last: scripts/last-dotplot@5aaae537a930

     1 #! /usr/bin/env python

     2 # Author: Martin C. Frith 2008

     3 # SPDX-License-Identifier: GPL-3.0-or-later

     5 # Read pair-wise alignments in MAF or LAST tabular format: write an

     6 # "Oxford grid", a.k.a. dotplot.

     8 # TODO: Currently, pixels with zero aligned nt-pairs are white, and

     9 # pixels with one or more aligned nt-pairs are black.  This can look

    10 # too crowded for large genome alignments.  I tried shading each pixel

    11 # according to the number of aligned nt-pairs within it, but the

    12 # result is too faint.  How can this be done better?

    14 import collections

    15 import functools

    16 import gzip

    17 from fnmatch import fnmatchcase

    18 import logging

    19 from operator import itemgetter

    20 import subprocess

    21 import itertools, optparse, os, re, sys

    23 # Try to make PIL/PILLOW work:

    24 try:

    25     from PIL import Image, ImageDraw, ImageFont, ImageColor

    26 except ImportError:

    27     import Image, ImageDraw, ImageFont, ImageColor

    29 try:

    30     from future_builtins import zip

    31 except ImportError:

    32     pass

    34 def myOpen(fileName):  # faster than fileinput

    35     if fileName is None:

    36         return []

    37     if fileName == "-":

    38         return sys.stdin

    39     if fileName.endswith(".gz"):

    40         return gzip.open(fileName, "rt")  # xxx dubious for Python2

    41     return open(fileName)

    43 def groupByFirstItem(things):

    44     for k, v in itertools.groupby(things, itemgetter(0)):

    45         yield k, [i[1:] for i in v]

    47 def croppedBlocks(blocks, ranges1, ranges2):

    48     headBeg1, headBeg2, headSize = blocks[0]

    49     for r1 in ranges1:

    50         for r2 in ranges2:

    51             cropBeg1, cropEnd1 = r1

    52             if headBeg1 < 0:

    53                 cropBeg1, cropEnd1 = -cropEnd1, -cropBeg1

    54             cropBeg2, cropEnd2 = r2

    55             if headBeg2 < 0:

    56                 cropBeg2, cropEnd2 = -cropEnd2, -cropBeg2

    57             for beg1, beg2, size in blocks:

    58                 b1 = max(cropBeg1, beg1)

    59                 e1 = min(cropEnd1, beg1 + size)

    60                 if b1 >= e1: continue

    61                 offset = beg2 - beg1

    62                 b2 = max(cropBeg2, b1 + offset)

    63                 e2 = min(cropEnd2, e1 + offset)

    64                 if b2 >= e2: continue

    65                 yield b2 - offset, b2, e2 - b2

    67 def tabBlocks(beg1, beg2, blocks):

    68     '''Get the gapless blocks of an alignment, from LAST tabular format.'''

    69     for i in blocks.split(","):

    70         if ":" in i:

    71             x, y = i.split(":")

    72             beg1 += int(x)

    73             beg2 += int(y)

    74         else:

    75             size = int(i)

    76             yield beg1, beg2, size

    77             beg1 += size

    78             beg2 += size

    80 def mafBlocks(beg1, beg2, seq1, seq2):

    81     '''Get the gapless blocks of an alignment, from MAF format.'''

    82     size = 0

    83     for x, y in zip(seq1, seq2):

    84         if x == "-":

    85             if size:

    86                 yield beg1, beg2, size

    87                 beg1 += size

    88                 beg2 += size

    89                 size = 0

    90             beg2 += 1

    91         elif y == "-":

    92             if size:

    93                 yield beg1, beg2, size

    94                 beg1 += size

    95                 beg2 += size

    96                 size = 0

    97             beg1 += 1

    98         else:

    99             size += 1

   100     if size: yield beg1, beg2, size

   102 def alignmentFromSegment(qrySeqName, qrySeqLen, segment):

   103     refSeqLen = sys.maxsize  # XXX

   104     refSeqName, refSeqBeg, qrySeqBeg, size = segment

   105     block = refSeqBeg, qrySeqBeg, size

   106     return refSeqName, refSeqLen, qrySeqName, qrySeqLen, [block]

   108 def alignmentInput(lines):

   109     '''Get alignments and sequence lengths, from MAF or tabular format.'''

