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 jalview.analysis.AlignmentUtils;
24 import jalview.analysis.SequenceIdMatcher;
25 import jalview.api.AlignViewportI;
26 import jalview.api.FeatureColourI;
27 import jalview.api.FeatureRenderer;
28 import jalview.api.FeaturesSourceI;
29 import jalview.datamodel.AlignedCodonFrame;
30 import jalview.datamodel.Alignment;
31 import jalview.datamodel.AlignmentI;
32 import jalview.datamodel.SequenceDummy;
33 import jalview.datamodel.SequenceFeature;
34 import jalview.datamodel.SequenceI;
35 import jalview.datamodel.features.FeatureMatcherSet;
36 import jalview.datamodel.features.FeatureMatcherSetI;
37 import jalview.io.gff.GffHelperBase;
38 import jalview.io.gff.GffHelperFactory;
39 import jalview.io.gff.GffHelperI;
40 import jalview.schemes.FeatureColour;
41 import jalview.util.ColorUtils;
42 import jalview.util.MapList;
43 import jalview.util.ParseHtmlBodyAndLinks;
44 import jalview.util.StringUtils;
46 import java.awt.Color;
47 import java.io.IOException;
48 import java.util.ArrayList;
49 import java.util.Arrays;
50 import java.util.Collections;
51 import java.util.HashMap;
52 import java.util.List;
54 import java.util.Map.Entry;
57 * Parses and writes features files, which may be in Jalview, GFF2 or GFF3
58 * format. These are tab-delimited formats but with differences in the use of
61 * A Jalview feature file may define feature colours and then declare that the
62 * remainder of the file is in GFF format with the line 'GFF'.
64 * GFF3 files may include alignment mappings for features, which Jalview will
65 * attempt to model, and may include sequence data following a ##FASTA line.
72 public class FeaturesFile extends AlignFile implements FeaturesSourceI
74 private static final String TAB_REGEX = "\\t";
76 private static final String STARTGROUP = "STARTGROUP";
78 private static final String ENDGROUP = "ENDGROUP";
80 private static final String STARTFILTERS = "STARTFILTERS";
82 private static final String ENDFILTERS = "ENDFILTERS";
84 private static final String ID_NOT_SPECIFIED = "ID_NOT_SPECIFIED";
86 private static final String NOTE = "Note";
88 protected static final String GFF_VERSION = "##gff-version";
90 private AlignmentI lastmatchedAl = null;
92 private SequenceIdMatcher matcher = null;
94 protected AlignmentI dataset;
96 protected int gffVersion;
99 * Creates a new FeaturesFile object.
101 public FeaturesFile()
106 * Constructor which does not parse the file immediately
110 * @throws IOException
112 public FeaturesFile(String file, DataSourceType paste)
115 super(false, file, paste);
120 * @throws IOException
122 public FeaturesFile(FileParse source) throws IOException
128 * Constructor that optionally parses the file immediately
130 * @param parseImmediately
133 * @throws IOException
135 public FeaturesFile(boolean parseImmediately, String file,
136 DataSourceType type) throws IOException
138 super(parseImmediately, file, type);
142 * Parse GFF or sequence features file using case-independent matching,
146 * - alignment/dataset containing sequences that are to be annotated
148 * - hashtable to store feature colour definitions
150 * - process html strings into plain text
151 * @return true if features were added
153 public boolean parse(AlignmentI align,
154 Map<String, FeatureColourI> colours, boolean removeHTML)
156 return parse(align, colours, removeHTML, false);
160 * Extends the default addProperties by also adding peptide-to-cDNA mappings
161 * (if any) derived while parsing a GFF file
164 public void addProperties(AlignmentI al)
166 super.addProperties(al);
167 if (dataset != null && dataset.getCodonFrames() != null)
169 AlignmentI ds = (al.getDataset() == null) ? al : al.getDataset();
170 for (AlignedCodonFrame codons : dataset.getCodonFrames())
172 ds.addCodonFrame(codons);
178 * Parse GFF or Jalview format sequence features file
181 * - alignment/dataset containing sequences that are to be annotated
183 * - map to store feature colour definitions
185 * - process html strings into plain text
186 * @param relaxedIdmatching
187 * - when true, ID matches to compound sequence IDs are allowed
