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.ScoreMatrix;
24 import jalview.analysis.scoremodels.ScoreModels;
25 import jalview.datamodel.AlignmentAnnotation;
26 import jalview.datamodel.AlignmentI;
27 import jalview.datamodel.Mapping;
28 import jalview.datamodel.Sequence;
29 import jalview.datamodel.SequenceI;
30 import jalview.schemes.ResidueProperties;
31 import jalview.util.Comparison;
32 import jalview.util.Format;
33 import jalview.util.MapList;
34 import jalview.util.MessageManager;
36 import java.awt.Color;
37 import java.awt.Graphics;
38 import java.util.ArrayList;
39 import java.util.Arrays;
40 import java.util.List;
41 import java.util.StringTokenizer;
51 public static final String PEP = "pep";
53 public static final String DNA = "dna";
55 private static final String NEWLINE = System.lineSeparator();
57 static String[] dna = { "A", "C", "G", "T", "-" };
59 // "C", "T", "A", "G", "-"};
60 static String[] pep = { "A", "R", "N", "D", "C", "Q", "E", "G", "H", "I",
61 "L", "K", "M", "F", "P", "S", "T", "W", "Y", "V", "B", "Z", "X", "-" };
91 public String astr1 = "";
93 public String astr2 = "";
102 public int seq2start;
110 public float maxscore;
120 float[][] lookup = ResidueProperties.getBLOSUM62();
122 String[] intToStr = pep;
126 StringBuffer output = new StringBuffer();
130 private int[] charToInt;
133 * Creates a new AlignSeq object.
142 public AlignSeq(SequenceI s1, SequenceI s2, String type)
144 SeqInit(s1, s1.getSequenceAsString(), s2, s2.getSequenceAsString(),
149 * Creates a new AlignSeq object.
158 public AlignSeq(SequenceI s1, String string1, SequenceI s2,
159 String string2, String type)
161 SeqInit(s1, string1.toUpperCase(), s2, string2.toUpperCase(), type);
167 * @return DOCUMENT ME!
169 public float getMaxScore()
177 * @return DOCUMENT ME!
179 public int getSeq2Start()
187 * @return DOCUMENT ME!
189 public int getSeq2End()
197 * @return DOCUMENT ME!
199 public int getSeq1Start()
207 * @return DOCUMENT ME!
209 public int getSeq1End()
217 * @return DOCUMENT ME!
219 public String getOutput()
221 return output.toString();
227 * @return DOCUMENT ME!
229 public String getAStr1()
237 * @return DOCUMENT ME!
239 public String getAStr2()
247 * @return DOCUMENT ME!
249 public int[] getASeq1()
257 * @return DOCUMENT ME!
259 public int[] getASeq2()
267 * @return DOCUMENT ME!
269 public SequenceI getS1()
277 * @return DOCUMENT ME!
