X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FAlignSeq.java;h=07f43daefc960838d56330469f1da80e52917cfa;hb=fd8dc2879b13709703b57dd7ed4b3754b253e12f;hp=ed4e3bce1f0427ea70d523d268d0d5cd3ab2811a;hpb=67428050d76de7946d96d43ba16ee02d439026bd;p=jalview.git diff --git a/src/jalview/analysis/AlignSeq.java b/src/jalview/analysis/AlignSeq.java index ed4e3bc..07f43da 100755 --- a/src/jalview/analysis/AlignSeq.java +++ b/src/jalview/analysis/AlignSeq.java @@ -20,15 +20,15 @@ */ package jalview.analysis; -import jalview.analysis.scoremodels.PairwiseSeqScoreModel; +import jalview.analysis.scoremodels.PIDModel; import jalview.analysis.scoremodels.ScoreMatrix; import jalview.analysis.scoremodels.ScoreModels; +import jalview.analysis.scoremodels.SimilarityParams; import jalview.datamodel.AlignmentAnnotation; import jalview.datamodel.AlignmentI; import jalview.datamodel.Mapping; import jalview.datamodel.Sequence; import jalview.datamodel.SequenceI; -import jalview.schemes.ResidueProperties; import jalview.util.Comparison; import jalview.util.Format; import jalview.util.MapList; @@ -55,12 +55,6 @@ public class AlignSeq private static final String NEWLINE = System.lineSeparator(); - static String[] dna = { "A", "C", "G", "T", "-" }; - - // "C", "T", "A", "G", "-"}; - static String[] pep = { "A", "R", "N", "D", "C", "Q", "E", "G", "H", "I", - "L", "K", "M", "F", "P", "S", "T", "W", "Y", "V", "B", "Z", "X", "-" }; - float[][] score; float[][] E; @@ -118,33 +112,24 @@ public class AlignSeq int gapExtend = 20; - float[][] lookup = ScoreModels.getInstance().getBlosum62().getMatrix(); - - // ResidueProperties.getBLOSUM62(); - - String[] intToStr = pep; - - int defInt = 23; - StringBuffer output = new StringBuffer(); - String type; + String type; // AlignSeq.PEP or AlignSeq.DNA + + private ScoreMatrix scoreMatrix; - private int[] charToInt; + private static final int GAP_INDEX = -1; /** * Creates a new AlignSeq object. * - * @param s1 - * DOCUMENT ME! - * @param s2 - * DOCUMENT ME! - * @param type - * DOCUMENT ME! + * @param s1 first sequence for alignment + * @param s2 second sequence for alignment + * @param type molecule type, either AlignSeq.PEP or AlignSeq.DNA */ public AlignSeq(SequenceI s1, SequenceI s2, String type) { - SeqInit(s1, s1.getSequenceAsString(), s2, s2.getSequenceAsString(), + seqInit(s1, s1.getSequenceAsString(), s2, s2.getSequenceAsString(), type); } @@ -161,7 +146,7 @@ public class AlignSeq public AlignSeq(SequenceI s1, String string1, SequenceI s2, String string2, String type) { - SeqInit(s1, string1.toUpperCase(), s2, string2.toUpperCase(), type); + seqInit(s1, string1.toUpperCase(), s2, string2.toUpperCase(), type); } /** @@ -265,26 +250,6 @@ public class AlignSeq } /** - * DOCUMENT ME! - * - * @return DOCUMENT ME! - */ - public SequenceI getS1() - { - return s1; - } - - /** - * DOCUMENT ME! - * - * @return DOCUMENT ME! - */ - public SequenceI getS2() - { - return s2; - } - - /** * * @return aligned instance of Seq 1 */ @@ -326,36 +291,13 @@ public class AlignSeq * @param type * DNA or PEPTIDE */ - public void SeqInit(SequenceI s1, String string1, SequenceI s2, + public void seqInit(SequenceI s1, String string1, SequenceI s2, String string2, String type) { this.s1 = s1; this.s2 = s2; setDefaultParams(type); - SeqInit(string1, string2); - } - - /** - * Construct score matrix for sequences with custom substitution matrix - * - * @param s1 - * - sequence 1 - * @param string1 - * - string to use for s1 - * @param s2 - * - sequence 2 - * @param string2 - * - string to use for s2 - * @param scoreMatrix - * - substitution matrix to use for alignment - */ - public void SeqInit(SequenceI s1, String string1, SequenceI s2, - String string2, ScoreMatrix scoreMatrix) - { - this.s1 = s1; - this.s2 = s2; - setType(scoreMatrix.isDNA() ? AlignSeq.DNA : AlignSeq.PEP); - lookup = scoreMatrix.getMatrix(); + seqInit(string1, string2); } /** @@ -365,7 +307,7 @@ public class AlignSeq * @param string1 * @param string2 */ - private void SeqInit(String string1, String string2) + private void seqInit(String string1, String string2) { s1str = extractGaps(jalview.util.Comparison.GapChars, string1); s2str = extractGaps(jalview.util.Comparison.GapChars, string2); @@ -378,84 +320,31 @@ public class AlignSeq return; } - // System.out.println("lookuip " + rt.freeMemory() + " "+ rt.totalMemory()); - seq1 = new int[s1str.length()]; - - // System.out.println("seq1 " + rt.freeMemory() +" " + rt.totalMemory()); - seq2 = new int[s2str.length()]; - - // System.out.println("seq2 " + rt.freeMemory() + " " + rt.totalMemory()); score = new float[s1str.length()][s2str.length()]; - // System.out.println("score " + rt.freeMemory() + " " + rt.totalMemory()); E = new float[s1str.length()][s2str.length()]; - // System.out.println("E " + rt.freeMemory() + " " + rt.totalMemory()); F = new float[s1str.length()][s2str.length()]; traceback = new int[s1str.length()][s2str.length()]; - // System.out.println("F " + rt.freeMemory() + " " + rt.totalMemory()); - seq1 = stringToInt(s1str, type); - - // System.out.println("seq1 " + rt.freeMemory() + " " + rt.totalMemory()); - seq2 = stringToInt(s2str, type); - - // System.out.println("Seq2 " + rt.freeMemory() + " " + rt.totalMemory()); - // long tstart = System.currentTimeMillis(); - // calcScoreMatrix(); - // long tend = System.currentTimeMillis(); - // System.out.println("Time take to calculate score matrix = " + - // (tend-tstart) + " ms"); - // printScoreMatrix(score); - // System.out.println(); - // printScoreMatrix(traceback); - // System.out.println(); - // printScoreMatrix(E); - // System.out.println(); - // /printScoreMatrix(F); - // System.out.println(); - // tstart = System.currentTimeMillis(); - // traceAlignment(); - // tend = System.currentTimeMillis(); - // System.out.println("Time take to traceback alignment = " + (tend-tstart) - // + " ms"); - } - - private void setDefaultParams(String type) - { - setType(type); + seq1 = indexEncode(s1str); - if (type.equals(AlignSeq.PEP)) - { - lookup = ScoreModels.getInstance().getDefaultModel(true).getMatrix(); - } - else if (type.equals(AlignSeq.DNA)) - { - lookup = ScoreModels.getInstance().getDefaultModel(false).getMatrix(); - } + seq2 = indexEncode(s2str); } - private void setType(String type2) + private void setDefaultParams(String moleculeType) { - this.type = type2; - if (type.equals(AlignSeq.PEP)) - { - intToStr = pep; - charToInt = ResidueProperties.aaIndex; - defInt = ResidueProperties.maxProteinIndex; - } - else if (type.equals(AlignSeq.DNA)) - { - intToStr = dna; - charToInt = ResidueProperties.nucleotideIndex; - defInt = ResidueProperties.maxNucleotideIndex; - } - else + if (!PEP.equals(moleculeType) && !DNA.equals(moleculeType)) { output.append("Wrong type = dna or pep only"); throw new Error(MessageManager.formatMessage( - "error.unknown_type_dna_or_pep", new String[] { type2 })); + "error.unknown_type_dna_or_pep", + new String[] { moleculeType })); } + + type = moleculeType; + scoreMatrix = ScoreModels.getInstance().getDefaultModel( + PEP.equals(type)); } /** @@ -464,7 +353,7 @@ public class AlignSeq public void traceAlignment() { // Find