188 * @return true if features were added
190 public boolean parse(AlignmentI align,
191 Map<String, FeatureColourI> colours, boolean removeHTML,
192 boolean relaxedIdmatching)
194 return parse(align, colours, null, removeHTML, relaxedIdmatching);
198 * Parse GFF or Jalview format sequence features file
201 * - alignment/dataset containing sequences that are to be annotated
203 * - map to store feature colour definitions
205 * - map to store feature filter definitions
207 * - process html strings into plain text
208 * @param relaxedIdmatching
209 * - when true, ID matches to compound sequence IDs are allowed
210 * @return true if features were added
212 public boolean parse(AlignmentI align,
213 Map<String, FeatureColourI> colours,
214 Map<String, FeatureMatcherSetI> filters, boolean removeHTML,
215 boolean relaxedIdmatching)
217 Map<String, String> gffProps = new HashMap<>();
219 * keep track of any sequences we try to create from the data
221 List<SequenceI> newseqs = new ArrayList<>();
227 String featureGroup = null;
229 while ((line = nextLine()) != null)
231 // skip comments/process pragmas
232 if (line.length() == 0 || line.startsWith("#"))
234 if (line.toLowerCase().startsWith("##"))
236 processGffPragma(line, gffProps, align, newseqs);
241 gffColumns = line.split(TAB_REGEX);
242 if (gffColumns.length == 1)
244 if (line.trim().equalsIgnoreCase("GFF"))
247 * Jalview features file with appended GFF
248 * assume GFF2 (though it may declare ##gff-version 3)
255 if (gffColumns.length > 0 && gffColumns.length < 4)
258 * if 2 or 3 tokens, we anticipate either 'startgroup', 'endgroup' or
259 * a feature type colour specification
261 String ft = gffColumns[0];
262 if (ft.equalsIgnoreCase(STARTFILTERS))
264 parseFilters(filters);
267 if (ft.equalsIgnoreCase(STARTGROUP))
269 featureGroup = gffColumns[1];
271 else if (ft.equalsIgnoreCase(ENDGROUP))
273 // We should check whether this is the current group,
274 // but at present there's no way of showing more than 1 group
279 String colscheme = gffColumns[1];
280 FeatureColourI colour = FeatureColour
281 .parseJalviewFeatureColour(colscheme);
284 colours.put(ft, colour);
291 * if not a comment, GFF pragma, startgroup, endgroup or feature
292 * colour specification, that just leaves a feature details line
293 * in either Jalview or GFF format
297 parseJalviewFeature(line, gffColumns, align, colours, removeHTML,
298 relaxedIdmatching, featureGroup);
302 parseGff(gffColumns, align, relaxedIdmatching, newseqs);
306 } catch (Exception ex)
308 // should report somewhere useful for UI if necessary
309 warningMessage = ((warningMessage == null) ? "" : warningMessage)
310 + "Parsing error at\n" + line;
311 System.out.println("Error parsing feature file: " + ex + "\n" + line);
312 ex.printStackTrace(System.err);
318 * experimental - add any dummy sequences with features to the alignment
319 * - we need them for Ensembl feature extraction - though maybe not otherwise
321 for (SequenceI newseq : newseqs)
323 if (newseq.getFeatures().hasFeatures())
325 align.addSequence(newseq);
332 * Reads input lines from STARTFILTERS to ENDFILTERS and adds a feature type
333 * filter to the map for each line parsed. After exit from this method,
334 * nextLine() should return the line after ENDFILTERS (or we are already at
335 * end of file if ENDFILTERS was missing).
338 * @throws IOException
340 protected void parseFilters(Map<String, FeatureMatcherSetI> filters)
344 while ((line = nextLine()) != null)
346 if (line.toUpperCase().startsWith(ENDFILTERS))
350 String[] tokens = line.split(TAB_REGEX);
351 if (tokens.length != 2)
353 System.err.println(String.format("Invalid token count %d for %d",
354 tokens.length, line));
358 String featureType = tokens[0];
359 FeatureMatcherSetI fm = FeatureMatcherSet.fromString(tokens[1]);
360 if (fm != null && filters != null)
362 filters.put(featureType, fm);
369 * Try to parse a Jalview format feature specification and add it as a
370 * sequence feature to any matching sequences in the alignment. Returns true
371 * if successful (a feature was added), or false if not.
