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.ResidueCount;
29 import jalview.datamodel.SequenceI;
30 import jalview.datamodel.ResidueCount.SymbolCounts;
31 import jalview.ext.android.SparseIntArray;
32 import jalview.util.Comparison;
33 import jalview.util.Format;
34 import jalview.util.MappingUtils;
35 import jalview.util.QuickSort;
37 import java.util.Arrays;
38 import java.util.Hashtable;
39 import java.util.List;
42 * Takes in a vector or array of sequences and column start and column end and
43 * returns a new Hashtable[] of size maxSeqLength, if Hashtable not supplied.
44 * This class is used extensively in calculating alignment colourschemes that
45 * depend on the amount of conservation in each alignment column.
50 public class AAFrequency
52 public static final String PROFILE = "P";
55 * Quick look-up of String value of char 'A' to 'Z'
57 private static final String[] CHARS = new String['Z' - 'A' + 1];
61 for (char c = 'A'; c <= 'Z'; c++)
63 CHARS[c - 'A'] = String.valueOf(c);
67 public static final Profile[] calculate(List<SequenceI> list,
70 return calculate(list, start, end, false);
73 public static final Profile[] calculate(List<SequenceI> sequences,
74 int start, int end, boolean profile)
76 SequenceI[] seqs = new SequenceI[sequences.size()];
78 synchronized (sequences)
80 for (int i = 0; i < sequences.size(); i++)
82 seqs[i] = sequences.get(i);
83 if (seqs[i].getLength() > width)
85 width = seqs[i].getLength();
89 Profile[] reply = new Profile[width];
96 calculate(seqs, start, end, reply, profile);
102 * Calculate the consensus symbol(s) for each column in the given range.
106 * start column (inclusive, base zero)
108 * end column (exclusive)
110 * array in which to store profile per column
111 * @param saveFullProfile
112 * if true, store all symbol counts
114 public static final void calculate(final SequenceI[] sequences,
115 int start, int end, Profile[] result, boolean saveFullProfile)
117 // long now = System.currentTimeMillis();
118 int seqCount = sequences.length;
119 boolean nucleotide = false;
120 int nucleotideCount = 0;
121 int peptideCount = 0;
123 for (int column = start; column < end; column++)
126 * Apply a heuristic to detect nucleotide data (which can
127 * be counted in more compact arrays); here we test for
128 * more than 90% nucleotide; recheck every 10 columns in case
129 * of misleading data e.g. highly conserved Alanine in peptide!
130 * Mistakenly guessing nucleotide has a small performance cost,
131 * as it will result in counting in sparse arrays.
132 * Mistakenly guessing peptide has a small space cost,
133 * as it will use a larger than necessary array to hold counts.
135 if (nucleotideCount > 100 && column % 10 == 0)
137 nucleotide = (9 * peptideCount < nucleotideCount);
139 ResidueCount residueCounts = new ResidueCount(nucleotide);
141 for (int row = 0; row < seqCount; row++)
143 if (sequences[row] == null)
146 .println("WARNING: Consensus skipping null sequence - possible race condition.");
149 char[] seq = sequences[row].getSequence();
150 if (seq.length > column)
152 char c = seq[column];
153 residueCounts.add(c);
154 if (Comparison.isNucleotide(c))
158 else if (!Comparison.isGap(c))
166 * count a gap if the sequence doesn't reach this column
168 residueCounts.addGap();
172 int maxCount = residueCounts.getModalCount();
173 String maxResidue = residueCounts.getResiduesForCount(maxCount);
174 int gapCount = residueCounts.getGapCount();
175 Profile profile = new Profile(seqCount, gapCount, maxCount,
180 profile.setCounts(residueCounts);
183 result[column] = profile;
185 // long elapsed = System.currentTimeMillis() - now;
186 // System.out.println(elapsed);
190 * Make an estimate of the profile size we are going to compute i.e. how many
191 * different characters may be present in it. Overestimating has a cost of
192 * using more memory than necessary. Underestimating has a cost of needing to
193 * extend the SparseIntArray holding the profile counts.
