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 for (int i = startCol; i < endCol; i++)
257 ProfileI profile = profiles.get(i);
261 * happens if sequences calculated over were
262 * shorter than alignment width
264 consensus.annotations[i] = null;
268 final int dp = getPercentageDp(nseq);
270 float value = profile.getPercentageIdentity(ignoreGaps);
272 String description = getTooltip(profile, value, showSequenceLogo,
275 String modalResidue = profile.getModalResidue();
276 if ("".equals(modalResidue))
280 else if (modalResidue.length() > 1)
284 consensus.annotations[i] = new Annotation(modalResidue, description,
287 // long elapsed = System.currentTimeMillis() - now;
288 // System.out.println(-elapsed);
292 * Derive the gap count annotation row.
295 * the annotation row to add annotations to
297 * the source consensus data
299 * start column (inclusive)
301 * end column (exclusive)
303 public static void completeGapAnnot(AlignmentAnnotation gaprow,
304 ProfilesI profiles, int startCol, int endCol, long nseq)
306 if (gaprow == null || gaprow.annotations == null
307 || gaprow.annotations.length < endCol)
310 * called with a bad alignment annotation row
311 * wait for it to be initialised properly
315 // always set ranges again
316 gaprow.graphMax = nseq;
318 double scale = 0.8 / nseq;
319 for (int i = startCol; i < endCol; i++)
321 ProfileI profile = profiles.get(i);
325 * happens if sequences calculated over were
326 * shorter than alignment width
328 gaprow.annotations[i] = null;
332 final int gapped = profile.getNonGapped();
334 String description = "" + gapped;
336 gaprow.annotations[i] = new Annotation("", description, '\0', gapped,
337 jalview.util.ColorUtils.bleachColour(Color.DARK_GRAY,
338 (float) scale * gapped));
343 * Returns a tooltip showing either
345 * <li>the full profile (percentages of all residues present), if
346 * showSequenceLogo is true, or</li>
347 * <li>just the modal (most common) residue(s), if showSequenceLogo is
350 * Percentages are as a fraction of all sequence, or only ungapped sequences
351 * if ignoreGaps is true.
355 * @param showSequenceLogo
358 * the number of decimal places to format percentages to
361 static String getTooltip(ProfileI profile, float pid,
362 boolean showSequenceLogo, boolean ignoreGaps, int dp)
364 ResidueCount counts = profile.getCounts();
366 String description = null;
367 if (counts != null && showSequenceLogo)
369 int normaliseBy = ignoreGaps ? profile.getNonGapped()
370 : profile.getHeight();
371 description = counts.getTooltip(normaliseBy, dp);
375 StringBuilder sb = new StringBuilder(64);
376 String maxRes = profile.getModalResidue();
377 if (maxRes.length() > 1)
379 sb.append("[").append(maxRes).append("]");
385 if (maxRes.length() > 0)
388 Format.appendPercentage(sb, pid, dp);
391 description = sb.toString();
397 * Returns the sorted profile for the given consensus data. The returned array
401 * [profileType, numberOfValues, totalPercent, charValue1, percentage1, charValue2, percentage2, ...]
402 * in descending order of percentage value
406 * the data object from which to extract and sort values
408 * if true, only non-gapped values are included in percentage
412 public static int[] extractProfile(ProfileI profile, boolean ignoreGaps)
414 ResidueCount counts = profile.getCounts();
420 SymbolCounts symbolCounts = counts.getSymbolCounts();
421 char[] symbols = symbolCounts.symbols;
422 int[] values = symbolCounts.values;
423 QuickSort.sort(values, symbols);
424 int totalPercentage = 0;
425 final int divisor = ignoreGaps ? profile.getNonGapped()
426 : profile.getHeight();
429 * traverse the arrays in reverse order (highest counts first)
431 int[] result = new int[3 + 2 * symbols.length];
432 int nextArrayPos = 3;
433 int nonZeroCount = 0;
435 for (int i = symbols.length - 1; i >= 0; i--)
437 int theChar = symbols[i];
438 int charCount = values[i];
439 final int percentage = (charCount * 100) / divisor;
443 * this count (and any remaining) round down to 0% - discard
448 result[nextArrayPos++] = theChar;
449 result[nextArrayPos++] = percentage;
450 totalPercentage += percentage;
454 * truncate array if any zero values were discarded
456 if (nonZeroCount < symbols.length)
458 int[] tmp = new int[3 + 2 * nonZeroCount];
459 System.arraycopy(result, 0, tmp, 0, tmp.length);
464 * fill in 'header' values
466 result[0] = AlignmentAnnotation.SEQUENCE_PROFILE;
467 result[1] = nonZeroCount;
468 result[2] = totalPercentage;
474 * Extract a sorted extract of cDNA codon profile data. The returned array
478 * [profileType, numberOfValues, totalPercentage, charValue1, percentage1, charValue2, percentage2, ...]