   110     mafCount = 0

   111     qrySeqName = ""

   112     segments = []

   113     for line in lines:

   114         w = line.split()

   115         if line[0] == "#":

   116             pass

   117         elif len(w) == 1:

   118             for i in segments:

   119                 yield alignmentFromSegment(qrySeqName, qrySeqLen, i)

   120             qrySeqName = w[0]

   121             qrySeqLen = 0

   122             segments = []

   123         elif len(w) == 2 and qrySeqName and w[1].isdigit():

   124             qrySeqLen += int(w[1])

   125         elif len(w) == 4 and qrySeqName and w[1].isdigit() and w[3].isdigit():

   126             refSeqName, refSeqBeg, refSeqEnd = w[0], int(w[1]), int(w[3])

   127             size = abs(refSeqEnd - refSeqBeg)

   128             if refSeqBeg > refSeqEnd:

   129                 refSeqBeg = -refSeqBeg

   130             segments.append((refSeqName, refSeqBeg, qrySeqLen, size))

   131             qrySeqLen += size

   132         elif line[0].isdigit():  # tabular format

   133             chr1, beg1, seqlen1 = w[1], int(w[2]), int(w[5])

   134             if w[4] == "-": beg1 -= seqlen1

   135             chr2, beg2, seqlen2 = w[6], int(w[7]), int(w[10])

   136             if w[9] == "-": beg2 -= seqlen2

   137             blocks = tabBlocks(beg1, beg2, w[11])

   138             yield chr1, seqlen1, chr2, seqlen2, blocks

   139         elif line[0] == "s":  # MAF format

   140             if mafCount == 0:

   141                 chr1, beg1, seqlen1, seq1 = w[1], int(w[2]), int(w[5]), w[6]

   142                 if w[4] == "-": beg1 -= seqlen1

   143                 mafCount = 1

   144             else:

   145                 chr2, beg2, seqlen2, seq2 = w[1], int(w[2]), int(w[5]), w[6]

   146                 if w[4] == "-": beg2 -= seqlen2

   147                 blocks = mafBlocks(beg1, beg2, seq1, seq2)

   148                 yield chr1, seqlen1, chr2, seqlen2, blocks

   149                 mafCount = 0

   150     for i in segments:

   151         yield alignmentFromSegment(qrySeqName, qrySeqLen, i)

   153 def seqRequestFromText(text):

   154     if ":" in text:

   155         pattern, interval = text.rsplit(":", 1)

   156         if "-" in interval:

   157             beg, end = interval.rsplit("-", 1)

   158             return pattern, int(beg), int(end)  # beg may be negative

   159     return text, 0, sys.maxsize

   161 def rangesFromSeqName(seqRequests, name, seqLen):

   162     if seqRequests:

   163         base = name.split(".")[-1]  # allow for names like hg19.chr7

   164         for pat, beg, end in seqRequests:

   165             if fnmatchcase(name, pat) or fnmatchcase(base, pat):

   166                 yield max(beg, 0), min(end, seqLen)

   167     else:

   168         yield 0, seqLen

   170 def updateSeqs(coverDict, seqRanges, seqName, ranges, coveredRange):

   171     beg, end = coveredRange

   172     if beg < 0:

   173         coveredRange = -end, -beg

   174     if seqName in coverDict:

   175         coverDict[seqName].append(coveredRange)

   176     else:

   177         coverDict[seqName] = [coveredRange]

   178         for beg, end in ranges:

   179             r = seqName, beg, end

   180             seqRanges.append(r)

   182 def readAlignments(fileName, opts):

   183     '''Get alignments and sequence limits, from MAF or tabular format.'''

   184     seqRequests1 = [seqRequestFromText(i) for i in opts.seq1]

   185     seqRequests2 = [seqRequestFromText(i) for i in opts.seq2]

   187     alignments = []

   188     seqRanges1 = []

   189     seqRanges2 = []

   190     coverDict1 = {}

   191     coverDict2 = {}

   192     lines = myOpen(fileName)

   193     for seqName1, seqLen1, seqName2, seqLen2, blocks in alignmentInput(lines):

   194         ranges1 = sorted(rangesFromSeqName(seqRequests1, seqName1, seqLen1))

   195         if not ranges1: continue

   196         ranges2 = sorted(rangesFromSeqName(seqRequests2, seqName2, seqLen2))

   197         if not ranges2: continue

   198         b = list(croppedBlocks(list(blocks), ranges1, ranges2))

   199         if not b: continue

   200         aln = seqName1, seqName2, b

   201         alignments.append(aln)

   202         coveredRange1 = b[0][0], b[-1][0] + b[-1][2]

   203         updateSeqs(coverDict1, seqRanges1, seqName1, ranges1, coveredRange1)

   204         coveredRange2 = b[0][1], b[-1][1] + b[-1][2]

   205         updateSeqs(coverDict2, seqRanges2, seqName2, ranges2, coveredRange2)

   206     return alignments, seqRanges1, coverDict1, seqRanges2, coverDict2

   208 def nameAndRangesFromDict(cropDict, seqName):

