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.
23 import java.util.ArrayList;
24 import java.util.Arrays;
25 import java.util.List;
28 * A simple way of bijectively mapping a non-contiguous linear range to another
29 * non-contiguous linear range.
31 * Use at your own risk!
33 * TODO: efficient implementation of private posMap method
35 * TODO: test/ensure that sense of from and to ratio start position is conserved
36 * (codon start position recovery)
42 * Subregions (base 1) described as { [start1, end1], [start2, end2], ...}
44 private List<int[]> fromShifts;
47 * Same format as fromShifts, for the 'mapped to' sequence
49 private List<int[]> toShifts;
52 * number of steps in fromShifts to one toRatio unit
54 private int fromRatio;
57 * number of steps in toShifts to one fromRatio
62 * lowest and highest value in the from Map
64 private int fromLowest;
66 private int fromHighest;
69 * lowest and highest value in the to Map
73 private int toHighest;
80 fromShifts = new ArrayList<>();
81 toShifts = new ArrayList<>();
85 * Two MapList objects are equal if they are the same object, or they both
86 * have populated shift ranges and all values are the same.
89 public boolean equals(Object o)
91 if (o == null || !(o instanceof MapList))
96 MapList obj = (MapList) o;
101 if (obj.fromRatio != fromRatio || obj.toRatio != toRatio
102 || obj.fromShifts == null || obj.toShifts == null)
106 return Arrays.deepEquals(fromShifts.toArray(), obj.fromShifts.toArray())
107 && Arrays.deepEquals(toShifts.toArray(),
108 obj.toShifts.toArray());
112 * Returns a hashcode made from the fromRatio, toRatio, and from/to ranges
115 public int hashCode()
117 int hashCode = 31 * fromRatio;
118 hashCode = 31 * hashCode + toRatio;
119 for (int[] shift : fromShifts)
121 hashCode = 31 * hashCode + shift[0];
122 hashCode = 31 * hashCode + shift[1];
124 for (int[] shift : toShifts)
126 hashCode = 31 * hashCode + shift[0];
127 hashCode = 31 * hashCode + shift[1];
134 * Returns the 'from' ranges as {[start1, end1], [start2, end2], ...}
138 public List<int[]> getFromRanges()
144 * Returns the 'to' ranges as {[start1, end1], [start2, end2], ...}
148 public List<int[]> getToRanges()
154 * Flattens a list of [start, end] into a single [start1, end1, start2,
160 protected static int[] getRanges(List<int[]> shifts)
162 int[] rnges = new int[2 * shifts.size()];
164 for (int[] r : shifts)
174 * @return length of mapped phrase in from
176 public int getFromRatio()
183 * @return length of mapped phrase in to
185 public int getToRatio()
190 public int getFromLowest()
195 public int getFromHighest()
200 public int getToLowest()
205 public int getToHighest()
211 * Constructor given from and to ranges as [start1, end1, start2, end2,...].
212 * There is no validation check that the ranges do not overlap each other.
215 * contiguous regions as [start1, end1, start2, end2, ...]
217 * same format as 'from'
219 * phrase length in 'from' (e.g. 3 for dna)
221 * phrase length in 'to' (e.g. 1 for protein)
223 public MapList(int from[], int to[], int fromRatio, int toRatio)
226 this.fromRatio = fromRatio;
227 this.toRatio = toRatio;
228 fromLowest = Integer.MAX_VALUE;
229 fromHighest = Integer.MIN_VALUE;
231 for (int i = 0; i < from.length; i += 2)
234 * note lowest and highest values - bearing in mind the
235 * direction may be reversed
237 fromLowest = Math.min(fromLowest, Math.min(from[i], from[i + 1]));
238 fromHighest = Math.max(fromHighest, Math.max(from[i], from[i + 1]));
239 fromShifts.add(new int[] { from[i], from[i + 1] });
242 toLowest = Integer.MAX_VALUE;
243 toHighest = Integer.MIN_VALUE;
244 for (int i = 0; i < to.length; i += 2)
246 toLowest = Math.min(toLowest, Math.min(to[i], to[i + 1]));
247 toHighest = Math.max(toHighest, Math.max(to[i], to[i + 1]));
248 toShifts.add(new int[] { to[i], to[i + 1] });
253 * Copy constructor. Creates an identical mapping.