479 * in descending order of percentage value, where the character values encode codon triplets
485 public static int[] extractCdnaProfile(
486 Hashtable<String, Object> hashtable, boolean ignoreGaps)
488 // this holds #seqs, #ungapped, and then codon count, indexed by encoded
490 int[] codonCounts = (int[]) hashtable.get(PROFILE);
491 int[] sortedCounts = new int[codonCounts.length - 2];
492 System.arraycopy(codonCounts, 2, sortedCounts, 0,
493 codonCounts.length - 2);
495 int[] result = new int[3 + 2 * sortedCounts.length];
496 // first value is just the type of profile data
497 result[0] = AlignmentAnnotation.CDNA_PROFILE;
499 char[] codons = new char[sortedCounts.length];
500 for (int i = 0; i < codons.length; i++)
502 codons[i] = (char) i;
504 QuickSort.sort(sortedCounts, codons);
505 int totalPercentage = 0;
506 int distinctValuesCount = 0;
508 int divisor = ignoreGaps ? codonCounts[1] : codonCounts[0];
509 for (int i = codons.length - 1; i >= 0; i--)
511 final int codonCount = sortedCounts[i];
514 break; // nothing else of interest here
516 final int percentage = codonCount * 100 / divisor;
520 * this (and any remaining) values rounded down to 0 - discard
524 distinctValuesCount++;
525 result[j++] = codons[i];
526 result[j++] = percentage;
527 totalPercentage += percentage;
529 result[2] = totalPercentage;
532 * Just return the non-zero values
534 // todo next value is redundant if we limit the array to non-zero counts
535 result[1] = distinctValuesCount;
536 return Arrays.copyOfRange(result, 0, j);
540 * Compute a consensus for the cDNA coding for a protein alignment.
543 * the protein alignment (which should hold mappings to cDNA
546 * the consensus data stores to be populated (one per column)
548 public static void calculateCdna(AlignmentI alignment,
549 Hashtable<String, Object>[] hconsensus)
551 final char gapCharacter = alignment.getGapCharacter();
552 List<AlignedCodonFrame> mappings = alignment.getCodonFrames();
553 if (mappings == null || mappings.isEmpty())
558 int cols = alignment.getWidth();
559 for (int col = 0; col < cols; col++)
561 // todo would prefer a Java bean for consensus data
562 Hashtable<String, Object> columnHash = new Hashtable<>();
563 // #seqs, #ungapped seqs, counts indexed by (codon encoded + 1)
564 int[] codonCounts = new int[66];
565 codonCounts[0] = alignment.getSequences().size();
566 int ungappedCount = 0;
567 for (SequenceI seq : alignment.getSequences())
569 if (seq.getCharAt(col) == gapCharacter)
573 List<char[]> codons = MappingUtils.findCodonsFor(seq, col,
575 for (char[] codon : codons)
577 int codonEncoded = CodingUtils.encodeCodon(codon);
578 if (codonEncoded >= 0)
580 codonCounts[codonEncoded + 2]++;
586 codonCounts[1] = ungappedCount;
587 // todo: sort values here, save counts and codons?
588 columnHash.put(PROFILE, codonCounts);
589 hconsensus[col] = columnHash;
594 * Derive displayable cDNA consensus annotation from computed consensus data.