   209     if seqName in cropDict:

   210         return seqName, cropDict[seqName]

   211     n = seqName.split(".")[-1]

   212     if n in cropDict:

   213         return n, cropDict[n]

   214     return seqName, []

   216 def rangesForSecondaryAlignments(primaryRanges, seqLen):

   217     if primaryRanges:

   218         return primaryRanges

   219     return [(0, seqLen)]

   221 def readSecondaryAlignments(opts, cropRanges1, cropRanges2):

   222     cropDict1 = dict(groupByFirstItem(cropRanges1))

   223     cropDict2 = dict(groupByFirstItem(cropRanges2))

   225     alignments = []

   226     seqRanges1 = []

   227     seqRanges2 = []

   228     coverDict1 = {}

   229     coverDict2 = {}

   230     lines = myOpen(opts.alignments)

   231     for seqName1, seqLen1, seqName2, seqLen2, blocks in alignmentInput(lines):

   232         seqName1, ranges1 = nameAndRangesFromDict(cropDict1, seqName1)

   233         seqName2, ranges2 = nameAndRangesFromDict(cropDict2, seqName2)

   234         if not ranges1 and not ranges2:

   235             continue

   236         r1 = rangesForSecondaryAlignments(ranges1, seqLen1)

   237         r2 = rangesForSecondaryAlignments(ranges2, seqLen2)

   238         b = list(croppedBlocks(list(blocks), r1, r2))

   239         if not b: continue

   240         aln = seqName1, seqName2, b

   241         alignments.append(aln)

   242         if not ranges1:

   243             coveredRange1 = b[0][0], b[-1][0] + b[-1][2]

   244             updateSeqs(coverDict1, seqRanges1, seqName1, r1, coveredRange1)

   245         if not ranges2:

   246             coveredRange2 = b[0][1], b[-1][1] + b[-1][2]

   247             updateSeqs(coverDict2, seqRanges2, seqName2, r2, coveredRange2)

   248     return alignments, seqRanges1, coverDict1, seqRanges2, coverDict2

   250 def twoValuesFromOption(text, separator):

   251     if separator in text:

   252         return text.split(separator)

   253     return text, text

   255 def mergedRanges(ranges):

   256     oldBeg, maxEnd = ranges[0]

   257     for beg, end in ranges:

   258         if beg > maxEnd:

   259             yield oldBeg, maxEnd

   260             oldBeg = beg

   261             maxEnd = end

   262         elif end > maxEnd:

   263             maxEnd = end

   264     yield oldBeg, maxEnd

   266 def mergedRangesPerSeq(coverDict):

   267     for k, v in coverDict.items():

   268         v.sort()

   269         yield k, list(mergedRanges(v))

   271 def coveredLength(mergedCoverDict):

   272     return sum(sum(e - b for b, e in v) for v in mergedCoverDict.values())

   274 def trimmed(seqRanges, coverDict, minAlignedBases, maxGapFrac, endPad, midPad):

   275     maxEndGapFrac, maxMidGapFrac = twoValuesFromOption(maxGapFrac, ",")

   276     maxEndGap = max(float(maxEndGapFrac) * minAlignedBases, endPad * 1.0)

   277     maxMidGap = max(float(maxMidGapFrac) * minAlignedBases, midPad * 2.0)

   279     for seqName, rangeBeg, rangeEnd in seqRanges:

   280         seqBlocks = coverDict[seqName]

   281         blocks = [i for i in seqBlocks if i[0] < rangeEnd and i[1] > rangeBeg]

   282         if blocks[0][0] - rangeBeg > maxEndGap:

   283             rangeBeg = blocks[0][0] - endPad

   284         for j, y in enumerate(blocks):

   285             if j:

   286                 x = blocks[j - 1]

   287                 if y[0] - x[1] > maxMidGap:

   288                     yield seqName, rangeBeg, x[1] + midPad

   289                     rangeBeg = y[0] - midPad

   290         if rangeEnd - blocks[-1][1] > maxEndGap:

   291             rangeEnd = blocks[-1][1] + endPad

   292         yield seqName, rangeBeg, rangeEnd

   294 def rangesWithStrandInfo(seqRanges, strandOpt, alignments, seqIndex):

   295     if strandOpt == "1":

   296         forwardMinusReverse = collections.defaultdict(int)

   297         for i in alignments:

   298             blocks = i[2]

   299             beg1, beg2, size = blocks[0]

   300             numOfAlignedLetterPairs = sum(i[2] for i in blocks)

   301             if (beg1 < 0) != (beg2 < 0):  # opposite-strand alignment

   302                 numOfAlignedLetterPairs *= -1

   303             forwardMinusReverse[i[seqIndex]] += numOfAlignedLetterPairs

   304     strandNum = 0

   305     for seqName, beg, end in seqRanges:

   306         if strandOpt == "1":

   307             strandNum = 1 if forwardMinusReverse[seqName] >= 0 else 2

   308         yield seqName, beg, end, strandNum

   310 def natural_sort_key(my_string):

   311     '''Return a sort key for "natural" ordering, e.g. chr9 < chr10.'''