195 * @param profileSizes
196 * counts of sizes of profiles so far encountered
199 static int estimateProfileSize(SparseIntArray profileSizes)
201 if (profileSizes.size() == 0)
207 * could do a statistical heuristic here e.g. 75%ile
208 * for now just return the largest value
210 return profileSizes.keyAt(profileSizes.size() - 1);
214 * Derive the consensus annotations to be added to the alignment for display.
215 * This does not recompute the raw data, but may be called on a change in
216 * display options, such as 'ignore gaps', which may in turn result in a
217 * change in the derived values.
220 * the annotation row to add annotations to
222 * the source consensus data
228 * if true, normalise residue percentages ignoring gaps
229 * @param showSequenceLogo
230 * if true include all consensus symbols, else just show modal
233 * number of sequences
235 public static void completeConsensus(AlignmentAnnotation consensus,
236 Profile[] profiles, int iStart, int width, boolean ignoreGaps,
237 boolean showSequenceLogo, long nseq)
239 // long now = System.currentTimeMillis();
240 if (consensus == null || consensus.annotations == null
241 || consensus.annotations.length < width)
244 * called with a bad alignment annotation row
245 * wait for it to be initialised properly
250 final int dp = getPercentageDp(nseq);
252 for (int i = iStart; i < width; i++)
255 if (i >= profiles.length || ((profile = profiles[i]) == null))
258 * happens if sequences calculated over were
259 * shorter than alignment width
261 consensus.annotations[i] = null;
265 float value = profile.getPercentageIdentity(ignoreGaps);
267 String description = getTooltip(profile, value, showSequenceLogo,
270 String modalResidue = profile.getModalResidue();
271 if ("".equals(modalResidue))
275 else if (modalResidue.length() > 1)
279 consensus.annotations[i] = new Annotation(modalResidue,
280 description, ' ', value);
282 // long elapsed = System.currentTimeMillis() - now;
283 // System.out.println(-elapsed);
287 * Returns a tooltip showing either
289 * <li>the full profile (percentages of all residues present), if
290 * showSequenceLogo is true, or</li>
291 * <li>just the modal (most common) residue(s), if showSequenceLogo is false</li>
293 * Percentages are as a fraction of all sequence, or only ungapped sequences
294 * if ignoreGaps is true.
298 * @param showSequenceLogo
301 * the number of decimal places to format percentages to
304 static String getTooltip(Profile profile, float pid,
305 boolean showSequenceLogo, boolean ignoreGaps, int dp)
307 ResidueCount counts = profile.getCounts();
309 String description = null;
310 if (counts != null && showSequenceLogo)
312 int normaliseBy = ignoreGaps ? profile.getNonGapped() : profile
314 description = counts.getTooltip(normaliseBy, dp);
318 StringBuilder sb = new StringBuilder(64);
319 String maxRes = profile.getModalResidue();
320 if (maxRes.length() > 1)
322 sb.append("[").append(maxRes).append("]");
328 if (maxRes.length() > 0)
331 Format.appendPercentage(sb, pid, dp);
334 description = sb.toString();
340 * Returns the sorted profile for the given consensus data. The returned array
344 * [profileType, numberOfValues, nonGapCount, charValue1, percentage1, charValue2, percentage2, ...]
345 * in descending order of percentage value
349 * the data object from which to extract and sort values
351 * if true, only non-gapped values are included in percentage
355 public static int[] extractProfile(Profile profile,
358 int[] rtnval = new int[64];
359 ResidueCount counts = profile.getCounts();
365 SymbolCounts symbolCounts = counts.getSymbolCounts();
366 char[] symbols = symbolCounts.symbols;
367 int[] values = symbolCounts.values;
368 QuickSort.sort(values, symbols);
369 int nextArrayPos = 2;
370 int totalPercentage = 0;
371 final int divisor = ignoreGaps ? profile.getNonGapped() : profile
375 * traverse the arrays in reverse order (highest counts first)
377 for (int i = symbols.length - 1; i >= 0; i--)
379 int theChar = symbols[i];
380 int charCount = values[i];
382 rtnval[nextArrayPos++] = theChar;
383 final int percentage = (charCount * 100) / divisor;
384 rtnval[nextArrayPos++] = percentage;
385 totalPercentage += percentage;
387 rtnval[0] = symbols.length;
388 rtnval[1] = totalPercentage;
389 int[] result = new int[rtnval.length + 1];
390 result[0] = AlignmentAnnotation.SEQUENCE_PROFILE;
391 System.arraycopy(rtnval, 0, result, 1, rtnval.length);
397 * Extract a sorted extract of cDNA codon profile data. The returned array
401 * [profileType, numberOfValues, totalCount, charValue1, percentage1, charValue2, percentage2, ...]