376 * @param featureColours
378 * @param relaxedIdmatching
379 * @param featureGroup
381 protected boolean parseJalviewFeature(String line, String[] gffColumns,
382 AlignmentI alignment, Map<String, FeatureColourI> featureColours,
383 boolean removeHTML, boolean relaxedIdMatching,
387 * tokens: description seqid seqIndex start end type [score]
389 if (gffColumns.length < 6)
391 System.err.println("Ignoring feature line '" + line
392 + "' with too few columns (" + gffColumns.length + ")");
395 String desc = gffColumns[0];
396 String seqId = gffColumns[1];
397 SequenceI seq = findSequence(seqId, alignment, null, relaxedIdMatching);
399 if (!ID_NOT_SPECIFIED.equals(seqId))
401 seq = findSequence(seqId, alignment, null, relaxedIdMatching);
407 String seqIndex = gffColumns[2];
410 int idx = Integer.parseInt(seqIndex);
411 seq = alignment.getSequenceAt(idx);
412 } catch (NumberFormatException ex)
414 System.err.println("Invalid sequence index: " + seqIndex);
420 System.out.println("Sequence not found: " + line);
424 int startPos = Integer.parseInt(gffColumns[3]);
425 int endPos = Integer.parseInt(gffColumns[4]);
427 String ft = gffColumns[5];
429 if (!featureColours.containsKey(ft))
432 * Perhaps an old style groups file with no colours -
433 * synthesize a colour from the feature type
435 Color colour = ColorUtils.createColourFromName(ft);
436 featureColours.put(ft, new FeatureColour(colour));
438 SequenceFeature sf = null;
439 if (gffColumns.length > 6)
441 float score = Float.NaN;
444 score = new Float(gffColumns[6]).floatValue();
445 } catch (NumberFormatException ex)
447 sf = new SequenceFeature(ft, desc, startPos, endPos, featureGroup);
449 sf = new SequenceFeature(ft, desc, startPos, endPos, score,
454 sf = new SequenceFeature(ft, desc, startPos, endPos, featureGroup);
457 parseDescriptionHTML(sf, removeHTML);
459 seq.addSequenceFeature(sf);
462 && (seq = alignment.findName(seq, seqId, false)) != null)
464 seq.addSequenceFeature(new SequenceFeature(sf));
470 * clear any temporary handles used to speed up ID matching
472 protected void resetMatcher()
474 lastmatchedAl = null;
479 * Returns a sequence matching the given id, as follows
481 * <li>strict matching is on exact sequence name</li>
482 * <li>relaxed matching allows matching on a token within the sequence name,
484 * <li>first tries to find a match in the alignment sequences</li>
485 * <li>else tries to find a match in the new sequences already generated while
486 * parsing the features file</li>
487 * <li>else creates a new placeholder sequence, adds it to the new sequences
488 * list, and returns it</li>
494 * @param relaxedIdMatching
498 protected SequenceI findSequence(String seqId, AlignmentI align,
499 List<SequenceI> newseqs, boolean relaxedIdMatching)
501 // TODO encapsulate in SequenceIdMatcher, share the matcher
502 // with the GffHelper (removing code duplication)
503 SequenceI match = null;
504 if (relaxedIdMatching)
506 if (lastmatchedAl != align)
508 lastmatchedAl = align;
509 matcher = new SequenceIdMatcher(align.getSequencesArray());
512 matcher.addAll(newseqs);
515 match = matcher.findIdMatch(seqId);
519 match = align.findName(seqId, true);
520 if (match == null && newseqs != null)
522 for (SequenceI m : newseqs)
524 if (seqId.equals(m.getName()))
532 if (match == null && newseqs != null)
534 match = new SequenceDummy(seqId);
535 if (relaxedIdMatching)
537 matcher.addAll(Arrays.asList(new SequenceI[] { match }));
539 // add dummy sequence to the newseqs list
545 public void parseDescriptionHTML(SequenceFeature sf, boolean removeHTML)
547 if (sf.getDescription() == null)
551 ParseHtmlBodyAndLinks parsed = new ParseHtmlBodyAndLinks(
552 sf.getDescription(), removeHTML, newline);
556 sf.setDescription(parsed.getNonHtmlContent());
559 for (String link : parsed.getLinks())
566 * Returns contents of a Jalview format features file, for visible features,
567 * as filtered by type and group. Features with a null group are displayed if
568 * their feature type is visible. Non-positional features may optionally be
569 * included (with no check on type or group).
