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.datamodel.AlignedCodonFrame;
24 import jalview.datamodel.AlignmentAnnotation;
25 import jalview.datamodel.AlignmentI;
26 import jalview.datamodel.Annotation;
27 import jalview.datamodel.Profile;
28 import jalview.datamodel.ProfileI;
29 import jalview.datamodel.Profiles;
30 import jalview.datamodel.ProfilesI;
31 import jalview.datamodel.ResidueCount;
32 import jalview.datamodel.ResidueCount.SymbolCounts;
33 import jalview.datamodel.SequenceI;
34 import jalview.ext.android.SparseIntArray;
35 import jalview.util.Comparison;
36 import jalview.util.Format;
37 import jalview.util.MappingUtils;
38 import jalview.util.QuickSort;
40 import java.awt.Color;
41 import java.util.Arrays;
42 import java.util.Hashtable;
43 import java.util.List;
46 * Takes in a vector or array of sequences and column start and column end and
47 * returns a new Hashtable[] of size maxSeqLength, if Hashtable not supplied.
48 * This class is used extensively in calculating alignment colourschemes that
49 * depend on the amount of conservation in each alignment column.
54 public class AAFrequency
56 public static final String PROFILE = "P";
59 * Quick look-up of String value of char 'A' to 'Z'
61 private static final String[] CHARS = new String['Z' - 'A' + 1];
65 for (char c = 'A'; c <= 'Z'; c++)
67 CHARS[c - 'A'] = String.valueOf(c);
71 public static final ProfilesI calculate(List<SequenceI> list, int start,
74 return calculate(list, start, end, false);
77 public static final ProfilesI calculate(List<SequenceI> sequences,
78 int start, int end, boolean profile)
80 SequenceI[] seqs = new SequenceI[sequences.size()];
82 synchronized (sequences)
84 for (int i = 0; i < sequences.size(); i++)
86 seqs[i] = sequences.get(i);
87 int length = seqs[i].getLength();
99 ProfilesI reply = calculate(seqs, width, start, end, profile);
105 * Calculate the consensus symbol(s) for each column in the given range.
109 * the full width of the alignment
111 * start column (inclusive, base zero)
113 * end column (exclusive)
114 * @param saveFullProfile
115 * if true, store all symbol counts
117 public static final ProfilesI calculate(final SequenceI[] sequences,
118 int width, int start, int end, boolean saveFullProfile)
120 // long now = System.currentTimeMillis();
121 int seqCount = sequences.length;
122 boolean nucleotide = false;
123 int nucleotideCount = 0;
124 int peptideCount = 0;
126 ProfileI[] result = new ProfileI[width];
128 for (int column = start; column < end; column++)
131 * Apply a heuristic to detect nucleotide data (which can
132 * be counted in more compact arrays); here we test for
133 * more than 90% nucleotide; recheck every 10 columns in case
134 * of misleading data e.g. highly conserved Alanine in peptide!
135 * Mistakenly guessing nucleotide has a small performance cost,
136 * as it will result in counting in sparse arrays.
137 * Mistakenly guessing peptide has a small space cost,
138 * as it will use a larger than necessary array to hold counts.
140 if (nucleotideCount > 100 && column % 10 == 0)
142 nucleotide = (9 * peptideCount < nucleotideCount);
144 ResidueCount residueCounts = new ResidueCount(nucleotide);
146 for (int row = 0; row < seqCount; row++)
148 if (sequences[row] == null)
151 "WARNING: Consensus skipping null sequence - possible race condition.");
154 if (sequences[row].getLength() > column)
156 char c = sequences[row].getCharAt(column);
157 residueCounts.add(c);
158 if (Comparison.isNucleotide(c))
162 else if (!Comparison.isGap(c))
170 * count a gap if the sequence doesn't reach this column
172 residueCounts.addGap();
176 int maxCount = residueCounts.getModalCount();
177 String maxResidue = residueCounts.getResiduesForCount(maxCount);
178 int gapCount = residueCounts.getGapCount();
179 ProfileI profile = new Profile(seqCount, gapCount, maxCount,
184 profile.setCounts(residueCounts);
187 result[column] = profile;
189 return new Profiles(result);
190 // long elapsed = System.currentTimeMillis() - now;
191 // System.out.println(elapsed);
195 * Make an estimate of the profile size we are going to compute i.e. how many
196 * different characters may be present in it. Overestimating has a cost of
197 * using more memory than necessary. Underestimating has a cost of needing to
198 * extend the SparseIntArray holding the profile counts.