279 public SequenceI getS2()
286 * @return aligned instance of Seq 1
288 public SequenceI getAlignedSeq1()
290 SequenceI alSeq1 = new Sequence(s1.getName(), getAStr1());
291 alSeq1.setStart(s1.getStart() + getSeq1Start() - 1);
292 alSeq1.setEnd(s1.getStart() + getSeq1End() - 1);
293 alSeq1.setDatasetSequence(s1.getDatasetSequence() == null ? s1 : s1
294 .getDatasetSequence());
300 * @return aligned instance of Seq 2
302 public SequenceI getAlignedSeq2()
304 SequenceI alSeq2 = new Sequence(s2.getName(), getAStr2());
305 alSeq2.setStart(s2.getStart() + getSeq2Start() - 1);
306 alSeq2.setEnd(s2.getStart() + getSeq2End() - 1);
307 alSeq2.setDatasetSequence(s2.getDatasetSequence() == null ? s2 : s2
308 .getDatasetSequence());
313 * Construct score matrix for sequences with standard DNA or PEPTIDE matrix
318 * - string to use for s1
322 * - string to use for s2
326 public void SeqInit(SequenceI s1, String string1, SequenceI s2,
327 String string2, String type)
331 setDefaultParams(type);
332 SeqInit(string1, string2);
336 * Construct score matrix for sequences with custom substitution matrix
341 * - string to use for s1
345 * - string to use for s2
347 * - substitution matrix to use for alignment
349 public void SeqInit(SequenceI s1, String string1, SequenceI s2,
350 String string2, ScoreMatrix scoreMatrix)
354 setType(scoreMatrix.isDNA() ? AlignSeq.DNA : AlignSeq.PEP);
355 lookup = scoreMatrix.getMatrix();
359 * construct score matrix for string1 and string2 (after removing any existing
365 private void SeqInit(String string1, String string2)
367 s1str = extractGaps(jalview.util.Comparison.GapChars, string1);
368 s2str = extractGaps(jalview.util.Comparison.GapChars, string2);
370 if (s1str.length() == 0 || s2str.length() == 0)
372 output.append("ALL GAPS: "
373 + (s1str.length() == 0 ? s1.getName() : " ")
374 + (s2str.length() == 0 ? s2.getName() : ""));
378 // System.out.println("lookuip " + rt.freeMemory() + " "+ rt.totalMemory());
379 seq1 = new int[s1str.length()];
381 // System.out.println("seq1 " + rt.freeMemory() +" " + rt.totalMemory());
382 seq2 = new int[s2str.length()];
384 // System.out.println("seq2 " + rt.freeMemory() + " " + rt.totalMemory());
385 score = new float[s1str.length()][s2str.length()];
387 // System.out.println("score " + rt.freeMemory() + " " + rt.totalMemory());
388 E = new float[s1str.length()][s2str.length()];
390 // System.out.println("E " + rt.freeMemory() + " " + rt.totalMemory());
391 F = new float[s1str.length()][s2str.length()];
392 traceback = new int[s1str.length()][s2str.length()];
394 // System.out.println("F " + rt.freeMemory() + " " + rt.totalMemory());
395 seq1 = stringToInt(s1str, type);
397 // System.out.println("seq1 " + rt.freeMemory() + " " + rt.totalMemory());
398 seq2 = stringToInt(s2str, type);
400 // System.out.println("Seq2 " + rt.freeMemory() + " " + rt.totalMemory());
401 // long tstart = System.currentTimeMillis();
402 // calcScoreMatrix();
403 // long tend = System.currentTimeMillis();
404 // System.out.println("Time take to calculate score matrix = " +
405 // (tend-tstart) + " ms");
406 // printScoreMatrix(score);
407 // System.out.println();
408 // printScoreMatrix(traceback);
409 // System.out.println();
410 // printScoreMatrix(E);
411 // System.out.println();
412 // /printScoreMatrix(F);
413 // System.out.println();
414 // tstart = System.currentTimeMillis();
416 // tend = System.currentTimeMillis();
417 // System.out.println("Time take to traceback alignment = " + (tend-tstart)
421 private void setDefaultParams(String type)
425 if (type.equals(AlignSeq.PEP))