   312     parts = re.split(r'(\d+)', my_string)

   313     parts[1::2] = map(int, parts[1::2])

   314     return parts

   316 def nameKey(oneSeqRanges):

   317     return natural_sort_key(oneSeqRanges[0][0])

   319 def sizeKey(oneSeqRanges):

   320     return sum(b - e for n, b, e, s in oneSeqRanges), nameKey(oneSeqRanges)

   322 def alignmentKey(seqNamesToLists, oneSeqRanges):

   323     seqName = oneSeqRanges[0][0]

   324     alignmentsOfThisSequence = seqNamesToLists[seqName]

   325     numOfAlignedLetterPairs = sum(i[3] for i in alignmentsOfThisSequence)

   326     toMiddle = numOfAlignedLetterPairs // 2

   327     for i in alignmentsOfThisSequence:

   328         toMiddle -= i[3]

   329         if toMiddle < 0:

   330             return i[1:3]  # sequence-rank and "position" of this alignment

   332 def rankAndFlipPerSeq(seqRanges):

   333     rangesGroupedBySeqName = itertools.groupby(seqRanges, itemgetter(0))

   334     for rank, group in enumerate(rangesGroupedBySeqName):

   335         seqName, ranges = group

   336         strandNum = next(ranges)[3]

   337         flip = 1 if strandNum < 2 else -1

   338         yield seqName, (rank, flip)

   340 def alignmentSortData(alignments, seqIndex, otherNamesToRanksAndFlips):

   341     otherIndex = 1 - seqIndex

   342     for i in alignments:

   343         blocks = i[2]

   344         otherRank, otherFlip = otherNamesToRanksAndFlips[i[otherIndex]]

   345         otherPos = otherFlip * abs(blocks[0][otherIndex] +

   346                                    blocks[-1][otherIndex] + blocks[-1][2])

   347         numOfAlignedLetterPairs = sum(i[2] for i in blocks)

   348         yield i[seqIndex], otherRank, otherPos, numOfAlignedLetterPairs

   350 def mySortedRanges(seqRanges, sortOpt, seqIndex, alignments, otherRanges):

   351     rangesGroupedBySeqName = itertools.groupby(seqRanges, itemgetter(0))

   352     g = [list(ranges) for seqName, ranges in rangesGroupedBySeqName]

   353     for i in g:

   354         if i[0][3] > 1:

   355             i.reverse()

   356     if sortOpt == "1":

   357         g.sort(key=nameKey)

   358     if sortOpt == "2":

   359         g.sort(key=sizeKey)

   360     if sortOpt == "3":

   361         otherNamesToRanksAndFlips = dict(rankAndFlipPerSeq(otherRanges))

   362         alns = sorted(alignmentSortData(alignments, seqIndex,

   363                                         otherNamesToRanksAndFlips))

   364         alnsGroupedBySeqName = itertools.groupby(alns, itemgetter(0))

   365         seqNamesToLists = dict((k, list(v)) for k, v in alnsGroupedBySeqName)

   366         g.sort(key=functools.partial(alignmentKey, seqNamesToLists))

   367     return [j for i in g for j in i]

   369 def allSortedRanges(opts, alignments, alignmentsB,

   370                     seqRanges1, seqRangesB1, seqRanges2, seqRangesB2):

   371     o1, oB1 = twoValuesFromOption(opts.strands1, ":")

   372     o2, oB2 = twoValuesFromOption(opts.strands2, ":")

   373     if o1 == "1" and o2 == "1":

   374         raise Exception("the strand options have circular dependency")

   375     seqRanges1 = list(rangesWithStrandInfo(seqRanges1, o1, alignments, 0))

   376     seqRanges2 = list(rangesWithStrandInfo(seqRanges2, o2, alignments, 1))

   377     seqRangesB1 = list(rangesWithStrandInfo(seqRangesB1, oB1, alignmentsB, 0))

   378     seqRangesB2 = list(rangesWithStrandInfo(seqRangesB2, oB2, alignmentsB, 1))

   380     o1, oB1 = twoValuesFromOption(opts.sort1, ":")

   381     o2, oB2 = twoValuesFromOption(opts.sort2, ":")