573 * @param includeNonPositional
574 * if true, include non-positional features (regardless of group or
578 public String printJalviewFormat(SequenceI[] sequences,
579 FeatureRenderer fr, boolean includeNonPositional)
581 Map<String, FeatureColourI> visibleColours = fr
582 .getDisplayedFeatureCols();
583 Map<String, FeatureMatcherSetI> featureFilters = fr.getFeatureFilters();
584 List<String> visibleFeatureGroups = fr.getDisplayedFeatureGroups();
586 if (!includeNonPositional
587 && (visibleColours == null || visibleColours.isEmpty()))
589 // no point continuing.
590 return "No Features Visible";
594 * write out feature colours (if we know them)
596 // TODO: decide if feature links should also be written here ?
597 StringBuilder out = new StringBuilder(256);
598 if (visibleColours != null)
600 for (Entry<String, FeatureColourI> featureColour : visibleColours
603 FeatureColourI colour = featureColour.getValue();
604 out.append(colour.toJalviewFormat(featureColour.getKey())).append(
609 String[] types = visibleColours == null ? new String[0]
610 : visibleColours.keySet()
611 .toArray(new String[visibleColours.keySet().size()]);
614 * feature filters if any
616 outputFeatureFilters(out, visibleColours, featureFilters);
619 * sort groups alphabetically, and ensure that features with a
620 * null or empty group are output after those in named groups
622 List<String> sortedGroups = new ArrayList<>(visibleFeatureGroups);
623 sortedGroups.remove(null);
624 sortedGroups.remove("");
625 Collections.sort(sortedGroups);
626 sortedGroups.add(null);
627 sortedGroups.add("");
629 boolean foundSome = false;
632 * first output any non-positional features
634 if (includeNonPositional)
636 for (int i = 0; i < sequences.length; i++)
638 String sequenceName = sequences[i].getName();
639 for (SequenceFeature feature : sequences[i].getFeatures()
640 .getNonPositionalFeatures())
643 out.append(formatJalviewFeature(sequenceName, feature));
649 * positional features within groups
651 foundSome |= outputFeaturesByGroup(out, fr, sortedGroups, types,
654 return foundSome ? out.toString() : "No Features Visible";
658 * Outputs any feature filters defined for visible feature types, sandwiched by
659 * STARTFILTERS and ENDFILTERS lines
663 * @param featureFilters
665 void outputFeatureFilters(StringBuilder out,
666 Map<String, FeatureColourI> visible,
667 Map<String, FeatureMatcherSetI> featureFilters)
669 if (visible == null || featureFilters == null
670 || featureFilters.isEmpty())
675 boolean first = true;
676 for (String featureType : visible.keySet())
678 FeatureMatcherSetI filter = featureFilters.get(featureType);
684 out.append(newline).append(STARTFILTERS).append(newline);
686 out.append(featureType).append(TAB).append(filter.toStableString())
692 out.append(ENDFILTERS).append(newline);
698 * Appends output of visible sequence features within feature groups to the
699 * output buffer. Groups other than the null or empty group are sandwiched by
700 * STARTGROUP and ENDGROUP lines. Answers true if at least one feature was
701 * written, else false.
706 * @param featureTypes
710 private boolean outputFeaturesByGroup(StringBuilder out,
711 FeatureRenderer fr, List<String> groups, String[] featureTypes,
712 SequenceI[] sequences)
714 boolean foundSome = false;
715 for (String group : groups)
717 boolean firstInGroup = true;
718 boolean isNamedGroup = (group != null && !"".equals(group));
721 * output positional features within groups
723 for (int i = 0; i < sequences.length; i++)
725 String sequenceName = sequences[i].getName();
726 List<SequenceFeature> features = new ArrayList<>();
727 if (featureTypes.length > 0)
729 features.addAll(sequences[i].getFeatures().getFeaturesForGroup(
730 true, group, featureTypes));
733 for (SequenceFeature sequenceFeature : features)
735 if (fr.isVisible(sequenceFeature))
738 if (firstInGroup && isNamedGroup)
740 out.append(newline).append(STARTGROUP).append(TAB)
741 .append(group).append(newline);
743 firstInGroup = false;
744 out.append(formatJalviewFeature(sequenceName, sequenceFeature));
749 if (isNamedGroup && !firstInGroup)
751 out.append(ENDGROUP).append(TAB).append(group).append(newline);
759 * @param sequenceName