427 lookup = ScoreModels.getInstance().getDefaultModel(true).getMatrix();
429 else if (type.equals(AlignSeq.DNA))
431 lookup = ScoreModels.getInstance().getDefaultModel(false).getMatrix();
435 private void setType(String type2)
438 if (type.equals(AlignSeq.PEP))
441 charToInt = ResidueProperties.aaIndex;
442 defInt = ResidueProperties.maxProteinIndex;
444 else if (type.equals(AlignSeq.DNA))
447 charToInt = ResidueProperties.nucleotideIndex;
448 defInt = ResidueProperties.maxNucleotideIndex;
452 output.append("Wrong type = dna or pep only");
453 throw new Error(MessageManager.formatMessage(
454 "error.unknown_type_dna_or_pep", new String[] { type2 }));
461 public void traceAlignment()
463 // Find the maximum score along the rhs or bottom row
466 for (int i = 0; i < seq1.length; i++)
468 if (score[i][seq2.length - 1] > max)
470 max = score[i][seq2.length - 1];
472 maxj = seq2.length - 1;
476 for (int j = 0; j < seq2.length; j++)
478 if (score[seq1.length - 1][j] > max)
480 max = score[seq1.length - 1][j];
481 maxi = seq1.length - 1;
486 // System.out.println(maxi + " " + maxj + " " + score[maxi][maxj]);
490 maxscore = score[i][j] / 10;
495 aseq1 = new int[seq1.length + seq2.length];
496 aseq2 = new int[seq1.length + seq2.length];
498 count = (seq1.length + seq2.length) - 1;
500 while ((i > 0) && (j > 0))
502 if ((aseq1[count] != defInt) && (i >= 0))
504 aseq1[count] = seq1[i];
505 astr1 = s1str.charAt(i) + astr1;
508 if ((aseq2[count] != defInt) && (j > 0))
510 aseq2[count] = seq2[j];
511 astr2 = s2str.charAt(j) + astr2;
514 trace = findTrace(i, j);
524 aseq1[count] = defInt;
525 astr1 = "-" + astr1.substring(1);
527 else if (trace == -1)
530 aseq2[count] = defInt;
531 astr2 = "-" + astr2.substring(1);
540 if (aseq1[count] != defInt)
542 aseq1[count] = seq1[i];
543 astr1 = s1str.charAt(i) + astr1;
546 if (aseq2[count] != defInt)
548 aseq2[count] = seq2[j];
549 astr2 = s2str.charAt(j) + astr2;
556 public void printAlignment(java.io.PrintStream os)
558 // TODO: Use original sequence characters rather than re-translated
559 // characters in output
560 // Find the biggest id length for formatting purposes
561 String s1id = s1.getName(), s2id = s2.getName();
562 int maxid = s1.getName().length();
563 if (s2.getName().length() > maxid)
565 maxid = s2.getName().length();
570 // JAL-527 - truncate the sequence ids
571 if (s1.getName().length() > maxid)
573 s1id = s1.getName().substring(0, 30);
575 if (s2.getName().length() > maxid)
577 s2id = s2.getName().substring(0, 30);
580 int len = 72 - maxid - 1;
581 int nochunks = ((aseq1.length - count) / len)
582 + ((aseq1.length - count) % len > 0 ? 1 : 0);
585 output.append("Score = ").append(score[maxi][maxj]).append(NEWLINE);
586 output.append("Length of alignment = ")
587 .append(String.valueOf(aseq1.length - count)).append(NEWLINE);
588 output.append("Sequence ");
589 output.append(new Format("%" + maxid + "s").form(s1.getName()));
590 output.append(" : ").append(String.valueOf(s1.getStart()))
591 .append(" - ").append(String.valueOf(s1.getEnd()));
592 output.append(" (Sequence length = ")
593 .append(String.valueOf(s1str.length())).append(")")
595 output.append("Sequence ");
596 output.append(new Format("%" + maxid + "s").form(s2.getName()));
597 output.append(" : ").append(String.valueOf(s2.getStart()))
598 .append(" - ").append(String.valueOf(s2.getEnd()));
599 output.append(" (Sequence length = ")
600 .append(String.valueOf(s2str.length())).append(")")
601 .append(NEWLINE).append(NEWLINE);
603 for (int j = 0; j < nochunks; j++)
605 // Print the first aligned sequence