257 public MapList(MapList map)
260 // TODO not used - remove?
261 this.fromLowest = map.fromLowest;
262 this.fromHighest = map.fromHighest;
263 this.toLowest = map.toLowest;
264 this.toHighest = map.toHighest;
266 this.fromRatio = map.fromRatio;
267 this.toRatio = map.toRatio;
268 if (map.fromShifts != null)
270 for (int[] r : map.fromShifts)
272 fromShifts.add(new int[] { r[0], r[1] });
275 if (map.toShifts != null)
277 for (int[] r : map.toShifts)
279 toShifts.add(new int[] { r[0], r[1] });
285 * Constructor given ranges as lists of [start, end] positions. There is no
286 * validation check that the ranges do not overlap each other.
293 public MapList(List<int[]> fromRange, List<int[]> toRange, int fromRatio,
297 fromRange = coalesceRanges(fromRange);
298 toRange = coalesceRanges(toRange);
299 this.fromShifts = fromRange;
300 this.toShifts = toRange;
301 this.fromRatio = fromRatio;
302 this.toRatio = toRatio;
304 fromLowest = Integer.MAX_VALUE;
305 fromHighest = Integer.MIN_VALUE;
306 for (int[] range : fromRange)
308 if (range.length != 2)
310 // throw new IllegalArgumentException(range);
311 System.err.println("Invalid format for fromRange "
312 + Arrays.toString(range) + " may cause errors");
314 fromLowest = Math.min(fromLowest, Math.min(range[0], range[1]));
315 fromHighest = Math.max(fromHighest, Math.max(range[0], range[1]));
318 toLowest = Integer.MAX_VALUE;
319 toHighest = Integer.MIN_VALUE;
320 for (int[] range : toRange)
322 if (range.length != 2)
324 // throw new IllegalArgumentException(range);
325 System.err.println("Invalid format for toRange "
326 + Arrays.toString(range) + " may cause errors");
328 toLowest = Math.min(toLowest, Math.min(range[0], range[1]));
329 toHighest = Math.max(toHighest, Math.max(range[0], range[1]));
334 * Consolidates a list of ranges so that any contiguous ranges are merged.
335 * This assumes the ranges are already in start order (does not sort them).
337 * The main use case for this method is when mapping cDNA sequence to its
338 * protein product, based on CDS feature ranges which derive from spliced
339 * exons, but are contiguous on the cDNA sequence. For example
342 * CDS 1-20 // from exon1
343 * CDS 21-35 // from exon2
344 * CDS 36-71 // from exon3
345 * 'coalesce' to range 1-71
349 * @return the same list (if unchanged), else a new merged list, leaving the
350 * input list unchanged
352 public static List<int[]> coalesceRanges(final List<int[]> ranges)
354 if (ranges == null || ranges.size() < 2)
359 boolean changed = false;
360 List<int[]> merged = new ArrayList<>();
361 int[] lastRange = ranges.get(0);
362 int lastDirection = lastRange[1] >= lastRange[0] ? 1 : -1;
363 lastRange = new int[] { lastRange[0], lastRange[1] };
364 merged.add(lastRange);
365 boolean first = true;
367 for (final int[] range : ranges)
375 int direction = range[1] >= range[0] ? 1 : -1;
378 * if next range is in the same direction as last and contiguous,
379 * just update the end position of the last range
381 boolean sameDirection = range[1] == range[0]
382 || direction == lastDirection;
383 boolean extending = range[0] == lastRange[1] + lastDirection;
384 if (sameDirection && extending)
386 lastRange[1] = range[1];
391 lastRange = new int[] { range[0], range[1] };
392 merged.add(lastRange);
393 // careful: merging [5, 5] after [7, 6] should keep negative direction
394 lastDirection = (range[1] == range[0]) ? lastDirection : direction;
398 return changed ? merged : ranges;
402 * get all mapped positions from 'from' to 'to'
404 * @return int[][] { int[] { fromStart, fromFinish, toStart, toFinish }, int
405 * [fromFinish-fromStart+2] { toStart..toFinish mappings}}
407 protected int[][] makeFromMap()
409 // TODO not used - remove??