   382     if o1 == "3" and o2 == "3":

   383         raise Exception("the sort options have circular dependency")

   384     if o1 != "3":

   385         s1 = mySortedRanges(seqRanges1, o1, None, None, None)

   386     if o2 != "3":

   387         s2 = mySortedRanges(seqRanges2, o2, None, None, None)

   388     if o1 == "3":

   389         s1 = mySortedRanges(seqRanges1, o1, 0, alignments, s2)

   390     if o2 == "3":

   391         s2 = mySortedRanges(seqRanges2, o2, 1, alignments, s1)

   392     t1 = mySortedRanges(seqRangesB1, oB1, 0, alignmentsB, s2)

   393     t2 = mySortedRanges(seqRangesB2, oB2, 1, alignmentsB, s1)

   394     return s1 + t1, s2 + t2

   396 def sizesPerText(texts, font, textDraw):

   397     sizes = 0, 0

   398     for t in texts:

   399         if textDraw is not None:

   400             sizes = textDraw.textsize(t, font=font)

   401         yield t, sizes

   403 def prettyNum(n):

   404     t = str(n)

   405     groups = []

   406     while t:

   407         groups.append(t[-3:])

   408         t = t[:-3]

   409     return ",".join(reversed(groups))

   411 def sizeText(size):

   412     suffixes = "bp", "kb", "Mb", "Gb"

   413     for i, x in enumerate(suffixes):

   414         j = 10 ** (i * 3)

   415         if size < j * 10:

   416             return "%.2g" % (1.0 * size / j) + x

   417         if size < j * 1000 or i == len(suffixes) - 1:

   418             return "%.0f" % (1.0 * size / j) + x

   420 def labelText(seqRange, labelOpt):

   421     seqName, beg, end, strandNum = seqRange

   422     if labelOpt == 1:

   423         return seqName + ": " + sizeText(end - beg)

   424     if labelOpt == 2:

   425         return seqName + ":" + prettyNum(beg) + ": " + sizeText(end - beg)

   426     if labelOpt == 3:

   427         return seqName + ":" + prettyNum(beg) + "-" + prettyNum(end)

   428     return seqName

   430 def rangeLabels(seqRanges, labelOpt, font, textDraw, textRot):

   431     x = y = 0

   432     for r in seqRanges:

   433         text = labelText(r, labelOpt)

   434         if textDraw is not None:

   435             x, y = textDraw.textsize(text, font=font)

   436             if textRot:

   437                 x, y = y, x

   438         yield text, x, y, r[3]

   440 def dataFromRanges(sortedRanges, font, textDraw, labelOpt, textRot):

   441     for seqName, rangeBeg, rangeEnd, strandNum in sortedRanges:

   442         out = [seqName, str(rangeBeg), str(rangeEnd)]

   443         if strandNum > 0:

   444             out.append(".+-"[strandNum])

   445         logging.info("\t".join(out))

   446     logging.info("")

   447     rangeSizes = [e - b for n, b, e, s in sortedRanges]

   448     labs = list(rangeLabels(sortedRanges, labelOpt, font, textDraw, textRot))

   449     margin = max(i[2] for i in labs)

   450     # xxx the margin may be too big, because some labels may get omitted

   451     return rangeSizes, labs, margin

   453 def div_ceil(x, y):

   454     '''Return x / y rounded up.'''

   455     q, r = divmod(x, y)

   456     return q + (r != 0)

   458 def get_bp_per_pix(rangeSizes, pixTweenRanges, maxPixels):

   459     '''Get the minimum bp-per-pixel that fits in the size limit.'''

   460     logging.info("choosing bp per pixel...")

   461     numOfRanges = len(rangeSizes)

   462     maxPixelsInRanges = maxPixels - pixTweenRanges * (numOfRanges - 1)

   463     if maxPixelsInRanges < numOfRanges:

   464         raise Exception("can't fit the image: too many sequences?")

   465     negLimit = -maxPixelsInRanges

   466     negBpPerPix = sum(rangeSizes) // negLimit

   467     while True:

   468         if sum(i // negBpPerPix for i in rangeSizes) >= negLimit:

   469             return -negBpPerPix

   470         negBpPerPix -= 1

   472 def getRangePixBegs(rangePixLens, pixTweenRanges, margin):

   473     '''Get the start pixel for each range.'''

   474     rangePixBegs = []

   475     pix_tot = margin - pixTweenRanges

   476     for i in rangePixLens:

   477         pix_tot += pixTweenRanges

   478         rangePixBegs.append(pix_tot)

   479         pix_tot += i

   480     return rangePixBegs

   482 def pixelData(rangeSizes, bp_per_pix, pixTweenRanges, margin):

   483     '''Return pixel information about the ranges.'''

