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.ResidueCount;
30 import jalview.datamodel.SequenceI;
31 import jalview.datamodel.ResidueCount.SymbolCounts;
32 import jalview.ext.android.SparseIntArray;
33 import jalview.util.Comparison;
34 import jalview.util.Format;
35 import jalview.util.MappingUtils;
36 import jalview.util.QuickSort;
38 import java.util.Arrays;
39 import java.util.Hashtable;
40 import java.util.List;
43 * Takes in a vector or array of sequences and column start and column end and
44 * returns a new Hashtable[] of size maxSeqLength, if Hashtable not supplied.
45 * This class is used extensively in calculating alignment colourschemes that
46 * depend on the amount of conservation in each alignment column.
51 public class AAFrequency
53 public static final String PROFILE = "P";
56 * Quick look-up of String value of char 'A' to 'Z'
58 private static final String[] CHARS = new String['Z' - 'A' + 1];
62 for (char c = 'A'; c <= 'Z'; c++)
64 CHARS[c - 'A'] = String.valueOf(c);
68 public static final ProfileI[] calculate(List<SequenceI> list,
71 return calculate(list, start, end, false);
74 public static final ProfileI[] calculate(List<SequenceI> sequences,
75 int start, int end, boolean profile)
77 SequenceI[] seqs = new SequenceI[sequences.size()];
79 synchronized (sequences)
81 for (int i = 0; i < sequences.size(); i++)
83 seqs[i] = sequences.get(i);
84 if (seqs[i].getLength() > width)
86 width = seqs[i].getLength();
90 ProfileI[] reply = new ProfileI[width];
97 calculate(seqs, start, end, reply, profile);
103 * Calculate the consensus symbol(s) for each column in the given range.
107 * start column (inclusive, base zero)
109 * end column (exclusive)
111 * array in which to store profile per column
112 * @param saveFullProfile
113 * if true, store all symbol counts
115 public static final void calculate(final SequenceI[] sequences,
116 int start, int end, ProfileI[] result, boolean saveFullProfile)
118 // long now = System.currentTimeMillis();
119 int seqCount = sequences.length;
120 boolean nucleotide = false;
121 int nucleotideCount = 0;
122 int peptideCount = 0;
124 for (int column = start; column < end; column++)
127 * Apply a heuristic to detect nucleotide data (which can
128 * be counted in more compact arrays); here we test for
129 * more than 90% nucleotide; recheck every 10 columns in case
130 * of misleading data e.g. highly conserved Alanine in peptide!
131 * Mistakenly guessing nucleotide has a small performance cost,
132 * as it will result in counting in sparse arrays.
133 * Mistakenly guessing peptide has a small space cost,
134 * as it will use a larger than necessary array to hold counts.
136 if (nucleotideCount > 100 && column % 10 == 0)
138 nucleotide = (9 * peptideCount < nucleotideCount);
140 ResidueCount residueCounts = new ResidueCount(nucleotide);
142 for (int row = 0; row < seqCount; row++)
144 if (sequences[row] == null)
147 .println("WARNING: Consensus skipping null sequence - possible race condition.");
150 char[] seq = sequences[row].getSequence();
151 if (seq.length > column)
153 char c = seq[column];
154 residueCounts.add(c);
155 if (Comparison.isNucleotide(c))
159 else if (!Comparison.isGap(c))
167 * count a gap if the sequence doesn't reach this column
169 residueCounts.addGap();
173 int maxCount = residueCounts.getModalCount();
174 String maxResidue = residueCounts.getResiduesForCount(maxCount);
175 int gapCount = residueCounts.getGapCount();
176 ProfileI profile = new Profile(seqCount, gapCount, maxCount,
181 profile.setCounts(residueCounts);
184 result[column] = profile;
186 // long elapsed = System.currentTimeMillis() - now;
187 // System.out.println(elapsed);
191 * Make an estimate of the profile size we are going to compute i.e. how many
192 * different characters may be present in it. Overestimating has a cost of
193 * using more memory than necessary. Underestimating has a cost of needing to
194 * extend the SparseIntArray holding the profile counts.