410 return posMap(fromShifts, fromRatio, toShifts, toRatio);
414 * get all mapped positions from 'to' to 'from'
416 * @return int[to position]=position mapped in from
418 protected int[][] makeToMap()
420 // TODO not used - remove??
421 return posMap(toShifts, toRatio, fromShifts, fromRatio);
425 * construct an int map for intervals in intVals
428 * @return int[] { from, to pos in range }, int[range.to-range.from+1]
429 * returning mapped position
431 private int[][] posMap(List<int[]> shiftTo, int ratio,
432 List<int[]> shiftFrom, int toRatio)
434 // TODO not used - remove??
435 int iv = 0, ivSize = shiftTo.size();
440 int[] intv = shiftTo.get(iv++);
441 int from = intv[0], to = intv[1];
449 intv = shiftTo.get(iv++);
468 int mp[][] = new int[to - from + 2][];
469 for (int i = 0; i < mp.length; i++)
471 int[] m = shift(i + from, shiftTo, ratio, shiftFrom, toRatio);
492 int[][] map = new int[][] { new int[] { from, to, tF, tT },
493 new int[to - from + 2] };
498 for (int i = 0; i < mp.length; i++)
502 map[1][i] = mp[i][0] - tF;
506 map[1][i] = -1; // indicates an out of range mapping
516 * start position for shift (in original reference frame)
520 * public void addShift(int pos, int shift) { int sidx = 0; int[]
521 * rshift=null; while (sidx<shifts.size() && (rshift=(int[])
522 * shifts.elementAt(sidx))[0]<pos) sidx++; if (sidx==shifts.size())
523 * shifts.insertElementAt(new int[] { pos, shift}, sidx); else
524 * rshift[1]+=shift; }
528 * shift from pos to To(pos)
532 * @return int shifted position in To, frameshift in From, direction of mapped
535 public int[] shiftFrom(int pos)
537 return shift(pos, fromShifts, fromRatio, toShifts, toRatio);
541 * inverse of shiftFrom - maps pos in To to a position in From
545 * @return shifted position in From, frameshift in To, direction of mapped
548 public int[] shiftTo(int pos)
550 return shift(pos, toShifts, toRatio, fromShifts, fromRatio);
561 protected static int[] shift(int pos, List<int[]> shiftTo, int fromRatio,
562 List<int[]> shiftFrom, int toRatio)
564 // TODO: javadoc; tests
565 int[] fromCount = countPos(shiftTo, pos);
566 if (fromCount == null)
570 int fromRemainder = (fromCount[0] - 1) % fromRatio;
571 int toCount = 1 + (((fromCount[0] - 1) / fromRatio) * toRatio);
572 int[] toPos = traverseToPosition(shiftFrom, toCount);
575 return null; // throw new Error("Bad Mapping!");
577 // System.out.println(fromCount[0]+" "+fromCount[1]+" "+toCount);
578 return new int[] { toPos[0], fromRemainder, toPos[1] };
582 * count how many positions pos is along the series of intervals.
