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.
21 package jalview.analysis;
23 import jalview.analysis.scoremodels.PIDModel;
24 import jalview.analysis.scoremodels.ScoreMatrix;
25 import jalview.analysis.scoremodels.ScoreModels;
26 import jalview.analysis.scoremodels.SimilarityParams;
27 import jalview.datamodel.AlignmentAnnotation;
28 import jalview.datamodel.AlignmentI;
29 import jalview.datamodel.Mapping;
30 import jalview.datamodel.Sequence;
31 import jalview.datamodel.SequenceI;
32 import jalview.util.Comparison;
33 import jalview.util.Format;
34 import jalview.util.MapList;
35 import jalview.util.MessageManager;
37 import java.awt.Color;
38 import java.awt.Graphics;
39 import java.util.ArrayList;
40 import java.util.Arrays;
41 import java.util.List;
42 import java.util.StringTokenizer;
52 public static final String PEP = "pep";
54 public static final String DNA = "dna";
56 private static final String NEWLINE = System.lineSeparator();
86 public String astr1 = "";
88 public String astr2 = "";
105 public float maxscore;
119 StringBuffer output = new StringBuffer();
121 String type; // AlignSeq.PEP or AlignSeq.DNA
123 private ScoreMatrix scoreModel;
126 * Creates a new AlignSeq object.
128 * @param s1 first sequence for alignment
129 * @param s2 second sequence for alignment
130 * @param type molecule type, either AlignSeq.PEP or AlignSeq.DNA
132 public AlignSeq(SequenceI s1, SequenceI s2, String type)
134 seqInit(s1, s1.getSequenceAsString(), s2, s2.getSequenceAsString(),
139 * Creates a new AlignSeq object.
148 public AlignSeq(SequenceI s1, String string1, SequenceI s2,
149 String string2, String type)
151 seqInit(s1, string1.toUpperCase(), s2, string2.toUpperCase(), type);
157 * @return DOCUMENT ME!
159 public float getMaxScore()
167 * @return DOCUMENT ME!
169 public int getSeq2Start()
177 * @return DOCUMENT ME!
179 public int getSeq2End()
187 * @return DOCUMENT ME!
189 public int getSeq1Start()
197 * @return DOCUMENT ME!
199 public int getSeq1End()
207 * @return DOCUMENT ME!
209 public String getOutput()
211 return output.toString();
217 * @return DOCUMENT ME!
219 public String getAStr1()
227 * @return DOCUMENT ME!
229 public String getAStr2()
237 * @return DOCUMENT ME!
239 public int[] getASeq1()
247 * @return DOCUMENT ME!
249 public int[] getASeq2()
256 * @return aligned instance of Seq 1
258 public SequenceI getAlignedSeq1()
260 SequenceI alSeq1 = new Sequence(s1.getName(), getAStr1());
261 alSeq1.setStart(s1.getStart() + getSeq1Start() - 1);
262 alSeq1.setEnd(s1.getStart() + getSeq1End() - 1);
263 alSeq1.setDatasetSequence(s1.getDatasetSequence() == null ? s1 : s1
264 .getDatasetSequence());
270 * @return aligned instance of Seq 2
272 public SequenceI getAlignedSeq2()
274 SequenceI alSeq2 = new Sequence(s2.getName(), getAStr2());
275 alSeq2.setStart(s2.getStart() + getSeq2Start() - 1);
276 alSeq2.setEnd(s2.getStart() + getSeq2End() - 1);
277 alSeq2.setDatasetSequence(s2.getDatasetSequence() == null ? s2 : s2
278 .getDatasetSequence());
283 * Construct score matrix for sequences with standard DNA or PEPTIDE matrix
288 * - string to use for s1
292 * - string to use for s2
296 public void seqInit(SequenceI s1, String string1, SequenceI s2,
297 String string2, String type)
301 setDefaultParams(type);
302 seqInit(string1, string2);
306 * construct score matrix for string1 and string2 (after removing any existing
312 private void seqInit(String string1, String string2)