596 * @param consensusAnnotation
597 * the annotation row to be populated for display
598 * @param consensusData
599 * the computed consensus data
600 * @param showProfileLogo
601 * if true show all symbols present at each position, else only the
604 * the number of sequences in the alignment
606 public static void completeCdnaConsensus(
607 AlignmentAnnotation consensusAnnotation,
608 Hashtable<String, Object>[] consensusData,
609 boolean showProfileLogo, int nseqs)
611 if (consensusAnnotation == null
612 || consensusAnnotation.annotations == null
613 || consensusAnnotation.annotations.length < consensusData.length)
615 // called with a bad alignment annotation row - wait for it to be
616 // initialised properly
620 // ensure codon triplet scales with font size
621 consensusAnnotation.scaleColLabel = true;
622 for (int col = 0; col < consensusData.length; col++)
624 Hashtable<String, Object> hci = consensusData[col];
627 // gapped protein column?
630 // array holds #seqs, #ungapped, then codon counts indexed by codon
631 final int[] codonCounts = (int[]) hci.get(PROFILE);
635 * First pass - get total count and find the highest
637 final char[] codons = new char[codonCounts.length - 2];
638 for (int j = 2; j < codonCounts.length; j++)
640 final int codonCount = codonCounts[j];
641 codons[j - 2] = (char) (j - 2);
642 totalCount += codonCount;
646 * Sort array of encoded codons by count ascending - so the modal value
647 * goes to the end; start by copying the count (dropping the first value)
649 int[] sortedCodonCounts = new int[codonCounts.length - 2];
650 System.arraycopy(codonCounts, 2, sortedCodonCounts, 0,
651 codonCounts.length - 2);
652 QuickSort.sort(sortedCodonCounts, codons);
654 int modalCodonEncoded = codons[codons.length - 1];
655 int modalCodonCount = sortedCodonCounts[codons.length - 1];
656 String modalCodon = String
657 .valueOf(CodingUtils.decodeCodon(modalCodonEncoded));
658 if (sortedCodonCounts.length > 1 && sortedCodonCounts[codons.length
659 - 2] == sortedCodonCounts[codons.length - 1])
662 * two or more codons share the modal count
666 float pid = sortedCodonCounts[sortedCodonCounts.length - 1] * 100
667 / (float) totalCount;
670 * todo ? Replace consensus hashtable with sorted arrays of codons and
671 * counts (non-zero only). Include total count in count array [0].
675 * Scan sorted array backwards for most frequent values first. Show
676 * repeated values compactly.
678 StringBuilder mouseOver = new StringBuilder(32);
679 StringBuilder samePercent = new StringBuilder();
680 String percent = null;
681 String lastPercent = null;
682 int percentDecPl = getPercentageDp(nseqs);
684 for (int j = codons.length - 1; j >= 0; j--)
686 int codonCount = sortedCodonCounts[j];
690 * remaining codons are 0% - ignore, but finish off the last one if
693 if (samePercent.length() > 0)
695 mouseOver.append(samePercent).append(": ").append(percent)
700 int codonEncoded = codons[j];
701 final int pct = codonCount * 100 / totalCount;
702 String codon = String
703 .valueOf(CodingUtils.decodeCodon(codonEncoded));
704 StringBuilder sb = new StringBuilder();
705 Format.appendPercentage(sb, pct, percentDecPl);
706 percent = sb.toString();
707 if (showProfileLogo || codonCount == modalCodonCount)
709 if (percent.equals(lastPercent) && j > 0)
711 samePercent.append(samePercent.length() == 0 ? "" : ", ");
712 samePercent.append(codon);
716 if (samePercent.length() > 0)
718 mouseOver.append(samePercent).append(": ").append(lastPercent)
721 samePercent.setLength(0);
722 samePercent.append(codon);
724 lastPercent = percent;
728 consensusAnnotation.annotations[col] = new Annotation(modalCodon,
729 mouseOver.toString(), ' ', pid);
734 * Returns the number of decimal places to show for profile percentages. For
735 * less than 100 sequences, returns zero (the integer percentage value will be
736 * displayed). For 100-999 sequences, returns 1, for 1000-9999 returns 2, etc.
741 protected static int getPercentageDp(long nseq)