606 output.append(new Format("%" + (maxid) + "s").form(s1id)).append(" ");
608 for (int i = 0; i < len; i++)
610 if ((i + (j * len)) < astr1.length())
612 output.append(astr1.charAt(i + (j * len)));
616 output.append(NEWLINE);
617 output.append(new Format("%" + (maxid) + "s").form(" ")).append(" ");
619 // Print out the matching chars
620 for (int i = 0; i < len; i++)
622 if ((i + (j * len)) < astr1.length())
624 boolean sameChar = Comparison.isSameResidue(
625 astr1.charAt(i + (j * len)), astr2.charAt(i + (j * len)),
628 && !jalview.util.Comparison.isGap(astr1.charAt(i
634 else if (type.equals("pep"))
636 if (ResidueProperties.getPAM250(astr1.charAt(i + (j * len)),
637 astr2.charAt(i + (j * len))) > 0)
653 // Now print the second aligned sequence
654 output = output.append(NEWLINE);
655 output = output.append(new Format("%" + (maxid) + "s").form(s2id))
658 for (int i = 0; i < len; i++)
660 if ((i + (j * len)) < astr2.length())
662 output.append(astr2.charAt(i + (j * len)));
666 output.append(NEWLINE).append(NEWLINE);
669 pid = pid / (aseq1.length - count) * 100;
670 output = output.append(new Format("Percentage ID = %2.2f\n").form(pid));
673 os.print(output.toString());
674 } catch (Exception ex)
685 public void printScoreMatrix(int[][] mat)
690 for (int i = 0; i < n; i++)
692 // Print the top sequence
695 Format.print(System.out, "%8s", s2str.substring(0, 1));
697 for (int jj = 1; jj < m; jj++)
699 Format.print(System.out, "%5s", s2str.substring(jj, jj + 1));
702 System.out.println();
705 for (int j = 0; j < m; j++)
709 Format.print(System.out, "%3s", s1str.substring(i, i + 1));
712 Format.print(System.out, "%3d ", mat[i][j] / 10);
715 System.out.println();
727 * @return DOCUMENT ME!
729 public int findTrace(int i, int j)
732 float max = score[i - 1][j - 1] + (lookup[seq1[i]][seq2[j]] * 10);
739 else if (F[i][j] == max)
753 else if (E[i][j] == max)
770 public void calcScoreMatrix()
775 // top left hand element
776 score[0][0] = lookup[seq1[0]][seq2[0]] * 10;
777 E[0][0] = -gapExtend;
780 // Calculate the top row first
781 for (int j = 1; j < m; j++)
783 // What should these values be? 0 maybe
784 E[0][j] = max(score[0][j - 1] - gapOpen, E[0][j - 1] - gapExtend);
785 F[0][j] = -gapExtend;
787 score[0][j] = max(lookup[seq1[0]][seq2[j]] * 10, -gapOpen, -gapExtend);
792 // Now do the left hand column
793 for (int i = 1; i < n; i++)
796 F[i][0] = max(score[i - 1][0] - gapOpen, F[i - 1][0] - gapExtend);
798 score[i][0] = max(lookup[seq1[i]][seq2[0]] * 10, E[i][0], F[i][0]);
799 traceback[i][0] = -1;
802 // Now do all the other rows
803 for (int i = 1; i < n; i++)
805 for (int j = 1; j < m; j++)
807 E[i][j] = max(score[i][j - 1] - gapOpen, E[i][j - 1] - gapExtend);
808 F[i][j] = max(score[i - 1][j] - gapOpen, F[i - 1][j] - gapExtend);
810 score[i][j] = max(score[i - 1][j - 1]
811 + (lookup[seq1[i]][seq2[j]] * 10), E[i][j], F[i][j]);
812 traceback[i][j] = findTrace(i, j);
818 * Returns the given sequence with all of the given gap characters removed.
821 * a string of characters to be treated as gaps
827 public static String extractGaps(String gapChars, String seq)
829 if (gapChars == null || seq == null)
833 StringTokenizer str = new StringTokenizer(seq, gapChars);
834 StringBuilder newString = new StringBuilder(seq.length());
836 while (str.hasMoreTokens())
838 newString.append(str.nextToken());
841 return newString.toString();
854 * @return DOCUMENT ME!
856 public float max(float f1, float f2, float f3)
881 * @return DOCUMENT ME!
883 public float max(float f1, float f2)
903 * @return DOCUMENT ME!