586 * @return number of positions or null if pos is not within intervals
588 protected static int[] countPos(List<int[]> shiftTo, int pos)
590 int count = 0, intv[], iv = 0, ivSize = shiftTo.size();
593 intv = shiftTo.get(iv++);
594 if (intv[0] <= intv[1])
596 if (pos >= intv[0] && pos <= intv[1])
598 return new int[] { count + pos - intv[0] + 1, +1 };
602 count += intv[1] - intv[0] + 1;
607 if (pos >= intv[1] && pos <= intv[0])
609 return new int[] { count + intv[0] - pos + 1, -1 };
613 count += intv[0] - intv[1] + 1;
621 * Reads through the given intervals until {@code count} positions have been
622 * traversed, and returns an array consisting of two values:
624 * <li>the value at the {@code count'th} position</li>
625 * <li>+1 if the last interval read is forwards, -1 if reverse direction</li>
627 * Returns null if the ranges include less than {@code count} positions, or if
631 * a list of [start, end] ranges
633 * the number of positions to traverse
636 protected static int[] traverseToPosition(List<int[]> intervals,
640 int ivSize = intervals.size();
650 int[] intv = intervals.get(iv++);
651 int diff = intv[1] - intv[0];
654 if (count <= traversed + 1 + diff)
656 return new int[] { intv[0] + (count - traversed - 1), +1 };
660 traversed += 1 + diff;
665 if (count <= traversed + 1 - diff)
667 return new int[] { intv[0] - (count - traversed - 1), -1 };
671 traversed += 1 - diff;
679 * find series of intervals mapping from start-end in the From map.
682 * position mapped 'to'
684 * position mapped 'to'
685 * @return series of [start, end] ranges in sequence mapped 'from'
687 public int[] locateInFrom(int start, int end)
689 // inefficient implementation
690 int fromStart[] = shiftTo(start);
691 // needs to be inclusive of end of symbol position
692 int fromEnd[] = shiftTo(end);
694 return getIntervals(fromShifts, fromStart, fromEnd, fromRatio);
698 * find series of intervals mapping from start-end in the to map.
701 * position mapped 'from'
703 * position mapped 'from'
704 * @return series of [start, end] ranges in sequence mapped 'to'
706 public int[] locateInTo(int start, int end)
708 int toStart[] = shiftFrom(start);
709 int toEnd[] = shiftFrom(end);
710 return getIntervals(toShifts, toStart, toEnd, toRatio);
714 * like shift - except returns the intervals in the given vector of shifts
715 * which were spanned in traversing fromStart to fromEnd
721 * @return series of from,to intervals from from first position of starting
722 * region to final position of ending region inclusive
724 protected static int[] getIntervals(List<int[]> shiftFrom,
725 int[] fromStart, int[] fromEnd, int fromRatio2)
727 if (fromStart == null || fromEnd == null)
731 int startpos, endpos;
732 startpos = fromStart[0]; // first position in fromStart
733 endpos = fromEnd[0]; // last position in fromEnd
734 int endindx = (fromRatio2 - 1); // additional positions to get to last
735 // position from endpos
736 int intv = 0, intvSize = shiftFrom.size();
737 int iv[], i = 0, fs = -1, fe_s = -1, fe = -1; // containing intervals
738 // search intervals to locate ones containing startpos and count endindx
739 // positions on from endpos
740 while (intv < intvSize && (fs == -1 || fe == -1))
742 iv = shiftFrom.get(intv++);
745 endpos = iv[0]; // start counting from beginning of interval
746 endindx--; // inclusive of endpos
750 if (fs == -1 && startpos >= iv[0] && startpos <= iv[1])
754 if (endpos >= iv[0] && endpos <= iv[1])
762 if (endpos + endindx <= iv[1])
765 endpos = endpos + endindx; // end of end token is within this
770 endindx -= iv[1] - endpos; // skip all this interval too
777 if (fs == -1 && startpos <= iv[0] && startpos >= iv[1])
781 if (endpos <= iv[0] && endpos >= iv[1])
789 if (endpos - endindx >= iv[1])
792 endpos = endpos - endindx; // end of end token is within this
797 endindx -= endpos - iv[1]; // skip all this interval too
804 if (fs == fe && fe == -1)
808 List<int[]> ranges = new ArrayList<>();
813 // truncate initial interval
814 iv = shiftFrom.get(intv++);
815 iv = new int[] { iv[0], iv[1] };// clone
822 ranges.add(iv); // add initial range
823 iv = shiftFrom.get(intv++); // get next interval
824 iv = new int[] { iv[0], iv[1] };// clone
831 ranges.add(iv); // add only - or final range
835 // walk from end of interval.