200 * @param profileSizes
201 * counts of sizes of profiles so far encountered
204 static int estimateProfileSize(SparseIntArray profileSizes)
206 if (profileSizes.size() == 0)
212 * could do a statistical heuristic here e.g. 75%ile
213 * for now just return the largest value
215 return profileSizes.keyAt(profileSizes.size() - 1);
219 * Derive the consensus annotations to be added to the alignment for display.
220 * This does not recompute the raw data, but may be called on a change in
221 * display options, such as 'ignore gaps', which may in turn result in a
222 * change in the derived values.
225 * the annotation row to add annotations to
227 * the source consensus data
229 * start column (inclusive)
231 * end column (exclusive)
233 * if true, normalise residue percentages ignoring gaps
234 * @param showSequenceLogo
235 * if true include all consensus symbols, else just show modal
238 * number of sequences
240 public static void completeConsensus(AlignmentAnnotation consensus,
241 ProfilesI profiles, int startCol, int endCol, boolean ignoreGaps,
242 boolean showSequenceLogo, long nseq)
244 // long now = System.currentTimeMillis();
245 if (consensus == null || consensus.annotations == null
246 || consensus.annotations.length < endCol)
249 * called with a bad alignment annotation row
250 * wait for it to be initialised properly
255 float threshhold = 0;
256 if (consensus.getThreshold()!=null)
258 threshhold = consensus.getThreshold().value;
260 for (int i = startCol; i < endCol; i++)
262 ProfileI profile = profiles.get(i);
266 * happens if sequences calculated over were
267 * shorter than alignment width
269 consensus.annotations[i] = null;
273 final int dp = getPercentageDp(nseq);
275 float value = profile.getPercentageIdentity(ignoreGaps);
277 String description = getTooltip(profile, value, showSequenceLogo,
280 String modalResidue = profile.getModalResidue();
281 if ("".equals(modalResidue) || threshhold>value)
285 else if (modalResidue.length() > 1)
289 consensus.annotations[i] = new Annotation(modalResidue, description,
292 // long elapsed = System.currentTimeMillis() - now;
293 // System.out.println(-elapsed);
297 * Derive the gap count annotation row.
300 * the annotation row to add annotations to
302 * the source consensus data
304 * start column (inclusive)
306 * end column (exclusive)
308 public static void completeGapAnnot(AlignmentAnnotation gaprow,
309 ProfilesI profiles, int startCol, int endCol, long nseq)
311 if (gaprow == null || gaprow.annotations == null
312 || gaprow.annotations.length < endCol)
315 * called with a bad alignment annotation row
316 * wait for it to be initialised properly
320 // always set ranges again
321 gaprow.graphMax = nseq;
323 double scale = 0.8 / nseq;
324 for (int i = startCol; i < endCol; i++)
326 ProfileI profile = profiles.get(i);
330 * happens if sequences calculated over were
331 * shorter than alignment width
333 gaprow.annotations[i] = null;
337 final int gapped = profile.getNonGapped();
339 String description = "" + gapped;
341 gaprow.annotations[i] = new Annotation("", description, '\0', gapped,
342 jalview.util.ColorUtils.bleachColour(Color.DARK_GRAY,
343 (float) scale * gapped));
348 * Returns a tooltip showing either
350 * <li>the full profile (percentages of all residues present), if
351 * showSequenceLogo is true, or</li>
352 * <li>just the modal (most common) residue(s), if showSequenceLogo is
355 * Percentages are as a fraction of all sequence, or only ungapped sequences
356 * if ignoreGaps is true.
360 * @param showSequenceLogo
363 * the number of decimal places to format percentages to
366 static String getTooltip(ProfileI profile, float pid,
367 boolean showSequenceLogo, boolean ignoreGaps, int dp)
369 ResidueCount counts = profile.getCounts();
371 String description = null;
372 if (counts != null && showSequenceLogo)
374 int normaliseBy = ignoreGaps ? profile.getNonGapped()
375 : profile.getHeight();
376 description = counts.getTooltip(normaliseBy, dp);
380 StringBuilder sb = new StringBuilder(64);
381 String maxRes = profile.getModalResidue();
382 if (maxRes.length() > 1)
384 sb.append("[").append(maxRes).append("]");
390 if (maxRes.length() > 0)
393 Format.appendPercentage(sb, pid, dp);
396 description = sb.toString();
402 * Returns the sorted profile for the given consensus data. The returned array
406 * [profileType, numberOfValues, totalPercent, charValue1, percentage1, charValue2, percentage2, ...]