314 s1str = extractGaps(jalview.util.Comparison.GapChars, string1);
315 s2str = extractGaps(jalview.util.Comparison.GapChars, string2);
317 if (s1str.length() == 0 || s2str.length() == 0)
319 output.append("ALL GAPS: "
320 + (s1str.length() == 0 ? s1.getName() : " ")
321 + (s2str.length() == 0 ? s2.getName() : ""));
325 seq1 = new int[s1str.length()];
327 seq2 = new int[s2str.length()];
329 score = new float[s1str.length()][s2str.length()];
331 E = new float[s1str.length()][s2str.length()];
333 F = new float[s1str.length()][s2str.length()];
334 traceback = new int[s1str.length()][s2str.length()];
336 seq1 = indexEncode(s1str);
338 seq2 = indexEncode(s2str);
341 private void setDefaultParams(String moleculeType)
343 if (!PEP.equals(moleculeType) && !DNA.equals(moleculeType))
345 output.append("Wrong type = dna or pep only");
346 throw new Error(MessageManager.formatMessage(
347 "error.unknown_type_dna_or_pep",
348 new String[] { moleculeType }));
352 scoreModel = ScoreModels.getInstance().getDefaultModel(
354 lookup = scoreModel.getMatrix();
355 gapIndex = scoreModel.getMatrixIndex(' ');
361 public void traceAlignment()
363 // Find the maximum score along the rhs or bottom row
364 float max = -Float.MAX_VALUE;
366 for (int i = 0; i < seq1.length; i++)
368 if (score[i][seq2.length - 1] > max)
370 max = score[i][seq2.length - 1];
372 maxj = seq2.length - 1;
376 for (int j = 0; j < seq2.length; j++)
378 if (score[seq1.length - 1][j] > max)
380 max = score[seq1.length - 1][j];
381 maxi = seq1.length - 1;
386 // System.out.println(maxi + " " + maxj + " " + score[maxi][maxj]);
390 maxscore = score[i][j] / 10;
395 aseq1 = new int[seq1.length + seq2.length];
396 aseq2 = new int[seq1.length + seq2.length];
398 StringBuilder sb1 = new StringBuilder(aseq1.length);
399 StringBuilder sb2 = new StringBuilder(aseq2.length);
401 count = (seq1.length + seq2.length) - 1;
403 while (i > 0 && j > 0)
405 aseq1[count] = seq1[i];
406 sb1.append(s1str.charAt(i));
407 aseq2[count] = seq2[j];
408 sb2.append(s2str.charAt(j));
410 trace = findTrace(i, j);
420 aseq1[count] = gapIndex;
421 sb1.replace(sb1.length() - 1, sb1.length(), "-");
423 else if (trace == -1)
426 aseq2[count] = gapIndex;
427 sb2.replace(sb2.length() - 1, sb2.length(), "-");
436 if (aseq1[count] != gapIndex)
438 aseq1[count] = seq1[i];
439 sb1.append(s1str.charAt(i));
442 if (aseq2[count] != gapIndex)
444 aseq2[count] = seq2[j];
445 sb2.append(s2str.charAt(j));
449 * we built the character strings backwards, so now
450 * reverse them to convert to sequence strings
452 astr1 = sb1.reverse().toString();
453 astr2 = sb2.reverse().toString();
459 public void printAlignment(java.io.PrintStream os)
461 // TODO: Use original sequence characters rather than re-translated
462 // characters in output
463 // Find the biggest id length for formatting purposes
464 String s1id = s1.getName(), s2id = s2.getName();
465 int maxid = s1.getName().length();
466 if (s2.getName().length() > maxid)
468 maxid = s2.getName().length();
473 // JAL-527 - truncate the sequence ids
474 if (s1.getName().length() > maxid)
476 s1id = s1.getName().substring(0, 30);
478 if (s2.getName().length() > maxid)
480 s2id = s2.getName().substring(0, 30);
483 int len = 72 - maxid - 1;
484 int nochunks = ((aseq1.length - count) / len)
485 + ((aseq1.length - count) % len > 0 ? 1 : 0);
488 output.append("Score = ").append(score[maxi][maxj]).append(NEWLINE);
489 output.append("Length of alignment = ")