196 * @param profileSizes
197 * counts of sizes of profiles so far encountered
200 static int estimateProfileSize(SparseIntArray profileSizes)
202 if (profileSizes.size() == 0)
208 * could do a statistical heuristic here e.g. 75%ile
209 * for now just return the largest value
211 return profileSizes.keyAt(profileSizes.size() - 1);
215 * Derive the consensus annotations to be added to the alignment for display.
216 * This does not recompute the raw data, but may be called on a change in
217 * display options, such as 'ignore gaps', which may in turn result in a
218 * change in the derived values.
221 * the annotation row to add annotations to
223 * the source consensus data
229 * if true, normalise residue percentages ignoring gaps
230 * @param showSequenceLogo
231 * if true include all consensus symbols, else just show modal
234 * number of sequences
236 public static void completeConsensus(AlignmentAnnotation consensus,
237 ProfileI[] profiles, int iStart, int width, boolean ignoreGaps,
238 boolean showSequenceLogo, long nseq)
240 // long now = System.currentTimeMillis();
241 if (consensus == null || consensus.annotations == null
242 || consensus.annotations.length < width)
245 * called with a bad alignment annotation row
246 * wait for it to be initialised properly
251 final int dp = getPercentageDp(nseq);
253 for (int i = iStart; i < width; i++)
256 if (i >= profiles.length || ((profile = profiles[i]) == null))
259 * happens if sequences calculated over were
260 * shorter than alignment width
262 consensus.annotations[i] = null;
266 float value = profile.getPercentageIdentity(ignoreGaps);
268 String description = getTooltip(profile, value, showSequenceLogo,
271 String modalResidue = profile.getModalResidue();
272 if ("".equals(modalResidue))
276 else if (modalResidue.length() > 1)
280 consensus.annotations[i] = new Annotation(modalResidue,
281 description, ' ', value);
283 // long elapsed = System.currentTimeMillis() - now;
284 // System.out.println(-elapsed);
288 * Returns a tooltip showing either
290 * <li>the full profile (percentages of all residues present), if
291 * showSequenceLogo is true, or</li>
292 * <li>just the modal (most common) residue(s), if showSequenceLogo is false</li>
294 * Percentages are as a fraction of all sequence, or only ungapped sequences
295 * if ignoreGaps is true.
299 * @param showSequenceLogo
302 * the number of decimal places to format percentages to
305 static String getTooltip(ProfileI profile, float pid,
306 boolean showSequenceLogo, boolean ignoreGaps, int dp)
308 ResidueCount counts = profile.getCounts();
310 String description = null;
311 if (counts != null && showSequenceLogo)
313 int normaliseBy = ignoreGaps ? profile.getNonGapped() : profile
315 description = counts.getTooltip(normaliseBy, dp);
319 StringBuilder sb = new StringBuilder(64);
320 String maxRes = profile.getModalResidue();
321 if (maxRes.length() > 1)
323 sb.append("[").append(maxRes).append("]");
329 if (maxRes.length() > 0)
332 Format.appendPercentage(sb, pid, dp);
335 description = sb.toString();
341 * Returns the sorted profile for the given consensus data. The returned array
345 * [profileType, numberOfValues, nonGapCount, charValue1, percentage1, charValue2, percentage2, ...]
346 * in descending order of percentage value
350 * the data object from which to extract and sort values
352 * if true, only non-gapped values are included in percentage
356 public static int[] extractProfile(ProfileI profile,
359 int[] rtnval = new int[64];
360 ResidueCount counts = profile.getCounts();
366 SymbolCounts symbolCounts = counts.getSymbolCounts();
367 char[] symbols = symbolCounts.symbols;
368 int[] values = symbolCounts.values;
369 QuickSort.sort(values, symbols);
370 int nextArrayPos = 2;
371 int totalPercentage = 0;
372 final int divisor = ignoreGaps ? profile.getNonGapped() : profile
376 * traverse the arrays in reverse order (highest counts first)
378 for (int i = symbols.length - 1; i >= 0; i--)
380 int theChar = symbols[i];
381 int charCount = values[i];
383 rtnval[nextArrayPos++] = theChar;
384 final int percentage = (charCount * 100) / divisor;
385 rtnval[nextArrayPos++] = percentage;
386 totalPercentage += percentage;
388 rtnval[0] = symbols.length;
389 rtnval[1] = totalPercentage;
390 int[] result = new int[rtnval.length + 1];
391 result[0] = AlignmentAnnotation.SEQUENCE_PROFILE;
392 System.arraycopy(rtnval, 0, result, 1, rtnval.length);
398 * Extract a sorted extract of cDNA codon profile data. The returned array
402 * [profileType, numberOfValues, totalCount, charValue1, percentage1, charValue2, percentage2, ...]