407 * in descending order of percentage value
411 * the data object from which to extract and sort values
413 * if true, only non-gapped values are included in percentage
417 public static int[] extractProfile(ProfileI profile, boolean ignoreGaps)
419 ResidueCount counts = profile.getCounts();
425 SymbolCounts symbolCounts = counts.getSymbolCounts();
426 char[] symbols = symbolCounts.symbols;
427 int[] values = symbolCounts.values;
428 QuickSort.sort(values, symbols);
429 int totalPercentage = 0;
430 final int divisor = ignoreGaps ? profile.getNonGapped()
431 : profile.getHeight();
434 * traverse the arrays in reverse order (highest counts first)
436 int[] result = new int[3 + 2 * symbols.length];
437 int nextArrayPos = 3;
438 int nonZeroCount = 0;
440 for (int i = symbols.length - 1; i >= 0; i--)
442 int theChar = symbols[i];
443 int charCount = values[i];
444 final int percentage = (charCount * 100) / divisor;
448 * this count (and any remaining) round down to 0% - discard
453 result[nextArrayPos++] = theChar;
454 result[nextArrayPos++] = percentage;
455 totalPercentage += percentage;
459 * truncate array if any zero values were discarded
461 if (nonZeroCount < symbols.length)
463 int[] tmp = new int[3 + 2 * nonZeroCount];
464 System.arraycopy(result, 0, tmp, 0, tmp.length);
469 * fill in 'header' values
471 result[0] = AlignmentAnnotation.SEQUENCE_PROFILE;
472 result[1] = nonZeroCount;
473 result[2] = totalPercentage;
479 * Extract a sorted extract of cDNA codon profile data. The returned array
483 * [profileType, numberOfValues, totalPercentage, charValue1, percentage1, charValue2, percentage2, ...]
484 * in descending order of percentage value, where the character values encode codon triplets
490 public static int[] extractCdnaProfile(
491 Hashtable<String, Object> hashtable, boolean ignoreGaps)
493 // this holds #seqs, #ungapped, and then codon count, indexed by encoded
495 int[] codonCounts = (int[]) hashtable.get(PROFILE);
496 int[] sortedCounts = new int[codonCounts.length - 2];
497 System.arraycopy(codonCounts, 2, sortedCounts, 0,
498 codonCounts.length - 2);
500 int[] result = new int[3 + 2 * sortedCounts.length];
501 // first value is just the type of profile data
502 result[0] = AlignmentAnnotation.CDNA_PROFILE;
504 char[] codons = new char[sortedCounts.length];
505 for (int i = 0; i < codons.length; i++)
507 codons[i] = (char) i;
509 QuickSort.sort(sortedCounts, codons);
510 int totalPercentage = 0;
511 int distinctValuesCount = 0;
513 int divisor = ignoreGaps ? codonCounts[1] : codonCounts[0];
514 for (int i = codons.length - 1; i >= 0; i--)
516 final int codonCount = sortedCounts[i];
519 break; // nothing else of interest here
521 final int percentage = codonCount * 100 / divisor;
525 * this (and any remaining) values rounded down to 0 - discard
529 distinctValuesCount++;
530 result[j++] = codons[i];
531 result[j++] = percentage;
532 totalPercentage += percentage;
534 result[2] = totalPercentage;
537 * Just return the non-zero values
539 // todo next value is redundant if we limit the array to non-zero counts
540 result[1] = distinctValuesCount;
541 return Arrays.copyOfRange(result, 0, j);
545 * Compute a consensus for the cDNA coding for a protein alignment.
548 * the protein alignment (which should hold mappings to cDNA
551 * the consensus data stores to be populated (one per column)
553 public static void calculateCdna(AlignmentI alignment,
554 Hashtable<String, Object>[] hconsensus)
556 final char gapCharacter = alignment.getGapCharacter();
557 List<AlignedCodonFrame> mappings = alignment.getCodonFrames();
558 if (mappings == null || mappings.isEmpty())
563 int cols = alignment.getWidth();
564 for (int col = 0; col < cols; col++)
566 // todo would prefer a Java bean for consensus data
567 Hashtable<String, Object> columnHash = new Hashtable<>();
568 // #seqs, #ungapped seqs, counts indexed by (codon encoded + 1)
569 int[] codonCounts = new int[66];
570 codonCounts[0] = alignment.getSequences().size();
571 int ungappedCount = 0;
572 for (SequenceI seq : alignment.getSequences())
574 if (seq.getCharAt(col) == gapCharacter)
578 List<char[]> codons = MappingUtils.findCodonsFor(seq, col,
580 for (char[] codon : codons)
582 int codonEncoded = CodingUtils.encodeCodon(codon);
583 if (codonEncoded >= 0)
585 codonCounts[codonEncoded + 2]++;
591 codonCounts[1] = ungappedCount;
592 // todo: sort values here, save counts and codons?