836 i = shiftFrom.size() - 1;
841 iv = shiftFrom.get(i);
842 iv = new int[] { iv[1], iv[0] };// reverse and clone
843 // truncate initial interval
849 { // fix apparent logic bug when fe==-1
850 ranges.add(iv); // add (truncated) reversed interval
851 iv = shiftFrom.get(i);
852 iv = new int[] { iv[1], iv[0] }; // reverse and clone
856 // interval is already reversed
859 ranges.add(iv); // add only - or final range
861 // create array of start end intervals.
863 if (ranges != null && ranges.size() > 0)
865 range = new int[ranges.size() * 2];
867 intvSize = ranges.size();
869 while (intv < intvSize)
871 iv = ranges.get(intv);
874 ranges.set(intv++, null); // remove
881 * get the 'initial' position of mpos in To
885 * @return position of first word in to reference frame
887 public int getToPosition(int mpos)
889 // TODO not used - remove??
890 int[] mp = shiftTo(mpos);
899 * get range of positions in To frame for the mpos word in From
903 * @return null or int[] first position in To for mpos, last position in to
906 public int[] getToWord(int mpos)
908 int[] mp = shiftTo(mpos);
911 return new int[] { mp[0], mp[0] + mp[2] * (getFromRatio() - 1) };
917 * get From position in the associated reference frame for position pos in the
918 * associated sequence.
923 public int getMappedPosition(int pos)
925 // TODO not used - remove??
926 int[] mp = shiftFrom(pos);
934 public int[] getMappedWord(int pos)
936 // TODO not used - remove??
937 int[] mp = shiftFrom(pos);
940 return new int[] { mp[0], mp[0] + mp[2] * (getToRatio() - 1) };
947 * @return a MapList whose From range is this maplist's To Range, and vice
950 public MapList getInverse()
952 return new MapList(getToRanges(), getFromRanges(), getToRatio(),
957 * test for containment rather than equivalence to another mapping
960 * to be tested for containment
961 * @return true if local or mapped range map contains or is contained by this
964 public boolean containsEither(boolean local, MapList map)
966 // TODO not used - remove?
969 return ((getFromLowest() >= map.getFromLowest()
970 && getFromHighest() <= map.getFromHighest())
971 || (getFromLowest() <= map.getFromLowest()
972 && getFromHighest() >= map.getFromHighest()));
976 return ((getToLowest() >= map.getToLowest()
977 && getToHighest() <= map.getToHighest())
978 || (getToLowest() <= map.getToLowest()
979 && getToHighest() >= map.getToHighest()));
984 * String representation - for debugging, not guaranteed not to change
987 public String toString()
989 StringBuilder sb = new StringBuilder(64);
991 for (int[] shift : fromShifts)
993 sb.append(" ").append(Arrays.toString(shift));
996 sb.append(fromRatio).append(":").append(toRatio);
998 for (int[] shift : toShifts)
1000 sb.append(" ").append(Arrays.toString(shift));
1003 return sb.toString();
1007 * Extend this map list by adding the given map's ranges. There is no
1008 * validation check that the ranges do not overlap existing ranges (or each
1009 * other), but contiguous ranges are merged.
1013 public void addMapList(MapList map)
1015 if (this.equals(map))
1019 this.fromLowest = Math.min(fromLowest, map.fromLowest);
1020 this.toLowest = Math.min(toLowest, map.toLowest);
1021 this.fromHighest = Math.max(fromHighest, map.fromHighest);
1022 this.toHighest = Math.max(toHighest, map.toHighest);
1024 for (int[] range : map.getFromRanges())
1026 addRange(range, fromShifts);
1028 for (int[] range : map.getToRanges())
1030 addRange(range, toShifts);
1035 * Adds the given range to a list of ranges. If the new range just extends
1036 * existing ranges, the current endpoint is updated instead.
1041 static void addRange(int[] range, List<int[]> addTo)
1044 * list is empty - add to it!