905 public int[] stringToInt(String s, String type)
907 int[] seq1 = new int[s.length()];
909 for (int i = 0; i < s.length(); i++)
911 // String ss = s.substring(i, i + 1).toUpperCase();
912 char c = s.charAt(i);
913 if ('a' <= c && c <= 'z')
921 seq1[i] = charToInt[c]; // set accordingly from setType
922 if (seq1[i] < 0 || seq1[i] > defInt) // set from setType: 23 for
923 // peptides, or 4 for NA.
928 } catch (Exception e)
951 public static void displayMatrix(Graphics g, int[][] mat, int n, int m,
954 // TODO method dosen't seem to be referenced anywhere delete??
958 for (int i = 0; i < n; i++)
960 for (int j = 0; j < m; j++)
962 if (mat[i][j] >= max)
967 if (mat[i][j] <= min)
974 System.out.println(max + " " + min);
976 for (int i = 0; i < n; i++)
978 for (int j = 0; j < m; j++)
983 // System.out.println(mat[i][j]);
984 float score = (float) (mat[i][j] - min) / (float) (max - min);
985 g.setColor(new Color(score, 0, 0));
986 g.fillRect(x, y, psize, psize);
988 // System.out.println(x + " " + y + " " + score);
994 * Compute a globally optimal needleman and wunsch alignment between two
1000 * AlignSeq.DNA or AlignSeq.PEP
1002 public static AlignSeq doGlobalNWAlignment(SequenceI s1, SequenceI s2,
1005 AlignSeq as = new AlignSeq(s1, s2, type);
1007 as.calcScoreMatrix();
1008 as.traceAlignment();
1014 * @return mapping from positions in S1 to corresponding positions in S2
1016 public jalview.datamodel.Mapping getMappingFromS1(boolean allowmismatch)
1018 ArrayList<Integer> as1 = new ArrayList<Integer>(), as2 = new ArrayList<Integer>();
1019 int pdbpos = s2.getStart() + getSeq2Start() - 2;
1020 int alignpos = s1.getStart() + getSeq1Start() - 2;
1021 int lp2 = pdbpos - 3, lp1 = alignpos - 3;
1022 boolean lastmatch = false;
1023 // and now trace the alignment onto the atom set.
1024 for (int i = 0; i < astr1.length(); i++)
1026 char c1 = astr1.charAt(i), c2 = astr2.charAt(i);
1037 if (allowmismatch || c1 == c2)
1039 // extend mapping interval
1040 if (lp1 + 1 != alignpos || lp2 + 1 != pdbpos)
1042 as1.add(Integer.valueOf(alignpos));
1043 as2.add(Integer.valueOf(pdbpos));
1051 // extend mapping interval
1054 as1.add(Integer.valueOf(lp1));
1055 as2.add(Integer.valueOf(lp2));
1060 // construct range pairs
1062 int[] mapseq1 = new int[as1.size() + (lastmatch ? 1 : 0)], mapseq2 = new int[as2
1063 .size() + (lastmatch ? 1 : 0)];
1065 for (Integer ip : as1)
1071 for (Integer ip : as2)
1078 mapseq1[mapseq1.length - 1] = alignpos;
1079 mapseq2[mapseq2.length - 1] = pdbpos;
1081 MapList map = new MapList(mapseq1, mapseq2, 1, 1);
1083 jalview.datamodel.Mapping mapping = new Mapping(map);
1089 * matches ochains against al and populates seqs with the best match between
1090 * each ochain and the set in al
1094 * @param dnaOrProtein
1095 * @param removeOldAnnots
1096 * when true, old annotation is cleared before new annotation
1098 * @return List<List<SequenceI> originals, List<SequenceI> replacement,
1099 * List<AlignSeq> alignment between each>
1101 public static List<List<? extends Object>> replaceMatchingSeqsWith(
1102 List<SequenceI> seqs, List<AlignmentAnnotation> annotations,
1103 List<SequenceI> ochains, AlignmentI al, String dnaOrProtein,
1104 boolean removeOldAnnots)