490 .append(String.valueOf(aseq1.length - count)).append(NEWLINE);
491 output.append("Sequence ");
492 output.append(new Format("%" + maxid + "s").form(s1.getName()));
493 output.append(" : ").append(String.valueOf(s1.getStart()))
494 .append(" - ").append(String.valueOf(s1.getEnd()));
495 output.append(" (Sequence length = ")
496 .append(String.valueOf(s1str.length())).append(")")
498 output.append("Sequence ");
499 output.append(new Format("%" + maxid + "s").form(s2.getName()));
500 output.append(" : ").append(String.valueOf(s2.getStart()))
501 .append(" - ").append(String.valueOf(s2.getEnd()));
502 output.append(" (Sequence length = ")
503 .append(String.valueOf(s2str.length())).append(")")
504 .append(NEWLINE).append(NEWLINE);
506 ScoreMatrix pam250 = ScoreModels.getInstance().getPam250();
508 for (int j = 0; j < nochunks; j++)
510 // Print the first aligned sequence
511 output.append(new Format("%" + (maxid) + "s").form(s1id)).append(" ");
513 for (int i = 0; i < len; i++)
515 if ((i + (j * len)) < astr1.length())
517 output.append(astr1.charAt(i + (j * len)));
521 output.append(NEWLINE);
522 output.append(new Format("%" + (maxid) + "s").form(" ")).append(" ");
525 * Print out the match symbols:
526 * | for exact match (ignoring case)
527 * . if PAM250 score is positive
530 for (int i = 0; i < len; i++)
532 if ((i + (j * len)) < astr1.length())
534 char c1 = astr1.charAt(i + (j * len));
535 char c2 = astr2.charAt(i + (j * len));
536 boolean sameChar = Comparison.isSameResidue(c1, c2, false);
537 if (sameChar && !Comparison.isGap(c1))
542 else if (type.equals("pep"))
544 if (pam250.getPairwiseScore(c1, c2) > 0)
560 // Now print the second aligned sequence
561 output = output.append(NEWLINE);
562 output = output.append(new Format("%" + (maxid) + "s").form(s2id))
565 for (int i = 0; i < len; i++)
567 if ((i + (j * len)) < astr2.length())
569 output.append(astr2.charAt(i + (j * len)));
573 output.append(NEWLINE).append(NEWLINE);
576 pid = pid / (aseq1.length - count) * 100;
577 output = output.append(new Format("Percentage ID = %2.2f\n").form(pid));
580 os.print(output.toString());
581 } catch (Exception ex)
594 * @return DOCUMENT ME!
596 public int findTrace(int i, int j)
599 float max = score[i - 1][j - 1] + (lookup[seq1[i]][seq2[j]] * 10);
606 else if (F[i][j] == max)
620 else if (E[i][j] == max)
637 public void calcScoreMatrix()
642 // top left hand element
643 score[0][0] = lookup[seq1[0]][seq2[0]] * 10;
644 E[0][0] = -gapExtend;
647 // Calculate the top row first
648 for (int j = 1; j < m; j++)
650 // What should these values be? 0 maybe
651 E[0][j] = max(score[0][j - 1] - gapOpen, E[0][j - 1] - gapExtend);
652 F[0][j] = -gapExtend;
654 score[0][j] = max(lookup[seq1[0]][seq2[j]] * 10, -gapOpen, -gapExtend);
659 // Now do the left hand column
660 for (int i = 1; i < n; i++)
663 F[i][0] = max(score[i - 1][0] - gapOpen, F[i - 1][0] - gapExtend);
665 score[i][0] = max(lookup[seq1[i]][seq2[0]] * 10, E[i][0], F[i][0]);
666 traceback[i][0] = -1;
669 // Now do all the other rows
670 for (int i = 1; i < n; i++)
672 for (int j = 1; j < m; j++)
674 E[i][j] = max(score[i][j - 1] - gapOpen, E[i][j - 1] - gapExtend);
675 F[i][j] = max(score[i - 1][j] - gapOpen, F[i - 1][j] - gapExtend);
677 score[i][j] = max(score[i - 1][j - 1]
678 + (lookup[seq1[i]][seq2[j]] * 10), E[i][j], F[i][j]);
679 traceback[i][j] = findTrace(i, j);
685 * Returns the given sequence with all of the given gap characters removed.