   484     rangePixLens = [div_ceil(i, bp_per_pix) for i in rangeSizes]

   485     rangePixBegs = getRangePixBegs(rangePixLens, pixTweenRanges, margin)

   486     tot_pix = rangePixBegs[-1] + rangePixLens[-1]

   487     return rangePixBegs, rangePixLens, tot_pix

   489 def drawLineForward(hits, width, bp_per_pix, beg1, beg2, size):

   490     while True:

   491         q1, r1 = divmod(beg1, bp_per_pix)

   492         q2, r2 = divmod(beg2, bp_per_pix)

   493         hits[q2 * width + q1] |= 1

   494         next_pix = min(bp_per_pix - r1, bp_per_pix - r2)

   495         if next_pix >= size: break

   496         beg1 += next_pix

   497         beg2 += next_pix

   498         size -= next_pix

   500 def drawLineReverse(hits, width, bp_per_pix, beg1, beg2, size):

   501     while True:

   502         q1, r1 = divmod(beg1, bp_per_pix)

   503         q2, r2 = divmod(beg2, bp_per_pix)

   504         hits[q2 * width + q1] |= 2

   505         next_pix = min(bp_per_pix - r1, r2 + 1)

   506         if next_pix >= size: break

   507         beg1 += next_pix

   508         beg2 -= next_pix

   509         size -= next_pix

   511 def strandAndOrigin(ranges, beg, size):

   512     isReverseStrand = (beg < 0)

   513     if isReverseStrand:

   514         beg = -(beg + size)

   515     for rangeBeg, rangeEnd, isReverseRange, origin in ranges:

   516         if rangeEnd > beg:  # assumes the ranges are sorted

   517             return (isReverseStrand != isReverseRange), origin

   519 def alignmentPixels(width, height, alignments, bp_per_pix,

   520                     rangeDict1, rangeDict2):

   521     hits = [0] * (width * height)  # the image data

   522     for seq1, seq2, blocks in alignments:

   523         beg1, beg2, size = blocks[0]

   524         isReverse1, ori1 = strandAndOrigin(rangeDict1[seq1], beg1, size)

   525         isReverse2, ori2 = strandAndOrigin(rangeDict2[seq2], beg2, size)

   526         for beg1, beg2, size in blocks:

   527             if isReverse1:

   528                 beg1 = -(beg1 + size)

   529                 beg2 = -(beg2 + size)

   530             if isReverse1 == isReverse2:

   531                 drawLineForward(hits, width, bp_per_pix,

   532                                 ori1 + beg1, ori2 + beg2, size)

   533             else:

   534                 drawLineReverse(hits, width, bp_per_pix,

   535                                 ori1 + beg1, ori2 - beg2 - 1, size)

   536     return hits

   538 def orientedBlocks(alignments, seqIndex):

   539     otherIndex = 1 - seqIndex

   540     for a in alignments:

   541         seq1, seq2, blocks = a

   542         for b in blocks:

   543             beg1, beg2, size = b

   544             if b[seqIndex] < 0:

   545                 b = -(beg1 + size), -(beg2 + size), size

   546             yield a[seqIndex], b[seqIndex], a[otherIndex], b[otherIndex], size

   548 def drawJoins(im, alignments, bpPerPix, seqIndex, rangeDict1, rangeDict2):

   549     blocks = orientedBlocks(alignments, seqIndex)

   550     oldSeq1 = ""

   551     for seq1, beg1, seq2, beg2, size in sorted(blocks):

   552         isReverse1, ori1 = strandAndOrigin(rangeDict1[seq1], beg1, size)

   553         isReverse2, ori2 = strandAndOrigin(rangeDict2[seq2], beg2, size)

   554         end1 = beg1 + size - 1

   555         end2 = beg2 + size - 1

   556         if isReverse1:

   557             beg1 = -(beg1 + 1)

   558             end1 = -(end1 + 1)

   559         if isReverse2:

   560             beg2 = -(beg2 + 1)

   561             end2 = -(end2 + 1)

   562         newPix1 = (ori1 + beg1) // bpPerPix

   563         newPix2 = (ori2 + beg2) // bpPerPix

   564         if seq1 == oldSeq1:

   565             lowerPix2 = min(oldPix2, newPix2)

   566             upperPix2 = max(oldPix2, newPix2)

   567             midPix1 = (oldPix1 + newPix1) // 2

   568             if isReverse1:

   569                 midPix1 = (oldPix1 + newPix1 + 1) // 2

   570                 oldPix1, newPix1 = newPix1, oldPix1

   571             if upperPix2 - lowerPix2 > 1 and oldPix1 <= newPix1 <= oldPix1 + 1:

   572                 if seqIndex == 0:

   573                     box = midPix1, lowerPix2, midPix1 + 1, upperPix2 + 1

   574                 else:

   575                     box = lowerPix2, midPix1, upperPix2 + 1, midPix1 + 1

   576                 im.paste("lightgray", box)

   577         oldPix1 = (ori1 + end1) // bpPerPix

   578         oldPix2 = (ori2 + end2) // bpPerPix

   579         oldSeq1 = seq1

   581 def expandedSeqDict(seqDict):

   582     '''Allow lookup by short sequence names, e.g. chr7 as well as hg19.chr7.'''