1046 if (addTo.size() == 0)
1052 int[] last = addTo.get(addTo.size() - 1);
1053 boolean lastForward = last[1] >= last[0];
1054 boolean newForward = range[1] >= range[0];
1057 * contiguous range in the same direction - just update endpoint
1059 if (lastForward == newForward && last[1] == range[0])
1066 * next range starts at +1 in forward sense - update endpoint
1068 if (lastForward && newForward && range[0] == last[1] + 1)
1075 * next range starts at -1 in reverse sense - update endpoint
1077 if (!lastForward && !newForward && range[0] == last[1] - 1)
1084 * just add the new range
1090 * Returns true if mapping is from forward strand, false if from reverse
1091 * strand. Result is just based on the first 'from' range that is not a single
1092 * position. Default is true unless proven to be false. Behaviour is not well
1093 * defined if the mapping has a mixture of forward and reverse ranges.
1097 public boolean isFromForwardStrand()
1099 return isForwardStrand(getFromRanges());
1103 * Returns true if mapping is to forward strand, false if to reverse strand.
1104 * Result is just based on the first 'to' range that is not a single position.
1105 * Default is true unless proven to be false. Behaviour is not well defined if
1106 * the mapping has a mixture of forward and reverse ranges.
1110 public boolean isToForwardStrand()
1112 return isForwardStrand(getToRanges());
1116 * A helper method that returns true unless at least one range has start >
1117 * end. Behaviour is undefined for a mixture of forward and reverse ranges.
1122 private boolean isForwardStrand(List<int[]> ranges)
1124 boolean forwardStrand = true;
1125 for (int[] range : ranges)
1127 if (range[1] > range[0])
1129 break; // forward strand confirmed
1131 else if (range[1] < range[0])
1133 forwardStrand = false;
1134 break; // reverse strand confirmed
1137 return forwardStrand;
1142 * @return true if from, or to is a three to 1 mapping
1144 public boolean isTripletMap()
1146 return (toRatio == 3 && fromRatio == 1)
1147 || (fromRatio == 3 && toRatio == 1);
1151 * Returns a map which is the composite of this one and the input map. That
1152 * is, the output map has the fromRanges of this map, and its toRanges are the
1153 * toRanges of this map as transformed by the input map.
1155 * Returns null if the mappings cannot be traversed (not all toRanges of this
1156 * map correspond to fromRanges of the input), or if this.toRatio does not
1157 * match map.fromRatio.
1161 * this: from [1-100] to [501-600]
1162 * input: from [10-40] to [60-90]
1163 * output: from [10-40] to [560-590]
1164 * Example 2 ('reverse strand exons'):
1165 * this: from [1-100] to [2000-1951], [1000-951] // transcript to loci
1166 * input: from [1-50] to [41-90] // CDS to transcript
1167 * output: from [10-40] to [1960-1951], [1000-971] // CDS to gene loci
1173 public MapList traverse(MapList map)
1181 * compound the ratios by this rule:
1182 * A:B with M:N gives A*M:B*N
1183 * reduced by greatest common divisor
1184 * so 1:3 with 3:3 is 3:9 or 1:3
1185 * 1:3 with 3:1 is 3:3 or 1:1
1186 * 1:3 with 1:3 is 1:9
1187 * 2:5 with 3:7 is 6:35
1189 int outFromRatio = getFromRatio() * map.getFromRatio();
1190 int outToRatio = getToRatio() * map.getToRatio();
1191 int gcd = MathUtils.gcd(outFromRatio, outToRatio);
1192 outFromRatio /= gcd;
1195 List<int[]> toRanges = new ArrayList<>();
1196 for (int[] range : getToRanges())
1198 int[] transferred = map.locateInTo(range[0], range[1]);
1199 if (transferred == null || transferred.length % 2 != 0)
1205 * convert [start1, end1, start2, end2, ...]
1206 * to [[start1, end1], [start2, end2], ...]
1208 for (int i = 0; i < transferred.length;)
1210 toRanges.add(new int[] { transferred[i], transferred[i + 1] });
1215 return new MapList(getFromRanges(), toRanges, outFromRatio, outToRatio);
1219 * Answers true if the mapping is from one contiguous range to another, else
1224 public boolean isContiguous()
1226 return fromShifts.size() == 1 && toShifts.size() == 1;