the maximum score along the rhs or bottom row - float max = -9999; + float max = -Float.MAX_VALUE; for (int i = 0; i < seq1.length; i++) { @@ -498,21 +387,17 @@ public class AlignSeq aseq1 = new int[seq1.length + seq2.length]; aseq2 = new int[seq1.length + seq2.length]; + StringBuilder sb1 = new StringBuilder(aseq1.length); + StringBuilder sb2 = new StringBuilder(aseq2.length); + count = (seq1.length + seq2.length) - 1; - while ((i > 0) && (j > 0)) + while (i > 0 && j > 0) { - if ((aseq1[count] != defInt) && (i >= 0)) - { - aseq1[count] = seq1[i]; - astr1 = s1str.charAt(i) + astr1; - } - - if ((aseq2[count] != defInt) && (j > 0)) - { - aseq2[count] = seq2[j]; - astr2 = s2str.charAt(j) + astr2; - } + aseq1[count] = seq1[i]; + sb1.append(s1str.charAt(i)); + aseq2[count] = seq2[j]; + sb2.append(s2str.charAt(j)); trace = findTrace(i, j); @@ -524,14 +409,14 @@ public class AlignSeq else if (trace == 1) { j--; - aseq1[count] = defInt; - astr1 = "-" + astr1.substring(1); + aseq1[count] = GAP_INDEX; + sb1.replace(sb1.length() - 1, sb1.length(), "-"); } else if (trace == -1) { i--; - aseq2[count] = defInt; - astr2 = "-" + astr2.substring(1); + aseq2[count] = GAP_INDEX; + sb2.replace(sb2.length() - 1, sb2.length(), "-"); } count--; @@ -540,17 +425,24 @@ public class AlignSeq seq1start = i + 1; seq2start = j + 1; - if (aseq1[count] != defInt) + if (aseq1[count] != GAP_INDEX) { aseq1[count] = seq1[i]; - astr1 = s1str.charAt(i) + astr1; + sb1.append(s1str.charAt(i)); } - if (aseq2[count] != defInt) + if (aseq2[count] != GAP_INDEX) { aseq2[count] = seq2[j]; - astr2 = s2str.charAt(j) + astr2; + sb2.append(s2str.charAt(j)); } + + /* + * we built the character strings backwards, so now + * reverse them to convert to sequence strings + */ + astr1 = sb1.reverse().toString(); + astr2 = sb2.reverse().toString(); } /** @@ -603,7 +495,7 @@ public class AlignSeq .append(String.valueOf(s2str.length())).append(")") .append(NEWLINE).append(NEWLINE); - PairwiseSeqScoreModel pam250 = ScoreModels.getInstance().getPam250(); + ScoreMatrix pam250 = ScoreModels.getInstance().getPam250(); for (int j = 0; j < nochunks; j++) { @@ -686,46 +578,6 @@ public class AlignSeq /** * DOCUMENT ME! * - * @param mat - * DOCUMENT ME! - */ - public void printScoreMatrix(int[][] mat) - { - int n = seq1.length; - int m = seq2.length; - - for (int i = 0; i < n; i++) - { - // Print the top sequence - if (i == 0) - { - Format.print(System.out, "%8s", s2str.substring(0, 1)); - - for (int jj = 1; jj < m; jj++) - { - Format.print(System.out, "%5s", s2str.substring(jj, jj + 1)); - } - - System.out.println(); - } - - for (int j = 0; j < m; j++) - { - if (j == 0) - { - Format.print(System.out, "%3s", s1str.substring(i, i + 1)); - } - - Format.print(System.out, "%3d ", mat[i][j] / 10); - } - - System.out.println(); - } - } - - /** - * DOCUMENT ME! - * * @param i * DOCUMENT ME! * @param j @@ -736,7 +588,10 @@ public class AlignSeq public int findTrace(int i, int j) { int t = 0; - float max = score[i - 1][j - 1] + (lookup[seq1[i]][seq2[j]] * 10); + // float pairwiseScore = lookup[seq1[i]][seq2[j]]; + float pairwiseScore = scoreMatrix.getPairwiseScore(s1str.charAt(i), + s2str.charAt(j)); + float max = score[i - 1][j - 1] + (pairwiseScore * 10); if (F[i][j] > max) { @@ -780,7 +635,8 @@ public class AlignSeq int m = seq2.length; // top left hand element - score[0][0] = lookup[seq1[0]][seq2[0]] * 10; + score[0][0] = scoreMatrix.getPairwiseScore(s1str.charAt(0), + s2str.charAt(0)) * 10; E[0][0] = -gapExtend; F[0][0] = 0; @@ -791,7 +647,9 @@ public class AlignSeq