402 * in descending order of percentage value, where the character values encode codon triplets
408 public static int[] extractCdnaProfile(Hashtable hashtable,
411 // this holds #seqs, #ungapped, and then codon count, indexed by encoded
413 int[] codonCounts = (int[]) hashtable.get(PROFILE);
414 int[] sortedCounts = new int[codonCounts.length - 2];
415 System.arraycopy(codonCounts, 2, sortedCounts, 0,
416 codonCounts.length - 2);
418 int[] result = new int[3 + 2 * sortedCounts.length];
419 // first value is just the type of profile data
420 result[0] = AlignmentAnnotation.CDNA_PROFILE;
422 char[] codons = new char[sortedCounts.length];
423 for (int i = 0; i < codons.length; i++)
425 codons[i] = (char) i;
427 QuickSort.sort(sortedCounts, codons);
428 int totalPercentage = 0;
429 int distinctValuesCount = 0;
431 int divisor = ignoreGaps ? codonCounts[1] : codonCounts[0];
432 for (int i = codons.length - 1; i >= 0; i--)
434 final int codonCount = sortedCounts[i];
437 break; // nothing else of interest here
439 distinctValuesCount++;
440 result[j++] = codons[i];
441 final int percentage = codonCount * 100 / divisor;
442 result[j++] = percentage;
443 totalPercentage += percentage;
445 result[2] = totalPercentage;
448 * Just return the non-zero values
450 // todo next value is redundant if we limit the array to non-zero counts
451 result[1] = distinctValuesCount;
452 return Arrays.copyOfRange(result, 0, j);
456 * Compute a consensus for the cDNA coding for a protein alignment.
459 * the protein alignment (which should hold mappings to cDNA
462 * the consensus data stores to be populated (one per column)
464 public static void calculateCdna(AlignmentI alignment,
465 Hashtable[] hconsensus)
467 final char gapCharacter = alignment.getGapCharacter();
468 List<AlignedCodonFrame> mappings = alignment.getCodonFrames();
469 if (mappings == null || mappings.isEmpty())
474 int cols = alignment.getWidth();
475 for (int col = 0; col < cols; col++)
477 // todo would prefer a Java bean for consensus data
478 Hashtable<String, int[]> columnHash = new Hashtable<String, int[]>();
479 // #seqs, #ungapped seqs, counts indexed by (codon encoded + 1)
480 int[] codonCounts = new int[66];
481 codonCounts[0] = alignment.getSequences().size();
482 int ungappedCount = 0;
483 for (SequenceI seq : alignment.getSequences())
485 if (seq.getCharAt(col) == gapCharacter)
489 List<char[]> codons = MappingUtils
490 .findCodonsFor(seq, col, mappings);
491 for (char[] codon : codons)
493 int codonEncoded = CodingUtils.encodeCodon(codon);
494 if (codonEncoded >= 0)
496 codonCounts[codonEncoded + 2]++;
501 codonCounts[1] = ungappedCount;
502 // todo: sort values here, save counts and codons?
503 columnHash.put(PROFILE, codonCounts);
504 hconsensus[col] = columnHash;
509 * Derive displayable cDNA consensus annotation from computed consensus data.