760 * @param sequenceFeature
762 protected String formatJalviewFeature(
763 String sequenceName, SequenceFeature sequenceFeature)
765 StringBuilder out = new StringBuilder(64);
766 if (sequenceFeature.description == null
767 || sequenceFeature.description.equals(""))
769 out.append(sequenceFeature.type).append(TAB);
773 if (sequenceFeature.links != null
774 && sequenceFeature.getDescription().indexOf("<html>") == -1)
776 out.append("<html>");
779 out.append(sequenceFeature.description);
780 if (sequenceFeature.links != null)
782 for (int l = 0; l < sequenceFeature.links.size(); l++)
784 String label = sequenceFeature.links.elementAt(l);
785 String href = label.substring(label.indexOf("|") + 1);
786 label = label.substring(0, label.indexOf("|"));
788 if (sequenceFeature.description.indexOf(href) == -1)
790 out.append(" <a href=\"" + href + "\">" + label + "</a>");
794 if (sequenceFeature.getDescription().indexOf("</html>") == -1)
796 out.append("</html>");
802 out.append(sequenceName);
803 out.append("\t-1\t");
804 out.append(sequenceFeature.begin);
806 out.append(sequenceFeature.end);
808 out.append(sequenceFeature.type);
809 if (!Float.isNaN(sequenceFeature.score))
812 out.append(sequenceFeature.score);
816 return out.toString();
820 * Parse method that is called when a GFF file is dragged to the desktop
825 AlignViewportI av = getViewport();
828 if (av.getAlignment() != null)
830 dataset = av.getAlignment().getDataset();
834 // working in the applet context ?
835 dataset = av.getAlignment();
840 dataset = new Alignment(new SequenceI[] {});
843 Map<String, FeatureColourI> featureColours = new HashMap<>();
844 boolean parseResult = parse(dataset, featureColours, false, true);
847 // pass error up somehow
851 // update viewport with the dataset data ?
855 setSeqs(dataset.getSequencesArray());
860 * Implementation of unused abstract method
862 * @return error message
865 public String print(SequenceI[] sqs, boolean jvsuffix)
867 System.out.println("Use printGffFormat() or printJalviewFormat()");
872 * Returns features output in GFF2 format
875 * the sequences whose features are to be output
877 * a map whose keys are the type names of visible features
878 * @param visibleFeatureGroups
879 * @param includeNonPositionalFeatures
882 public String printGffFormat(SequenceI[] sequences,
883 FeatureRenderer fr, boolean includeNonPositionalFeatures)
885 Map<String, FeatureColourI> visibleColours = fr.getDisplayedFeatureCols();
887 StringBuilder out = new StringBuilder(256);
889 out.append(String.format("%s %d\n", GFF_VERSION, gffVersion == 0 ? 2 : gffVersion));
891 if (!includeNonPositionalFeatures
892 && (visibleColours == null || visibleColours.isEmpty()))
894 return out.toString();
897 String[] types = visibleColours == null ? new String[0]
898 : visibleColours.keySet()
899 .toArray(new String[visibleColours.keySet().size()]);
901 for (SequenceI seq : sequences)
903 List<SequenceFeature> features = new ArrayList<>();
904 if (includeNonPositionalFeatures)
906 features.addAll(seq.getFeatures().getNonPositionalFeatures());
908 if (visibleColours != null && !visibleColours.isEmpty())
910 features.addAll(seq.getFeatures().getPositionalFeatures(types));
913 for (SequenceFeature sf : features)
915 if (!sf.isNonPositional() && !fr.isVisible(sf))
918 * feature hidden by group visibility, colour threshold,
919 * or feature filter condition
924 String source = sf.featureGroup;
927 source = sf.getDescription();
930 out.append(seq.getName());
936 out.append(sf.begin);
940 out.append(sf.score);
943 int strand = sf.getStrand();
944 out.append(strand == 1 ? "+" : (strand == -1 ? "-" : "."));
947 String phase = sf.getPhase();
948 out.append(phase == null ? "." : phase);
950 // miscellaneous key-values (GFF column 9)
951 String attributes = sf.getAttributes();
952 if (attributes != null)
954 out.append(TAB).append(attributes);
961 return out.toString();
965 * Returns a mapping given list of one or more Align descriptors (exonerate
968 * @param alignedRegions
969 * a list of "Align fromStart toStart fromCount"
970 * @param mapIsFromCdna
971 * if true, 'from' is dna, else 'from' is protein
973 * either 1 (forward) or -1 (reverse)
975 * @throws IOException
