findComplementaryFeatures(SequenceI seq,
727 * avoid duplication of features (e.g. peptide feature
728 * at all 3 mapped codon positions)
730 List<SequenceFeature> found = new ArrayList<>();
731 List<SequenceFeature> complementary = new ArrayList<>();
733 for (int pos = seq.getStart(); pos <= seq.getEnd(); pos++)
735 MappedFeatures mf = fr2.findComplementFeaturesAtResidue(seq, pos);
739 for (SequenceFeature sf : mf.features)
742 * make a virtual feature with local coordinates
744 if (!found.contains(sf))
746 String group = sf.getFeatureGroup();
752 int begin = sf.getBegin();
753 int end = sf.getEnd();
754 int[] range = mf.getMappedPositions(begin, end);
755 SequenceFeature sf2 = new SequenceFeature(sf, range[0],
756 range[1], group, sf.getScore());
757 complementary.add(sf2);
763 return complementary;
767 * Outputs any feature filters defined for visible feature types, sandwiched by
768 * STARTFILTERS and ENDFILTERS lines
772 * @param featureFilters
774 void outputFeatureFilters(StringBuilder out,
775 Map<String, FeatureColourI> visible,
776 Map<String, FeatureMatcherSetI> featureFilters)
778 if (visible == null || featureFilters == null
779 || featureFilters.isEmpty())
784 boolean first = true;
785 for (String featureType : visible.keySet())
787 FeatureMatcherSetI filter = featureFilters.get(featureType);
793 out.append(newline).append(STARTFILTERS).append(newline);
795 out.append(featureType).append(TAB).append(filter.toStableString())
801 out.append(ENDFILTERS).append(newline);
807 * Appends output of visible sequence features within feature groups to the
808 * output buffer. Groups other than the null or empty group are sandwiched by
809 * STARTGROUP and ENDGROUP lines. Answers the number of features written.
813 * @param featureTypes
815 * @param includeNonPositional
818 private int outputFeaturesByGroup(StringBuilder out,
819 FeatureRenderer fr, String[] featureTypes,
820 SequenceI[] sequences, boolean includeNonPositional)
822 List<String> featureGroups = fr.getFeatureGroups();
825 * sort groups alphabetically, and ensure that features with a
826 * null or empty group are output after those in named groups
828 List<String> sortedGroups = new ArrayList<>(featureGroups);
829 sortedGroups.remove(null);
830 sortedGroups.remove("");
831 Collections.sort(sortedGroups);
832 sortedGroups.add(null);
833 sortedGroups.add("");
836 List<String> visibleGroups = fr.getDisplayedFeatureGroups();
839 * loop over all groups (may be visible or not);
840 * non-positional features are output even if group is not visible
842 for (String group : sortedGroups)
844 boolean firstInGroup = true;
845 boolean isNullGroup = group == null || "".equals(group);
847 for (int i = 0; i < sequences.length; i++)
849 String sequenceName = sequences[i].getName();
850 List<SequenceFeature> features = new ArrayList<>();
853 * get any non-positional features in this group, if wanted
854 * (for any feature type, whether visible or not)
856 if (includeNonPositional)
858 features.addAll(sequences[i].getFeatures()
859 .getFeaturesForGroup(false, group));
863 * add positional features for visible feature types, but
864 * (for named groups) only if feature group is visible
866 if (featureTypes.length > 0
867 && (isNullGroup || visibleGroups.contains(group)))
869 features.addAll(sequences[i].getFeatures().getFeaturesForGroup(
870 true, group, featureTypes));
873 for (SequenceFeature sf : features)
875 if (sf.isNonPositional() || fr.isVisible(sf))
883 out.append(STARTGROUP).append(TAB).append(group)
887 firstInGroup = false;
888 formatJalviewFeature(out, sequenceName, sf);
893 if (!isNullGroup && !firstInGroup)
895 out.append(ENDGROUP).append(TAB).append(group).append(newline);
902 * Formats one feature in Jalview format and appends to the string buffer
905 * @param sequenceName
906 * @param sequenceFeature
908 protected void formatJalviewFeature(
909 StringBuilder out, String sequenceName,
910 SequenceFeature sequenceFeature)
912 if (sequenceFeature.description == null
913 || sequenceFeature.description.equals(""))
915 out.append(sequenceFeature.type).append(TAB);
919 if (sequenceFeature.links != null
920 && sequenceFeature.getDescription().indexOf("<html>") == -1)
922 out.append("<html>");
925 out.append(sequenceFeature.description);
926 if (sequenceFeature.links != null)
928 for (int l = 0; l < sequenceFeature.links.size(); l++)
930 String label = sequenceFeature.links.elementAt(l);
931 String href = label.substring(label.indexOf("|") + 1);
932 label = label.substring(0, label.indexOf("|"));
934 if (sequenceFeature.description.indexOf(href) == -1)
936 out.append(" <a href=\"").append(href).append("\">")
937 .append(label).append("</a>");
941 if (sequenceFeature.getDescription().indexOf("</html>") == -1)
943 out.append("</html>");
949 out.append(sequenceName);
950 out.append("\t-1\t");
951 out.append(sequenceFeature.begin);
953 out.append(sequenceFeature.end);
955 out.append(sequenceFeature.type);
956 if (!Float.isNaN(sequenceFeature.score))
959 out.append(sequenceFeature.score);
965 * Parse method that is called when a GFF file is dragged to the desktop
970 AlignViewportI av = getViewport();
973 if (av.getAlignment() != null)
975 dataset = av.getAlignment().getDataset();
979 // working in the applet context ?