403 * in descending order of percentage value, where the character values encode codon triplets
409 public static int[] extractCdnaProfile(Hashtable hashtable,
412 // this holds #seqs, #ungapped, and then codon count, indexed by encoded
414 int[] codonCounts = (int[]) hashtable.get(PROFILE);
415 int[] sortedCounts = new int[codonCounts.length - 2];
416 System.arraycopy(codonCounts, 2, sortedCounts, 0,
417 codonCounts.length - 2);
419 int[] result = new int[3 + 2 * sortedCounts.length];
420 // first value is just the type of profile data
421 result[0] = AlignmentAnnotation.CDNA_PROFILE;
423 char[] codons = new char[sortedCounts.length];
424 for (int i = 0; i < codons.length; i++)
426 codons[i] = (char) i;
428 QuickSort.sort(sortedCounts, codons);
429 int totalPercentage = 0;
430 int distinctValuesCount = 0;
432 int divisor = ignoreGaps ? codonCounts[1] : codonCounts[0];
433 for (int i = codons.length - 1; i >= 0; i--)
435 final int codonCount = sortedCounts[i];
438 break; // nothing else of interest here
440 distinctValuesCount++;
441 result[j++] = codons[i];
442 final int percentage = codonCount * 100 / divisor;
443 result[j++] = percentage;
444 totalPercentage += percentage;
446 result[2] = totalPercentage;
449 * Just return the non-zero values
451 // todo next value is redundant if we limit the array to non-zero counts
452 result[1] = distinctValuesCount;
453 return Arrays.copyOfRange(result, 0, j);
457 * Compute a consensus for the cDNA coding for a protein alignment.
460 * the protein alignment (which should hold mappings to cDNA
463 * the consensus data stores to be populated (one per column)
465 public static void calculateCdna(AlignmentI alignment,
466 Hashtable[] hconsensus)
468 final char gapCharacter = alignment.getGapCharacter();
469 List<AlignedCodonFrame> mappings = alignment.getCodonFrames();
470 if (mappings == null || mappings.isEmpty())
475 int cols = alignment.getWidth();
476 for (int col = 0; col < cols; col++)
478 // todo would prefer a Java bean for consensus data
479 Hashtable<String, int[]> columnHash = new Hashtable<String, int[]>();
480 // #seqs, #ungapped seqs, counts indexed by (codon encoded + 1)
481 int[] codonCounts = new int[66];
482 codonCounts[0] = alignment.getSequences().size();
483 int ungappedCount = 0;
484 for (SequenceI seq : alignment.getSequences())
486 if (seq.getCharAt(col) == gapCharacter)
490 List<char[]> codons = MappingUtils
491 .findCodonsFor(seq, col, mappings);
492 for (char[] codon : codons)
494 int codonEncoded = CodingUtils.encodeCodon(codon);
495 if (codonEncoded >= 0)
497 codonCounts[codonEncoded + 2]++;
502 codonCounts[1] = ungappedCount;
503 // todo: sort values here, save counts and codons?
504 columnHash.put(PROFILE, codonCounts);
505 hconsensus[col] = columnHash;
510 * Derive displayable cDNA consensus annotation from computed consensus data.