977 protected MapList constructCodonMappingFromAlign(
978 List<String> alignedRegions, boolean mapIsFromCdna, int strand)
983 throw new IOException(
984 "Invalid strand for a codon mapping (cannot be 0)");
986 int regions = alignedRegions.size();
987 // arrays to hold [start, end] for each aligned region
988 int[] fromRanges = new int[regions * 2]; // from dna
989 int[] toRanges = new int[regions * 2]; // to protein
990 int fromRangesIndex = 0;
991 int toRangesIndex = 0;
993 for (String range : alignedRegions)
996 * Align mapFromStart mapToStart mapFromCount
997 * e.g. if mapIsFromCdna
998 * Align 11270 143 120
1000 * 120 bases from pos 11270 align to pos 143 in peptide
1001 * if !mapIsFromCdna this would instead be
1002 * Align 143 11270 40
1004 String[] tokens = range.split(" ");
1005 if (tokens.length != 3)
1007 throw new IOException("Wrong number of fields for Align");
1014 fromStart = Integer.parseInt(tokens[0]);
1015 toStart = Integer.parseInt(tokens[1]);
1016 fromCount = Integer.parseInt(tokens[2]);
1017 } catch (NumberFormatException nfe)
1019 throw new IOException(
1020 "Invalid number in Align field: " + nfe.getMessage());
1024 * Jalview always models from dna to protein, so adjust values if the
1025 * GFF mapping is from protein to dna
1030 int temp = fromStart;
1031 fromStart = toStart;
1034 fromRanges[fromRangesIndex++] = fromStart;
1035 fromRanges[fromRangesIndex++] = fromStart + strand * (fromCount - 1);
1038 * If a codon has an intron gap, there will be contiguous 'toRanges';
1039 * this is handled for us by the MapList constructor.
1040 * (It is not clear that exonerate ever generates this case)
1042 toRanges[toRangesIndex++] = toStart;
1043 toRanges[toRangesIndex++] = toStart + (fromCount - 1) / 3;
1046 return new MapList(fromRanges, toRanges, 3, 1);
1050 * Parse a GFF format feature. This may include creating a 'dummy' sequence to
1051 * hold the feature, or for its mapped sequence, or both, to be resolved
1052 * either later in the GFF file (##FASTA section), or when the user loads
1053 * additional sequences.
1057 * @param relaxedIdMatching
1061 protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment,
1062 boolean relaxedIdMatching, List<SequenceI> newseqs)
1065 * GFF: seqid source type start end score strand phase [attributes]
1067 if (gffColumns.length < 5)
1069 System.err.println("Ignoring GFF feature line with too few columns ("
1070 + gffColumns.length + ")");
1075 * locate referenced sequence in alignment _or_
1076 * as a forward or external reference (SequenceDummy)
1078 String seqId = gffColumns[0];
1079 SequenceI seq = findSequence(seqId, alignment, newseqs,
1082 SequenceFeature sf = null;
1083 GffHelperI helper = GffHelperFactory.getHelper(gffColumns);
1088 sf = helper.processGff(seq, gffColumns, alignment, newseqs,
1092 seq.addSequenceFeature(sf);
1093 while ((seq = alignment.findName(seq, seqId, true)) != null)
1095 seq.addSequenceFeature(new SequenceFeature(sf));
1098 } catch (IOException e)
1100 System.err.println("GFF parsing failed with: " + e.getMessage());
1109 * Process the 'column 9' data of the GFF file. This is less formally defined,
1110 * and its interpretation will vary depending on the tool that has generated
1116 protected void processGffColumnNine(String attributes, SequenceFeature sf)
1118 sf.setAttributes(attributes);
1121 * Parse attributes in column 9 and add them to the sequence feature's
1122 * 'otherData' table; use Note as a best proxy for description
1124 char nameValueSeparator = gffVersion == 3 ? '=' : ' ';
1125 // TODO check we don't break GFF2 values which include commas here
1126 Map<String, List<String>> nameValues = GffHelperBase
1127 .parseNameValuePairs(attributes, ";", nameValueSeparator, ",");
1128 for (Entry<String, List<String>> attr : nameValues.entrySet())
1130 String values = StringUtils.listToDelimitedString(attr.getValue(),
1132 sf.setValue(attr.getKey(), values);
1133 if (NOTE.equals(attr.getKey()))
1135 sf.setDescription(values);
1141 * After encountering ##fasta in a GFF3 file, process the remainder of the
1142 * file as FAST sequence data. Any placeholder sequences created during
1143 * feature parsing are updated with the actual sequences.