980 dataset = av.getAlignment();
985 dataset = new Alignment(new SequenceI[] {});
988 Map<String, FeatureColourI> featureColours = new HashMap<>();
989 boolean parseResult = parse(dataset, featureColours, false, true);
992 // pass error up somehow
996 // update viewport with the dataset data ?
1000 setSeqs(dataset.getSequencesArray());
1005 * Implementation of unused abstract method
1007 * @return error message
1010 public String print(SequenceI[] sqs, boolean jvsuffix)
1012 System.out.println("Use printGffFormat() or printJalviewFormat()");
1017 * Returns features output in GFF2 format
1020 * the sequences whose features are to be
1023 * a map whose keys are the type names of
1025 * @param visibleFeatureGroups
1026 * @param includeNonPositionalFeatures
1027 * @param includeComplement
1030 public String printGffFormat(SequenceI[] sequences,
1031 FeatureRenderer fr, boolean includeNonPositionalFeatures,
1032 boolean includeComplement)
1034 FeatureRenderer fr2 = null;
1035 if (includeComplement)
1037 AlignViewportI comp = fr.getViewport().getCodingComplement();
1038 fr2 = Desktop.getAlignFrameFor(comp).getFeatureRenderer();
1041 Map<String, FeatureColourI> visibleColours = fr.getDisplayedFeatureCols();
1043 StringBuilder out = new StringBuilder(256);
1045 out.append(String.format("%s %d\n", GFF_VERSION, gffVersion == 0 ? 2 : gffVersion));
1047 String[] types = visibleColours == null ? new String[0]
1048 : visibleColours.keySet()
1049 .toArray(new String[visibleColours.keySet().size()]);
1051 for (SequenceI seq : sequences)
1053 List<SequenceFeature> seqFeatures = new ArrayList<>();
1054 List<SequenceFeature> features = new ArrayList<>();
1055 if (includeNonPositionalFeatures)
1057 features.addAll(seq.getFeatures().getNonPositionalFeatures());
1059 if (visibleColours != null && !visibleColours.isEmpty())
1061 features.addAll(seq.getFeatures().getPositionalFeatures(types));
1063 for (SequenceFeature sf : features)
1065 if (sf.isNonPositional() || fr.isVisible(sf))
1068 * drop features hidden by group visibility, colour threshold,
1069 * or feature filter condition
1071 seqFeatures.add(sf);
1075 if (includeComplement)
1077 seqFeatures.addAll(findComplementaryFeatures(seq, fr2));
1081 * sort features here if wanted
1083 for (SequenceFeature sf : seqFeatures)
1085 formatGffFeature(out, seq, sf);
1086 out.append(newline);
1090 return out.toString();
1094 * Formats one feature as GFF and appends to the string buffer
1096 private void formatGffFeature(StringBuilder out, SequenceI seq,
1099 String source = sf.featureGroup;
1102 source = sf.getDescription();
1105 out.append(seq.getName());
1109 out.append(sf.type);
1111 out.append(sf.begin);
1115 out.append(sf.score);
1118 int strand = sf.getStrand();
1119 out.append(strand == 1 ? "+" : (strand == -1 ? "-" : "."));
1122 String phase = sf.getPhase();
1123 out.append(phase == null ? "." : phase);
1125 if (sf.otherDetails != null && !sf.otherDetails.isEmpty())
1127 Map<String, Object> map = sf.otherDetails;
1128 formatAttributes(out, map);
1133 * A helper method that outputs attributes stored in the map as
1134 * semicolon-delimited values e.g.