   583     newDict = seqDict.copy()

   584     for name, x in seqDict.items():

   585         if "." in name:

   586             base = name.split(".")[-1]

   587             if base in newDict:  # an ambiguous case was found:

   588                 return seqDict   # so give up completely

   589             newDict[base] = x

   590     return newDict

   592 def readBed(fileName, rangeDict):

   593     for line in myOpen(fileName):

   594         w = line.split()

   595         if not w: continue

   596         if not w[1].isdigit():

   597             w = w[1:]

   598         seqName = w[0]

   599         if seqName not in rangeDict: continue

   600         seqRanges = rangeDict[seqName]

   601         beg = int(w[1])

   602         end = int(w[2])

   603         if all(beg >= i[2] or end <= i[1] for i in seqRanges):

   604             continue

   605         itemName = w[3] if len(w) > 3 and w[3] != "." else ""

   606         layer = 900

   607         color = "#fbf"

   608         if len(w) > 4:

   609             if w[4] != ".":

   610                 layer = float(w[4])

   611             if len(w) > 5:

   612                 if len(w) > 8 and w[8].count(",") == 2:

   613                     color = "rgb(" + w[8] + ")"

   614                 else:

   615                     strand = w[5]

   616                     isRev = (seqRanges[0][3] > 1)

   617                     if strand == "+" and not isRev or strand == "-" and isRev:

   618                         color = "#ffe8e8"

   619                     if strand == "-" and not isRev or strand == "+" and isRev:

   620                         color = "#e8e8ff"

   621         yield layer, color, seqName, beg, end, itemName

   623 def commaSeparatedInts(text):

   624     return map(int, text.rstrip(",").split(","))

   626 def readGenePred(opts, fileName, rangeDict):

   627     for line in myOpen(fileName):

   628         fields = line.split()  # xxx tab ???

   629         if not fields: continue

   630         geneName = fields[12 if len(fields) > 12 else 0]  # XXX ???

   631         if fields[2] not in "+-":

   632             fields = fields[1:]

   633         seqName = fields[1]

   634         if seqName not in rangeDict: continue

   635         seqRanges = rangeDict[seqName]

   636         #strand = fields[2]

   637         cdsBeg = int(fields[5])

   638         cdsEnd = int(fields[6])

   639         exonBegs = commaSeparatedInts(fields[8])

   640         exonEnds = commaSeparatedInts(fields[9])

   641         for beg, end in zip(exonBegs, exonEnds):

   642             if all(beg >= i[2] or end <= i[1] for i in seqRanges):

   643                 continue

   644             yield 300, opts.exon_color, seqName, beg, end, geneName

   645             b = max(beg, cdsBeg)

   646             e = min(end, cdsEnd)

   647             if b < e: yield 400, opts.cds_color, seqName, b, e, ""

   649 def readRmsk(fileName, rangeDict):

   650     for line in myOpen(fileName):

   651         fields = line.split()

   652         if len(fields) == 17:  # rmsk.txt

   653             seqName = fields[5]

   654             if seqName not in rangeDict: continue  # do this ASAP for speed

   655             beg = int(fields[6])

   656             end = int(fields[7])

   657             strand = fields[9]

   658             repeatName = fields[10]

   659             repeatClass = fields[11]

   660         elif len(fields) == 15:  # .out

   661             seqName = fields[4]

   662             if seqName not in rangeDict: continue

   663             beg = int(fields[5]) - 1

   664             end = int(fields[6])

   665             strand = fields[8]

   666             repeatName = fields[9]

   667             repeatClass = fields[10].split("/")[0]

   668         else:

   669             continue

   670         seqRanges = rangeDict[seqName]

   671         if all(beg >= i[2] or end <= i[1] for i in seqRanges):

   672             continue

   673         if repeatClass in ("Low_complexity", "Simple_repeat", "Satellite"):

   674             yield 200, "#fbf", seqName, beg, end, repeatName

   675         elif (strand == "+") != (seqRanges[0][3] > 1):

   676             yield 100, "#ffe8e8", seqName, beg, end, repeatName

   677         else:

   678             yield 100, "#e8e8ff", seqName, beg, end, repeatName

   680 def isExtraFirstGapField(fields):

   681     return fields[4].isdigit()

   683 def readGaps(opts, fileName, rangeDict):

   684     '''Read locations of unsequenced gaps, from an agp or gap file.'''

