X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FAAFrequency.java;h=900209d34fa6b0ee0eb23e40d8cde607ff50b7ef;hb=e5cba1e6a138feeeb8cf5b5823b3cafc89e71b9f;hp=30d53737cda0309f04a4ed591563656f30177356;hpb=0a5ce6145bb76fc7eb8a5cc2670e20453fbedd29;p=jalview.git diff --git a/src/jalview/analysis/AAFrequency.java b/src/jalview/analysis/AAFrequency.java index 30d5373..900209d 100755 --- a/src/jalview/analysis/AAFrequency.java +++ b/src/jalview/analysis/AAFrequency.java @@ -24,16 +24,22 @@ import jalview.datamodel.AlignedCodonFrame; import jalview.datamodel.AlignmentAnnotation; import jalview.datamodel.AlignmentI; import jalview.datamodel.Annotation; +import jalview.datamodel.HiddenMarkovModel; import jalview.datamodel.Profile; +import jalview.datamodel.ProfileI; +import jalview.datamodel.Profiles; +import jalview.datamodel.ProfilesI; import jalview.datamodel.ResidueCount; -import jalview.datamodel.SequenceI; import jalview.datamodel.ResidueCount.SymbolCounts; +import jalview.datamodel.SequenceI; import jalview.ext.android.SparseIntArray; +import jalview.schemes.ResidueProperties; import jalview.util.Comparison; import jalview.util.Format; import jalview.util.MappingUtils; import jalview.util.QuickSort; +import java.awt.Color; import java.util.Arrays; import java.util.Hashtable; import java.util.List; @@ -51,6 +57,12 @@ public class AAFrequency { public static final String PROFILE = "P"; + private static final String AMINO = "amino"; + + private static final String DNA = "DNA"; + + private static final String RNA = "RNA"; + /* * Quick look-up of String value of char 'A' to 'Z' */ @@ -64,13 +76,13 @@ public class AAFrequency } } - public static final Profile[] calculate(List list, - int start, int end) + public static final ProfilesI calculate(List list, int start, + int end) { return calculate(list, start, end, false); } - public static final Profile[] calculate(List sequences, + public static final ProfilesI calculate(List sequences, int start, int end, boolean profile) { SequenceI[] seqs = new SequenceI[sequences.size()]; @@ -80,39 +92,39 @@ public class AAFrequency for (int i = 0; i < sequences.size(); i++) { seqs[i] = sequences.get(i); - if (seqs[i].getLength() > width) + int length = seqs[i].getLength(); + if (length > width) { - width = seqs[i].getLength(); + width = length; } } - Profile[] reply = new Profile[width]; - if (end >= width) { end = width; } - calculate(seqs, start, end, reply, profile); + ProfilesI reply = calculate(seqs, width, start, end, profile); return reply; } } + /** * Calculate the consensus symbol(s) for each column in the given range. * * @param sequences + * @param width + * the full width of the alignment * @param start * start column (inclusive, base zero) * @param end * end column (exclusive) - * @param result - * array in which to store profile per column * @param saveFullProfile * if true, store all symbol counts */ - public static final void calculate(final SequenceI[] sequences, - int start, int end, Profile[] result, boolean saveFullProfile) + public static final ProfilesI calculate(final SequenceI[] sequences, + int width, int start, int end, boolean saveFullProfile) { // long now = System.currentTimeMillis(); int seqCount = sequences.length; @@ -120,6 +132,8 @@ public class AAFrequency int nucleotideCount = 0; int peptideCount = 0; + ProfileI[] result = new ProfileI[width]; + for (int column = start; column < end; column++) { /* @@ -172,7 +186,7 @@ public class AAFrequency int maxCount = residueCounts.getModalCount(); String maxResidue = residueCounts.getResiduesForCount(maxCount); int gapCount = residueCounts.getGapCount(); - Profile profile = new Profile(seqCount, gapCount, maxCount, + ProfileI profile = new Profile(seqCount, gapCount, maxCount, maxResidue); if (saveFullProfile) @@ -182,6 +196,7 @@ public class AAFrequency result[column] = profile; } + return new Profiles(result); // long elapsed = System.currentTimeMillis() - now; // System.out.println(elapsed); } @@ -220,10 +235,10 @@ public class AAFrequency * the annotation row to add annotations to * @param profiles * the source consensus data - * @param iStart - * start column - * @param width - * end column + * @param startCol + * start column (inclusive) + * @param endCol + * end column (exclusive) * @param ignoreGaps * if true, normalise residue percentages ignoring gaps * @param showSequenceLogo @@ -233,12 +248,12 @@ public class AAFrequency * number of sequences */ public static void completeConsensus(AlignmentAnnotation consensus, - Profile[] profiles, int iStart, int width, boolean ignoreGaps, + ProfilesI profiles, int startCol, int endCol, boolean ignoreGaps, boolean showSequenceLogo, long nseq) { // long now = System.currentTimeMillis(); if (consensus == null || consensus.annotations == null - || consensus.annotations.length < width) + || consensus.annotations.length < endCol) { /* * called with a bad alignment annotation row @@ -247,21 +262,21 @@ public class AAFrequency return; } - final int dp = getPercentageDp(nseq); - - for (int i = iStart; i < width; i++) + for (int i = startCol; i < endCol; i++) { - Profile profile; - if (i >= profiles.length || ((profile = profiles[i]) == null)) + ProfileI profile = profiles.get(i); + if (profile == null) { /* * happens if sequences calculated over were * shorter than alignment width */ consensus.annotations[i] = null; - continue; + return; } + final int dp = getPercentageDp(nseq); + float value = profile.getPercentageIdentity(ignoreGaps); String description = getTooltip(profile, value, showSequenceLogo, @@ -276,14 +291,65 @@ public class AAFrequency { modalResidue = "+"; } - consensus.annotations[i] = new Annotation(modalResidue, - description, ' ', value); + consensus.annotations[i] = new Annotation(modalResidue, description, + ' ', value); } // long elapsed = System.currentTimeMillis() - now; // System.out.println(-elapsed); } /** + * Derive the gap count annotation row. + * + * @param gaprow + * the annotation row to add annotations to + * @param profiles + * the source consensus data + * @param startCol + * start column (inclusive) + * @param endCol + * end column (exclusive) + */ + public static void completeGapAnnot(AlignmentAnnotation gaprow, + ProfilesI profiles, int startCol, int endCol, long nseq) + { + if (gaprow == null || gaprow.annotations == null + || gaprow.annotations.length < endCol) + { + /* + * called with a bad alignment annotation row + * wait for it to be initialised properly + */ + return; + } + // always set ranges again + gaprow.graphMax = nseq; + gaprow.graphMin = 0; + double scale = 0.8/nseq; + for (int i = startCol; i < endCol; i++) + { + ProfileI profile = profiles.get(i); + if (profile == null) + { + /* + * happens if sequences calculated over were + * shorter than alignment width + */ + gaprow.annotations[i] = null; + return; + } + + final int gapped = profile.getNonGapped(); + + String description = "" + gapped; + + gaprow.annotations[i] = new Annotation("", description, + '\0', gapped, jalview.util.ColorUtils.bleachColour( + Color.DARK_GRAY, (float) scale * gapped)); + } + } + + /** * Returns a tooltip showing either *
    *
  • the full profile (percentages of all residues present), if @@ -301,7 +367,7 @@ public class AAFrequency * the number of decimal places to format percentages to * @return */ - static String getTooltip(Profile profile, float pid, + static String getTooltip(ProfileI profile, float pid, boolean showSequenceLogo, boolean ignoreGaps, int dp) { ResidueCount counts = profile.getCounts(); @@ -352,8 +418,7 @@ public class AAFrequency * calculations * @return */ - public static int[] extractProfile(Profile profile, - boolean ignoreGaps) + public static int[] extractProfile(ProfileI profile, boolean ignoreGaps) { int[] rtnval = new int[64]; ResidueCount counts = profile.getCounts(); @@ -393,6 +458,7 @@ public class AAFrequency return result; } + /** * Extract a sorted extract of cDNA codon profile data. The returned array * contains @@ -475,7 +541,7 @@ public class AAFrequency for (int col = 0; col < cols; col++) { // todo would prefer a Java bean for consensus data - Hashtable columnHash = new Hashtable(); + Hashtable columnHash = new Hashtable<>(); // #seqs, #ungapped seqs, counts indexed by (codon encoded + 1) int[] codonCounts = new int[66]; codonCounts[0] = alignment.getSequences().size(); @@ -662,4 +728,95 @@ public class AAFrequency } return scale; } + + /** + * Produces a HMM profile for a column in an alignment + * + * @param aa + * Alignment annotation for which the profile is being calculated. + * @param column + * Column in the alignment the profile is being made for. + * @param removeBelowBackground + * Boolean indicating whether to ignore residues with probabilities + * less than their background frequencies. + * @return + */ + public static int[] getHMMProfileFor(HiddenMarkovModel hmm, int column, + boolean removeBelowBackground) + { + + if (hmm != null) + { + String alph = hmm.getAlphabetType(); + int size = hmm.getNumberOfSymbols(); + char symbols[] = new char[size]; + int values[] = new int[size]; + List charList = hmm.getSymbols(); + Integer totalCount = 0; + + for (int i = 0; i < size; i++) + { + char symbol = charList.get(i); + symbols[i] = symbol; + Double value; + + value = hmm.getMatchEmissionProbability(column, symbol); + double freq; + + if (AMINO.equals(alph) && removeBelowBackground) + { + freq = ResidueProperties.aminoBackgroundFrequencies.get(symbol); + if (value < freq) + { + value = 0d; + } + } + else if (DNA.equals(alph) && removeBelowBackground) + { + freq = ResidueProperties.dnaBackgroundFrequencies.get(symbol); + if (value < freq) + { + value = 0d; + } + } + else if (RNA.equals(alph) && removeBelowBackground) + { + freq = ResidueProperties.rnaBackgroundFrequencies + .get(symbol); + if (value < freq) + { + value = 0d; + } + } + value = value * 10000; + values[i] = value.intValue(); + totalCount += value.intValue(); + } + + QuickSort.sort(values, symbols); + + int[] profile = new int[3 + size * 2]; + + profile[0] = AlignmentAnnotation.SEQUENCE_PROFILE; + profile[1] = size; + profile[2] = totalCount / 100; + + if (totalCount != 0) + { + int arrayPos = 3; + for (int k = size - 1; k >= 0; k--) + { + Double percentage; + Integer value = values[k]; + percentage = (value.doubleValue() / totalCount.doubleValue()) + * 100d; + profile[arrayPos] = symbols[k]; + profile[arrayPos + 1] = percentage.intValue(); + arrayPos += 2; + } + } + return profile; + } + return null; + } }