1137 * AC_Male=0;AF_NFE=0.00000e 00;Hom_FIN=0;GQ_MEDIAN=9
1140 * A map-valued attribute is formatted as a comma-delimited list within braces,
1144 * jvmap_CSQ={ALLELE_NUM=1,UNIPARC=UPI0002841053,Feature=ENST00000585561}
1147 * The {@code jvmap_} prefix designates a values map and is removed if the value
1148 * is parsed when read in. (The GFF3 specification allows 'semi-structured data'
1149 * to be represented provided the attribute name begins with a lower case
1154 * @see http://gmod.org/wiki/GFF3#GFF3_Format
1156 void formatAttributes(StringBuilder sb, Map<String, Object> map)
1159 boolean first = true;
1160 for (String key : map.keySet())
1162 if (SequenceFeature.STRAND.equals(key)
1163 || SequenceFeature.PHASE.equals(key))
1166 * values stashed in map but output to their own columns
1177 Object value = map.get(key);
1178 if (value instanceof Map<?, ?>)
1180 formatMapAttribute(sb, key, (Map<?, ?>) value);
1184 String formatted = StringUtils.urlEncode(value.toString(),
1185 GffHelperI.GFF_ENCODABLE);
1186 sb.append(key).append(EQUALS).append(formatted);
1192 * Formats the map entries as
1195 * key=key1=value1,key2=value2,...
1198 * and appends this to the string buffer
1204 private void formatMapAttribute(StringBuilder sb, String key,
1207 if (map == null || map.isEmpty())
1213 * AbstractMap.toString would be a shortcut here, but more reliable
1214 * to code the required format in case toString changes in future
1216 sb.append(key).append(EQUALS);
1217 boolean first = true;
1218 for (Entry<?, ?> entry : map.entrySet())
1225 sb.append(entry.getKey().toString()).append(EQUALS);
1226 String formatted = StringUtils.urlEncode(entry.getValue().toString(),
1227 GffHelperI.GFF_ENCODABLE);
1228 sb.append(formatted);
1233 * Returns a mapping given list of one or more Align descriptors (exonerate
1236 * @param alignedRegions
1237 * a list of "Align fromStart toStart fromCount"
1238 * @param mapIsFromCdna
1239 * if true, 'from' is dna, else 'from' is protein
1241 * either 1 (forward) or -1 (reverse)
1243 * @throws IOException
1245 protected MapList constructCodonMappingFromAlign(
1246 List<String> alignedRegions, boolean mapIsFromCdna, int strand)
1251 throw new IOException(
1252 "Invalid strand for a codon mapping (cannot be 0)");
1254 int regions = alignedRegions.size();
1255 // arrays to hold [start, end] for each aligned region
1256 int[] fromRanges = new int[regions * 2]; // from dna
1257 int[] toRanges = new int[regions * 2]; // to protein
1258 int fromRangesIndex = 0;
1259 int toRangesIndex = 0;
1261 for (String range : alignedRegions)
1264 * Align mapFromStart mapToStart mapFromCount
1265 * e.g. if mapIsFromCdna
1266 * Align 11270 143 120
1268 * 120 bases from pos 11270 align to pos 143 in peptide
1269 * if !mapIsFromCdna this would instead be
1270 * Align 143 11270 40
1272 String[] tokens = range.split(" ");
1273 if (tokens.length != 3)
1275 throw new IOException("Wrong number of fields for Align");
1282 fromStart = Integer.parseInt(tokens[0]);
1283 toStart = Integer.parseInt(tokens[1]);
1284 fromCount = Integer.parseInt(tokens[2]);
1285 } catch (NumberFormatException nfe)
1287 throw new IOException(
1288 "Invalid number in Align field: " + nfe.getMessage());
1292 * Jalview always models from dna to protein, so adjust values if the
1293 * GFF mapping is from protein to dna
1298 int temp = fromStart;
1299 fromStart = toStart;
1302 fromRanges[fromRangesIndex++] = fromStart;
1303 fromRanges[fromRangesIndex++] = fromStart + strand * (fromCount - 1);
1306 * If a codon has an intron gap, there will be contiguous 'toRanges';
1307 * this is handled for us by the MapList constructor.
1308 * (It is not clear that exonerate ever generates this case)
1310 toRanges[toRangesIndex++] = toStart;
1311 toRanges[toRangesIndex++] = toStart + (fromCount - 1) / 3;
1314 return new MapList(fromRanges, toRanges, 3, 1);
1318 * Parse a GFF format feature. This may include creating a 'dummy' sequence to
1319 * hold the feature, or for its mapped sequence, or both, to be resolved
1320 * either later in the GFF file (##FASTA section), or when the user loads
1321 * additional sequences.