1106 List<SequenceI> orig = new ArrayList<SequenceI>(), repl = new ArrayList<SequenceI>();
1107 List<AlignSeq> aligs = new ArrayList<AlignSeq>();
1108 if (al != null && al.getHeight() > 0)
1110 ArrayList<SequenceI> matches = new ArrayList<SequenceI>();
1111 ArrayList<AlignSeq> aligns = new ArrayList<AlignSeq>();
1113 for (SequenceI sq : ochains)
1115 SequenceI bestm = null;
1116 AlignSeq bestaseq = null;
1117 float bestscore = 0;
1118 for (SequenceI msq : al.getSequences())
1120 AlignSeq aseq = doGlobalNWAlignment(msq, sq, dnaOrProtein);
1121 if (bestm == null || aseq.getMaxScore() > bestscore)
1123 bestscore = aseq.getMaxScore();
1128 System.out.println("Best Score for " + (matches.size() + 1) + " :"
1131 aligns.add(bestaseq);
1132 al.deleteSequence(bestm);
1134 for (int p = 0, pSize = seqs.size(); p < pSize; p++)
1136 SequenceI sq, sp = seqs.get(p);
1138 if ((q = ochains.indexOf(sp)) > -1)
1140 seqs.set(p, sq = matches.get(q));
1143 sq.setName(sp.getName());
1144 sq.setDescription(sp.getDescription());
1146 sq.transferAnnotation(sp,
1147 sp2sq = aligns.get(q).getMappingFromS1(false));
1148 aligs.add(aligns.get(q));
1150 for (int ap = 0; ap < annotations.size();)
1152 if (annotations.get(ap).sequenceRef == sp)
1158 if (removeOldAnnots)
1160 annotations.remove(ap);
1164 AlignmentAnnotation alan = annotations.remove(ap);
1165 alan.liftOver(sq, sp2sq);
1166 alan.setSequenceRef(sq);
1167 sq.addAlignmentAnnotation(alan);
1175 if (sq.getAnnotation() != null && sq.getAnnotation().length > 0)
1177 annotations.addAll(inspos == -1 ? annotations.size() : inspos,
1178 Arrays.asList(sq.getAnnotation()));
1183 return Arrays.asList(orig, repl, aligs);
1187 * compute the PID vector used by the redundancy filter.
1189 * @param originalSequences
1190 * - sequences in alignment that are to filtered
1192 * - null or strings to be analysed (typically, visible portion of
1193 * each sequence in alignment)
1195 * - first column in window for calculation
1197 * - last column in window for calculation
1199 * - if true then use ungapped sequence to compute PID
1200 * @return vector containing maximum PID for i-th sequence and any sequences
1201 * longer than that seuqence
1203 public static float[] computeRedundancyMatrix(
1204 SequenceI[] originalSequences, String[] omitHidden, int start,
1205 int end, boolean ungapped)
1207 int height = originalSequences.length;
1208 float[] redundancy = new float[height];
1209 int[] lngth = new int[height];
1210 for (int i = 0; i < height; i++)
1216 // long start = System.currentTimeMillis();
1220 for (int i = 0; i < height; i++)
1223 for (int j = 0; j < i; j++)
1230 if (omitHidden == null)
1232 seqi = originalSequences[i].getSequenceAsString(start, end);
1233 seqj = originalSequences[j].getSequenceAsString(start, end);
1237 seqi = omitHidden[i];
1238 seqj = omitHidden[j];
1242 String ug = AlignSeq.extractGaps(Comparison.GapChars, seqi);
1243 lngth[i] = ug.length();
1251 String ug = AlignSeq.extractGaps(Comparison.GapChars, seqj);
1252 lngth[j] = ug.length();
1258 pid = Comparison.PID(seqi, seqj);
1260 // use real sequence length rather than string length
1261 if (lngth[j] < lngth[i])
1263 redundancy[j] = Math.max(pid, redundancy[j]);
1267 redundancy[i] = Math.max(pid, redundancy[i]);