512 * @param consensusAnnotation
513 * the annotation row to be populated for display
514 * @param consensusData
515 * the computed consensus data
516 * @param showProfileLogo
517 * if true show all symbols present at each position, else only the
520 * the number of sequences in the alignment
522 public static void completeCdnaConsensus(
523 AlignmentAnnotation consensusAnnotation,
524 Hashtable[] consensusData, boolean showProfileLogo, int nseqs)
526 if (consensusAnnotation == null
527 || consensusAnnotation.annotations == null
528 || consensusAnnotation.annotations.length < consensusData.length)
530 // called with a bad alignment annotation row - wait for it to be
531 // initialised properly
535 // ensure codon triplet scales with font size
536 consensusAnnotation.scaleColLabel = true;
537 for (int col = 0; col < consensusData.length; col++)
539 Hashtable hci = consensusData[col];
542 // gapped protein column?
545 // array holds #seqs, #ungapped, then codon counts indexed by codon
546 final int[] codonCounts = (int[]) hci.get(PROFILE);
550 * First pass - get total count and find the highest
552 final char[] codons = new char[codonCounts.length - 2];
553 for (int j = 2; j < codonCounts.length; j++)
555 final int codonCount = codonCounts[j];
556 codons[j - 2] = (char) (j - 2);
557 totalCount += codonCount;
561 * Sort array of encoded codons by count ascending - so the modal value
562 * goes to the end; start by copying the count (dropping the first value)
564 int[] sortedCodonCounts = new int[codonCounts.length - 2];
565 System.arraycopy(codonCounts, 2, sortedCodonCounts, 0,
566 codonCounts.length - 2);
567 QuickSort.sort(sortedCodonCounts, codons);
569 int modalCodonEncoded = codons[codons.length - 1];
570 int modalCodonCount = sortedCodonCounts[codons.length - 1];
571 String modalCodon = String.valueOf(CodingUtils
572 .decodeCodon(modalCodonEncoded));
573 if (sortedCodonCounts.length > 1
574 && sortedCodonCounts[codons.length - 2] == sortedCodonCounts[codons.length - 1])
577 * two or more codons share the modal count
581 float pid = sortedCodonCounts[sortedCodonCounts.length - 1] * 100
582 / (float) totalCount;
585 * todo ? Replace consensus hashtable with sorted arrays of codons and
586 * counts (non-zero only). Include total count in count array [0].
590 * Scan sorted array backwards for most frequent values first. Show
591 * repeated values compactly.
593 StringBuilder mouseOver = new StringBuilder(32);
594 StringBuilder samePercent = new StringBuilder();
595 String percent = null;
596 String lastPercent = null;
597 int percentDecPl = getPercentageDp(nseqs);
599 for (int j = codons.length - 1; j >= 0; j--)
601 int codonCount = sortedCodonCounts[j];
605 * remaining codons are 0% - ignore, but finish off the last one if
608 if (samePercent.length() > 0)
610 mouseOver.append(samePercent).append(": ").append(percent)
615 int codonEncoded = codons[j];
616 final int pct = codonCount * 100 / totalCount;
617 String codon = String
618 .valueOf(CodingUtils.decodeCodon(codonEncoded));
619 StringBuilder sb = new StringBuilder();
620 Format.appendPercentage(sb, pct, percentDecPl);
621 percent = sb.toString();
622 if (showProfileLogo || codonCount == modalCodonCount)
624 if (percent.equals(lastPercent) && j > 0)
626 samePercent.append(samePercent.length() == 0 ? "" : ", ");
627 samePercent.append(codon);
631 if (samePercent.length() > 0)
633 mouseOver.append(samePercent).append(": ")
634 .append(lastPercent).append("% ");
636 samePercent.setLength(0);
637 samePercent.append(codon);
639 lastPercent = percent;
643 consensusAnnotation.annotations[col] = new Annotation(modalCodon,
644 mouseOver.toString(), ' ', pid);
649 * Returns the number of decimal places to show for profile percentages. For
650 * less than 100 sequences, returns zero (the integer percentage value will be
651 * displayed). For 100-999 sequences, returns 1, for 1000-9999 returns 2, etc.
656 protected static int getPercentageDp(long nseq)