1147 * @throws IOException
1149 protected void processAsFasta(AlignmentI align, List<SequenceI> newseqs)
1155 } catch (IOException q)
1158 FastaFile parser = new FastaFile(this);
1159 List<SequenceI> includedseqs = parser.getSeqs();
1161 SequenceIdMatcher smatcher = new SequenceIdMatcher(newseqs);
1164 * iterate over includedseqs, and replacing matching ones with newseqs
1165 * sequences. Generic iterator not used here because we modify
1166 * includedseqs as we go
1168 for (int p = 0, pSize = includedseqs.size(); p < pSize; p++)
1170 // search for any dummy seqs that this sequence can be used to update
1171 SequenceI includedSeq = includedseqs.get(p);
1172 SequenceI dummyseq = smatcher.findIdMatch(includedSeq);
1173 if (dummyseq != null && dummyseq instanceof SequenceDummy)
1175 // probably have the pattern wrong
1176 // idea is that a flyweight proxy for a sequence ID can be created for
1177 // 1. stable reference creation
1178 // 2. addition of annotation
1179 // 3. future replacement by a real sequence
1180 // current pattern is to create SequenceDummy objects - a convenience
1181 // constructor for a Sequence.
1182 // problem is that when promoted to a real sequence, all references
1183 // need to be updated somehow. We avoid that by keeping the same object.
1184 ((SequenceDummy) dummyseq).become(includedSeq);
1185 dummyseq.createDatasetSequence();
1188 * Update mappings so they are now to the dataset sequence
1190 for (AlignedCodonFrame mapping : align.getCodonFrames())
1192 mapping.updateToDataset(dummyseq);
1196 * replace parsed sequence with the realised forward reference
1198 includedseqs.set(p, dummyseq);
1201 * and remove from the newseqs list
1203 newseqs.remove(dummyseq);
1208 * finally add sequences to the dataset
1210 for (SequenceI seq : includedseqs)
1212 // experimental: mapping-based 'alignment' to query sequence
1213 AlignmentUtils.alignSequenceAs(seq, align,
1214 String.valueOf(align.getGapCharacter()), false, true);
1216 // rename sequences if GFF handler requested this
1217 // TODO a more elegant way e.g. gffHelper.postProcess(newseqs) ?
1218 List<SequenceFeature> sfs = seq.getFeatures().getPositionalFeatures();
1221 String newName = (String) sfs.get(0).getValue(
1222 GffHelperI.RENAME_TOKEN);
1223 if (newName != null)
1225 seq.setName(newName);
1228 align.addSequence(seq);
1233 * Process a ## directive
1239 * @throws IOException
1241 protected void processGffPragma(String line, Map<String, String> gffProps,
1242 AlignmentI align, List<SequenceI> newseqs) throws IOException
1245 if ("###".equals(line))
1247 // close off any open 'forward references'
1251 String[] tokens = line.substring(2).split(" ");
1252 String pragma = tokens[0];
1253 String value = tokens.length == 1 ? null : tokens[1];
1255 if ("gff-version".equalsIgnoreCase(pragma))
1261 // value may be e.g. "3.1.2"
1262 gffVersion = Integer.parseInt(value.split("\\.")[0]);
1263 } catch (NumberFormatException e)
1269 else if ("sequence-region".equalsIgnoreCase(pragma))
1271 // could capture <seqid start end> if wanted here
1273 else if ("feature-ontology".equalsIgnoreCase(pragma))
1275 // should resolve against the specified feature ontology URI
1277 else if ("attribute-ontology".equalsIgnoreCase(pragma))
1279 // URI of attribute ontology - not currently used in GFF3
1281 else if ("source-ontology".equalsIgnoreCase(pragma))
1283 // URI of source ontology - not currently used in GFF3
1285 else if ("species-build".equalsIgnoreCase(pragma))
1287 // save URI of specific NCBI taxon version of annotations
1288 gffProps.put("species-build", value);
1290 else if ("fasta".equalsIgnoreCase(pragma))
1292 // process the rest of the file as a fasta file and replace any dummy
1294 processAsFasta(align, newseqs);
1298 System.err.println("Ignoring unknown pragma: " + line);