1325 * @param relaxedIdMatching
1329 protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment,
1330 boolean relaxedIdMatching, List<SequenceI> newseqs)
1333 * GFF: seqid source type start end score strand phase [attributes]
1335 if (gffColumns.length < 5)
1337 System.err.println("Ignoring GFF feature line with too few columns ("
1338 + gffColumns.length + ")");
1343 * locate referenced sequence in alignment _or_
1344 * as a forward or external reference (SequenceDummy)
1346 String seqId = gffColumns[0];
1347 SequenceI seq = findSequence(seqId, alignment, newseqs,
1350 SequenceFeature sf = null;
1351 GffHelperI helper = GffHelperFactory.getHelper(gffColumns);
1356 sf = helper.processGff(seq, gffColumns, alignment, newseqs,
1360 seq.addSequenceFeature(sf);
1361 while ((seq = alignment.findName(seq, seqId, true)) != null)
1363 seq.addSequenceFeature(new SequenceFeature(sf));
1366 } catch (IOException e)
1368 System.err.println("GFF parsing failed with: " + e.getMessage());
1377 * After encountering ##fasta in a GFF3 file, process the remainder of the
1378 * file as FAST sequence data. Any placeholder sequences created during
1379 * feature parsing are updated with the actual sequences.
1383 * @throws IOException
1385 protected void processAsFasta(AlignmentI align, List<SequenceI> newseqs)
1391 } catch (IOException q)
1394 FastaFile parser = new FastaFile(this);
1395 List<SequenceI> includedseqs = parser.getSeqs();
1397 SequenceIdMatcher smatcher = new SequenceIdMatcher(newseqs);
1400 * iterate over includedseqs, and replacing matching ones with newseqs
1401 * sequences. Generic iterator not used here because we modify
1402 * includedseqs as we go
1404 for (int p = 0, pSize = includedseqs.size(); p < pSize; p++)
1406 // search for any dummy seqs that this sequence can be used to update
1407 SequenceI includedSeq = includedseqs.get(p);
1408 SequenceI dummyseq = smatcher.findIdMatch(includedSeq);
1409 if (dummyseq != null && dummyseq instanceof SequenceDummy)
1411 // probably have the pattern wrong
1412 // idea is that a flyweight proxy for a sequence ID can be created for
1413 // 1. stable reference creation
1414 // 2. addition of annotation
1415 // 3. future replacement by a real sequence
1416 // current pattern is to create SequenceDummy objects - a convenience
1417 // constructor for a Sequence.
1418 // problem is that when promoted to a real sequence, all references
1419 // need to be updated somehow. We avoid that by keeping the same object.
1420 ((SequenceDummy) dummyseq).become(includedSeq);
1421 dummyseq.createDatasetSequence();
1424 * Update mappings so they are now to the dataset sequence
1426 for (AlignedCodonFrame mapping : align.getCodonFrames())
1428 mapping.updateToDataset(dummyseq);
1432 * replace parsed sequence with the realised forward reference
1434 includedseqs.set(p, dummyseq);
1437 * and remove from the newseqs list
1439 newseqs.remove(dummyseq);
1444 * finally add sequences to the dataset
1446 for (SequenceI seq : includedseqs)
1448 // experimental: mapping-based 'alignment' to query sequence
1449 AlignmentUtils.alignSequenceAs(seq, align,
1450 String.valueOf(align.getGapCharacter()), false, true);
1452 // rename sequences if GFF handler requested this
1453 // TODO a more elegant way e.g. gffHelper.postProcess(newseqs) ?
1454 List<SequenceFeature> sfs = seq.getFeatures().getPositionalFeatures();
1457 String newName = (String) sfs.get(0).getValue(
1458 GffHelperI.RENAME_TOKEN);
1459 if (newName != null)
1461 seq.setName(newName);
1464 align.addSequence(seq);
1469 * Process a ## directive
1475 * @throws IOException
1477 protected void processGffPragma(String line, Map<String, String> gffProps,
1478 AlignmentI align, List<SequenceI> newseqs) throws IOException
1481 if ("###".equals(line))
1483 // close off any open 'forward references'
1487 String[] tokens = line.substring(2).split(" ");
1488 String pragma = tokens[0];
1489 String value = tokens.length == 1 ? null : tokens[1];
1491 if ("gff-version".equalsIgnoreCase(pragma))
1497 // value may be e.g. "3.1.2"
1498 gffVersion = Integer.parseInt(value.split("\\.")[0]);
1499 } catch (NumberFormatException e)
1505 else if ("sequence-region".equalsIgnoreCase(pragma))
1507 // could capture <seqid start end> if wanted here
1509 else if ("feature-ontology".equalsIgnoreCase(pragma))
1511 // should resolve against the specified feature ontology URI
1513 else if ("attribute-ontology".equalsIgnoreCase(pragma))
1515 // URI of attribute ontology - not currently used in GFF3
1517 else if ("source-ontology".equalsIgnoreCase(pragma))
1519 // URI of source ontology - not currently used in GFF3
1521 else if ("species-build".equalsIgnoreCase(pragma))
1523 // save URI of specific NCBI taxon version of annotations
1524 gffProps.put("species-build", value);
1526 else if ("fasta".equalsIgnoreCase(pragma))
1528 // process the rest of the file as a fasta file and replace any dummy
1530 processAsFasta(align, newseqs);
1534 System.err.println("Ignoring unknown pragma: " + line);