688 * a string of characters to be treated as gaps
694 public static String extractGaps(String gapChars, String seq)
696 if (gapChars == null || seq == null)
700 StringTokenizer str = new StringTokenizer(seq, gapChars);
701 StringBuilder newString = new StringBuilder(seq.length());
703 while (str.hasMoreTokens())
705 newString.append(str.nextToken());
708 return newString.toString();
721 * @return DOCUMENT ME!
723 private static float max(float f1, float f2, float f3)
748 * @return DOCUMENT ME!
750 private static float max(float f1, float f2)
763 * Converts the character string to an array of integers which are the
764 * corresponding indices to the characters in the score matrix
770 int[] indexEncode(String s)
772 int[] encoded = new int[s.length()];
774 for (int i = 0; i < s.length(); i++)
776 char c = s.charAt(i);
777 encoded[i] = scoreModel.getMatrixIndex(c);
797 public static void displayMatrix(Graphics g, int[][] mat, int n, int m,
800 // TODO method dosen't seem to be referenced anywhere delete??
804 for (int i = 0; i < n; i++)
806 for (int j = 0; j < m; j++)
808 if (mat[i][j] >= max)
813 if (mat[i][j] <= min)
820 System.out.println(max + " " + min);
822 for (int i = 0; i < n; i++)
824 for (int j = 0; j < m; j++)
829 // System.out.println(mat[i][j]);
830 float score = (float) (mat[i][j] - min) / (float) (max - min);
831 g.setColor(new Color(score, 0, 0));
832 g.fillRect(x, y, psize, psize);
834 // System.out.println(x + " " + y + " " + score);
840 * Compute a globally optimal needleman and wunsch alignment between two
846 * AlignSeq.DNA or AlignSeq.PEP
848 public static AlignSeq doGlobalNWAlignment(SequenceI s1, SequenceI s2,
851 AlignSeq as = new AlignSeq(s1, s2, type);
853 as.calcScoreMatrix();
860 * @return mapping from positions in S1 to corresponding positions in S2
862 public jalview.datamodel.Mapping getMappingFromS1(boolean allowmismatch)
864 ArrayList<Integer> as1 = new ArrayList<Integer>(), as2 = new ArrayList<Integer>();
865 int pdbpos = s2.getStart() + getSeq2Start() - 2;
866 int alignpos = s1.getStart() + getSeq1Start() - 2;
867 int lp2 = pdbpos - 3, lp1 = alignpos - 3;
868 boolean lastmatch = false;
869 // and now trace the alignment onto the atom set.
870 for (int i = 0; i < astr1.length(); i++)
872 char c1 = astr1.charAt(i), c2 = astr2.charAt(i);
883 if (allowmismatch || c1 == c2)
885 // extend mapping interval
886 if (lp1 + 1 != alignpos || lp2 + 1 != pdbpos)
888 as1.add(Integer.valueOf(alignpos));
889 as2.add(Integer.valueOf(pdbpos));
897 // extend mapping interval
900 as1.add(Integer.valueOf(lp1));
901 as2.add(Integer.valueOf(lp2));
906 // construct range pairs
908 int[] mapseq1 = new int[as1.size() + (lastmatch ? 1 : 0)], mapseq2 = new int[as2
909 .size() + (lastmatch ? 1 : 0)];
911 for (Integer ip : as1)
917 for (Integer ip : as2)
924 mapseq1[mapseq1.length - 1] = alignpos;
925 mapseq2[mapseq2.length - 1] = pdbpos;
927 MapList map = new MapList(mapseq1, mapseq2, 1, 1);
929 jalview.datamodel.Mapping mapping = new Mapping(map);
935 * matches ochains against al and populates seqs with the best match between
936 * each ochain and the set in al
940 * @param dnaOrProtein
941 * @param removeOldAnnots
942 * when true, old annotation is cleared before new annotation
944 * @return List<List<SequenceI> originals, List<SequenceI> replacement,
945 * List<AlignSeq> alignment between each>
947 public static List<List<? extends Object>> replaceMatchingSeqsWith(