511 * @param consensusAnnotation
512 * the annotation row to be populated for display
513 * @param consensusData
514 * the computed consensus data
515 * @param showProfileLogo
516 * if true show all symbols present at each position, else only the
519 * the number of sequences in the alignment
521 public static void completeCdnaConsensus(
522 AlignmentAnnotation consensusAnnotation,
523 Hashtable[] consensusData, boolean showProfileLogo, int nseqs)
525 if (consensusAnnotation == null
526 || consensusAnnotation.annotations == null
527 || consensusAnnotation.annotations.length < consensusData.length)
529 // called with a bad alignment annotation row - wait for it to be
530 // initialised properly
534 // ensure codon triplet scales with font size
535 consensusAnnotation.scaleColLabel = true;
536 for (int col = 0; col < consensusData.length; col++)
538 Hashtable hci = consensusData[col];
541 // gapped protein column?
544 // array holds #seqs, #ungapped, then codon counts indexed by codon
545 final int[] codonCounts = (int[]) hci.get(PROFILE);
549 * First pass - get total count and find the highest
551 final char[] codons = new char[codonCounts.length - 2];
552 for (int j = 2; j < codonCounts.length; j++)
554 final int codonCount = codonCounts[j];
555 codons[j - 2] = (char) (j - 2);
556 totalCount += codonCount;
560 * Sort array of encoded codons by count ascending - so the modal value
561 * goes to the end; start by copying the count (dropping the first value)
563 int[] sortedCodonCounts = new int[codonCounts.length - 2];
564 System.arraycopy(codonCounts, 2, sortedCodonCounts, 0,
565 codonCounts.length - 2);
566 QuickSort.sort(sortedCodonCounts, codons);
568 int modalCodonEncoded = codons[codons.length - 1];
569 int modalCodonCount = sortedCodonCounts[codons.length - 1];
570 String modalCodon = String.valueOf(CodingUtils
571 .decodeCodon(modalCodonEncoded));
572 if (sortedCodonCounts.length > 1
573 && sortedCodonCounts[codons.length - 2] == sortedCodonCounts[codons.length - 1])
576 * two or more codons share the modal count
580 float pid = sortedCodonCounts[sortedCodonCounts.length - 1] * 100
581 / (float) totalCount;
584 * todo ? Replace consensus hashtable with sorted arrays of codons and
585 * counts (non-zero only). Include total count in count array [0].
589 * Scan sorted array backwards for most frequent values first. Show
590 * repeated values compactly.
592 StringBuilder mouseOver = new StringBuilder(32);
593 StringBuilder samePercent = new StringBuilder();
594 String percent = null;
595 String lastPercent = null;
596 int percentDecPl = getPercentageDp(nseqs);
598 for (int j = codons.length - 1; j >= 0; j--)
600 int codonCount = sortedCodonCounts[j];
604 * remaining codons are 0% - ignore, but finish off the last one if
607 if (samePercent.length() > 0)
609 mouseOver.append(samePercent).append(": ").append(percent)
614 int codonEncoded = codons[j];
615 final int pct = codonCount * 100 / totalCount;
616 String codon = String
617 .valueOf(CodingUtils.decodeCodon(codonEncoded));
618 StringBuilder sb = new StringBuilder();
619 Format.appendPercentage(sb, pct, percentDecPl);
620 percent = sb.toString();
621 if (showProfileLogo || codonCount == modalCodonCount)
623 if (percent.equals(lastPercent) && j > 0)
625 samePercent.append(samePercent.length() == 0 ? "" : ", ");
626 samePercent.append(codon);
630 if (samePercent.length() > 0)
632 mouseOver.append(samePercent).append(": ")
633 .append(lastPercent).append("% ");
635 samePercent.setLength(0);
636 samePercent.append(codon);
638 lastPercent = percent;
642 consensusAnnotation.annotations[col] = new Annotation(modalCodon,
643 mouseOver.toString(), ' ', pid);
648 * Returns the number of decimal places to show for profile percentages. For
649 * less than 100 sequences, returns zero (the integer percentage value will be
650 * displayed). For 100-999 sequences, returns 1, for 1000-9999 returns 2, etc.
655 protected static int getPercentageDp(long nseq)