E[0][j] = max(score[0][j - 1] - gapOpen, E[0][j - 1] - gapExtend); F[0][j] = -gapExtend; - score[0][j] = max(lookup[seq1[0]][seq2[j]] * 10, -gapOpen, -gapExtend); + float pairwiseScore = scoreMatrix.getPairwiseScore(s1str.charAt(0), + s2str.charAt(j)); + score[0][j] = max(pairwiseScore * 10, -gapOpen, -gapExtend); traceback[0][j] = 1; } @@ -802,7 +660,9 @@ public class AlignSeq E[i][0] = -gapOpen; F[i][0] = max(score[i - 1][0] - gapOpen, F[i - 1][0] - gapExtend); - score[i][0] = max(lookup[seq1[i]][seq2[0]] * 10, E[i][0], F[i][0]); + float pairwiseScore = scoreMatrix.getPairwiseScore(s1str.charAt(i), + s2str.charAt(0)); + score[i][0] = max(pairwiseScore * 10, E[i][0], F[i][0]); traceback[i][0] = -1; } @@ -814,8 +674,10 @@ public class AlignSeq E[i][j] = max(score[i][j - 1] - gapOpen, E[i][j - 1] - gapExtend); F[i][j] = max(score[i - 1][j] - gapOpen, F[i - 1][j] - gapExtend); + float pairwiseScore = scoreMatrix.getPairwiseScore(s1str.charAt(i), + s2str.charAt(j)); score[i][j] = max(score[i - 1][j - 1] - + (lookup[seq1[i]][seq2[j]] * 10), E[i][j], F[i][j]); + + (pairwiseScore * 10), E[i][j], F[i][j]); traceback[i][j] = findTrace(i, j); } } @@ -860,7 +722,7 @@ public class AlignSeq * * @return DOCUMENT ME! */ - public float max(float f1, float f2, float f3) + private static float max(float f1, float f2, float f3) { float max = f1; @@ -887,7 +749,7 @@ public class AlignSeq * * @return DOCUMENT ME! */ - public float max(float f1, float f2) + private static float max(float f1, float f2) { float max = f1; @@ -900,45 +762,24 @@ public class AlignSeq } /** - * DOCUMENT ME! + * Converts the character string to an array of integers which are the + * corresponding indices to the characters in the score matrix * * @param s - * DOCUMENT ME! - * @param type - * DOCUMENT ME! * - * @return DOCUMENT ME! + * @return */ - public int[] stringToInt(String s, String type) + int[] indexEncode(String s) { - int[] seq1 = new int[s.length()]; + int[] encoded = new int[s.length()]; for (int i = 0; i < s.length(); i++) { - // String ss = s.substring(i, i + 1).toUpperCase(); char c = s.charAt(i); - if ('a' <= c && c <= 'z') - { - // TO UPPERCASE !!! - c -= ('a' - 'A'); - } - - try - { - seq1[i] = charToInt[c]; // set accordingly from setType - if (seq1[i] < 0 || seq1[i] > defInt) // set from setType: 23 for - // peptides, or 4 for NA. - { - seq1[i] = defInt; - } - - } catch (Exception e) - { - seq1[i] = defInt; - } + encoded[i] = scoreMatrix.getMatrixIndex(c); } - return seq1; + return encoded; } /** @@ -958,7 +799,7 @@ public class AlignSeq public static void displayMatrix(Graphics g, int[][] mat, int n, int m, int psize) { - // TODO method dosen't seem to be referenced anywhere delete?? + // TODO method doesn't seem to be referenced anywhere delete?? int max = -1000; int min = 1000; @@ -1132,8 +973,8 @@ public class AlignSeq bestm = msq; } } - System.out.println("Best Score for " + (matches.size() + 1) + " :" - + bestscore); + // System.out.println("Best Score for " + (matches.size() + 1) + " :" + // + bestscore); matches.add(bestm); aligns.add(bestaseq); al.deleteSequence(bestm); @@ -1222,6 +1063,8 @@ public class AlignSeq // long start = System.currentTimeMillis(); + SimilarityParams pidParams = new SimilarityParams(true, true, true, + true); float pid; String seqi, seqj; for (int i = 0; i < height; i++) @@ -1262,7 +1105,7 @@ public class AlignSeq seqj = ug; } } - pid = Comparison.PID(seqi, seqj); + pid = (float) PIDModel.computePID(seqi, seqj, pidParams); // use real sequence length rather than string length if (lngth[j] < lngth[i])