948 List<SequenceI> seqs, List<AlignmentAnnotation> annotations,
949 List<SequenceI> ochains, AlignmentI al, String dnaOrProtein,
950 boolean removeOldAnnots)
952 List<SequenceI> orig = new ArrayList<SequenceI>(), repl = new ArrayList<SequenceI>();
953 List<AlignSeq> aligs = new ArrayList<AlignSeq>();
954 if (al != null && al.getHeight() > 0)
956 ArrayList<SequenceI> matches = new ArrayList<SequenceI>();
957 ArrayList<AlignSeq> aligns = new ArrayList<AlignSeq>();
959 for (SequenceI sq : ochains)
961 SequenceI bestm = null;
962 AlignSeq bestaseq = null;
964 for (SequenceI msq : al.getSequences())
966 AlignSeq aseq = doGlobalNWAlignment(msq, sq, dnaOrProtein);
967 if (bestm == null || aseq.getMaxScore() > bestscore)
969 bestscore = aseq.getMaxScore();
974 System.out.println("Best Score for " + (matches.size() + 1) + " :"
977 aligns.add(bestaseq);
978 al.deleteSequence(bestm);
980 for (int p = 0, pSize = seqs.size(); p < pSize; p++)
982 SequenceI sq, sp = seqs.get(p);
984 if ((q = ochains.indexOf(sp)) > -1)
986 seqs.set(p, sq = matches.get(q));
989 sq.setName(sp.getName());
990 sq.setDescription(sp.getDescription());
992 sq.transferAnnotation(sp,
993 sp2sq = aligns.get(q).getMappingFromS1(false));
994 aligs.add(aligns.get(q));
996 for (int ap = 0; ap < annotations.size();)
998 if (annotations.get(ap).sequenceRef == sp)
1004 if (removeOldAnnots)
1006 annotations.remove(ap);
1010 AlignmentAnnotation alan = annotations.remove(ap);
1011 alan.liftOver(sq, sp2sq);
1012 alan.setSequenceRef(sq);
1013 sq.addAlignmentAnnotation(alan);
1021 if (sq.getAnnotation() != null && sq.getAnnotation().length > 0)
1023 annotations.addAll(inspos == -1 ? annotations.size() : inspos,
1024 Arrays.asList(sq.getAnnotation()));
1029 return Arrays.asList(orig, repl, aligs);
1033 * compute the PID vector used by the redundancy filter.
1035 * @param originalSequences
1036 * - sequences in alignment that are to filtered
1038 * - null or strings to be analysed (typically, visible portion of
1039 * each sequence in alignment)
1041 * - first column in window for calculation
1043 * - last column in window for calculation
1045 * - if true then use ungapped sequence to compute PID
1046 * @return vector containing maximum PID for i-th sequence and any sequences
1047 * longer than that seuqence
1049 public static float[] computeRedundancyMatrix(
1050 SequenceI[] originalSequences, String[] omitHidden, int start,
1051 int end, boolean ungapped)
1053 int height = originalSequences.length;
1054 float[] redundancy = new float[height];
1055 int[] lngth = new int[height];
1056 for (int i = 0; i < height; i++)
1062 // long start = System.currentTimeMillis();
1064 SimilarityParams pidParams = new SimilarityParams(true, true, true,
1068 for (int i = 0; i < height; i++)
1071 for (int j = 0; j < i; j++)
1078 if (omitHidden == null)
1080 seqi = originalSequences[i].getSequenceAsString(start, end);
1081 seqj = originalSequences[j].getSequenceAsString(start, end);
1085 seqi = omitHidden[i];
1086 seqj = omitHidden[j];
1090 String ug = AlignSeq.extractGaps(Comparison.GapChars, seqi);
1091 lngth[i] = ug.length();
1099 String ug = AlignSeq.extractGaps(Comparison.GapChars, seqj);
1100 lngth[j] = ug.length();
1106 pid = (float) PIDModel.computePID(seqi, seqj, pidParams);
1108 // use real sequence length rather than string length
1109 if (lngth[j] < lngth[i])
1111 redundancy[j] = Math.max(pid, redundancy[j]);
1115 redundancy[i] = Math.max(pid, redundancy[i]);