593 columnHash.put(PROFILE, codonCounts);
594 hconsensus[col] = columnHash;
599 * Derive displayable cDNA consensus annotation from computed consensus data.
601 * @param consensusAnnotation
602 * the annotation row to be populated for display
603 * @param consensusData
604 * the computed consensus data
605 * @param showProfileLogo
606 * if true show all symbols present at each position, else only the
609 * the number of sequences in the alignment
611 public static void completeCdnaConsensus(
612 AlignmentAnnotation consensusAnnotation,
613 Hashtable<String, Object>[] consensusData,
614 boolean showProfileLogo, int nseqs)
616 if (consensusAnnotation == null
617 || consensusAnnotation.annotations == null
618 || consensusAnnotation.annotations.length < consensusData.length)
620 // called with a bad alignment annotation row - wait for it to be
621 // initialised properly
625 // ensure codon triplet scales with font size
626 consensusAnnotation.scaleColLabel = true;
627 for (int col = 0; col < consensusData.length; col++)
629 Hashtable<String, Object> hci = consensusData[col];
632 // gapped protein column?
635 // array holds #seqs, #ungapped, then codon counts indexed by codon
636 final int[] codonCounts = (int[]) hci.get(PROFILE);
640 * First pass - get total count and find the highest
642 final char[] codons = new char[codonCounts.length - 2];
643 for (int j = 2; j < codonCounts.length; j++)
645 final int codonCount = codonCounts[j];
646 codons[j - 2] = (char) (j - 2);
647 totalCount += codonCount;
651 * Sort array of encoded codons by count ascending - so the modal value
652 * goes to the end; start by copying the count (dropping the first value)
654 int[] sortedCodonCounts = new int[codonCounts.length - 2];
655 System.arraycopy(codonCounts, 2, sortedCodonCounts, 0,
656 codonCounts.length - 2);
657 QuickSort.sort(sortedCodonCounts, codons);
659 int modalCodonEncoded = codons[codons.length - 1];
660 int modalCodonCount = sortedCodonCounts[codons.length - 1];
661 String modalCodon = String
662 .valueOf(CodingUtils.decodeCodon(modalCodonEncoded));
663 if (sortedCodonCounts.length > 1 && sortedCodonCounts[codons.length
664 - 2] == sortedCodonCounts[codons.length - 1])
667 * two or more codons share the modal count
671 float pid = sortedCodonCounts[sortedCodonCounts.length - 1] * 100
672 / (float) totalCount;
675 * todo ? Replace consensus hashtable with sorted arrays of codons and
676 * counts (non-zero only). Include total count in count array [0].
680 * Scan sorted array backwards for most frequent values first. Show
681 * repeated values compactly.
683 StringBuilder mouseOver = new StringBuilder(32);
684 StringBuilder samePercent = new StringBuilder();
685 String percent = null;
686 String lastPercent = null;
687 int percentDecPl = getPercentageDp(nseqs);
689 for (int j = codons.length - 1; j >= 0; j--)
691 int codonCount = sortedCodonCounts[j];
695 * remaining codons are 0% - ignore, but finish off the last one if
698 if (samePercent.length() > 0)
700 mouseOver.append(samePercent).append(": ").append(percent)
705 int codonEncoded = codons[j];
706 final int pct = codonCount * 100 / totalCount;
707 String codon = String
708 .valueOf(CodingUtils.decodeCodon(codonEncoded));
709 StringBuilder sb = new StringBuilder();
710 Format.appendPercentage(sb, pct, percentDecPl);
711 percent = sb.toString();
712 if (showProfileLogo || codonCount == modalCodonCount)
714 if (percent.equals(lastPercent) && j > 0)
716 samePercent.append(samePercent.length() == 0 ? "" : ", ");
717 samePercent.append(codon);
721 if (samePercent.length() > 0)
723 mouseOver.append(samePercent).append(": ").append(lastPercent)
726 samePercent.setLength(0);
727 samePercent.append(codon);
729 lastPercent = percent;
733 consensusAnnotation.annotations[col] = new Annotation(modalCodon,
734 mouseOver.toString(), ' ', pid);
739 * Returns the number of decimal places to show for profile percentages. For
740 * less than 100 sequences, returns zero (the integer percentage value will be
741 * displayed). For 100-999 sequences, returns 1, for 1000-9999 returns 2, etc.
746 protected static int getPercentageDp(long nseq)