   685     for line in myOpen(fileName):

   686         w = line.split()

   687         if not w or w[0][0] == "#": continue

   688         if isExtraFirstGapField(w): w = w[1:]

   689         if w[4] not in "NU": continue

   690         seqName = w[0]

   691         if seqName not in rangeDict: continue

   692         end = int(w[2])

   693         beg = end - int(w[5])  # zero-based coordinate

   694         if w[7] == "yes":

   695             yield 3000, opts.bridged_color, seqName, beg, end, ""

   696         else:

   697             yield 2000, opts.unbridged_color, seqName, beg, end, ""

   699 def bedBoxes(beds, rangeDict, limit, isTop, bpPerPix, textSizes):

   700     cover = [(limit, limit)]

   701     for layer, color, seqName, bedBeg, bedEnd, name in reversed(beds):

   702         for rangeBeg, rangeEnd, isReverseRange, origin in rangeDict[seqName]:

   703             beg = max(bedBeg, rangeBeg)

   704             end = min(bedEnd, rangeEnd)

   705             if beg >= end: continue

   706             if isReverseRange:

   707                 beg, end = -end, -beg

   708             if layer <= 1000:

   709                 # include partly-covered pixels

   710                 pixBeg = (origin + beg) // bpPerPix

   711                 pixEnd = div_ceil(origin + end, bpPerPix)

   712             else:

   713                 # exclude partly-covered pixels

   714                 pixBeg = div_ceil(origin + beg, bpPerPix)

   715                 pixEnd = (origin + end) // bpPerPix

   716                 if pixEnd <= pixBeg: continue

   717                 if bedEnd >= rangeEnd:  # include partly-covered end pixels

   718                     if isReverseRange:

   719                         pixBeg = (origin + beg) // bpPerPix

   720                     else:

   721                         pixEnd = div_ceil(origin + end, bpPerPix)

   722             textWidth, textHeight = textSizes[name]

   723             nameBeg = (pixBeg + pixEnd - textHeight) // 2

   724             nameEnd = nameBeg + textHeight

   725             if name and all(e <= nameBeg or b >= nameEnd for b, e in cover):

   726                 cover.append((nameBeg, nameEnd))

   727             else:

   728                 name = ""

   729             yield layer, color, isTop, pixBeg, pixEnd, name, nameBeg, textWidth

   731 def drawAnnotations(im, boxes, tMargin, bMarginBeg, lMargin, rMarginBeg):

   732     # xxx use partial transparency for different-color overlaps?

   733     for layer, color, isTop, beg, end, name, nameBeg, nameLen in boxes:

   734         if isTop:

   735             box = beg, tMargin, end, bMarginBeg

   736         else:

   737             box = lMargin, beg, rMarginBeg, end

   738         im.paste(color, box)

   740 def placedLabels(labels, rangePixBegs, rangePixLens, beg, end):

   741     '''Return axis labels with endpoint & sort-order information.'''

   742     maxWidth = end - beg

   743     for i, j, k in zip(labels, rangePixBegs, rangePixLens):

   744         text, textWidth, textHeight, strandNum = i

   745         if textWidth > maxWidth:

   746             continue

   747         labelBeg = j + (k - textWidth) // 2

   748         labelEnd = labelBeg + textWidth

   749         sortKey = textWidth - k

   750         if labelBeg < beg:

   751             sortKey += maxWidth * (beg - labelBeg)

   752             labelBeg = beg

   753             labelEnd = beg + textWidth

   754         if labelEnd > end:

   755             sortKey += maxWidth * (labelEnd - end)

   756             labelEnd = end

   757             labelBeg = end - textWidth

   758         yield sortKey, labelBeg, labelEnd, text, textHeight, strandNum

   760 def nonoverlappingLabels(labels, minPixTweenLabels):

   761     '''Get a subset of non-overlapping axis labels, greedily.'''

   762     out = []

   763     for i in labels:

   764         beg = i[1] - minPixTweenLabels

   765         end = i[2] + minPixTweenLabels

   766         if all(j[2] <= beg or j[1] >= end for j in out):

   767             out.append(i)

   768     return out

   770 def axisImage(labels, rangePixBegs, rangePixLens, textRot,

   771               textAln, font, image_mode, opts):

   772     '''Make an image of axis labels.'''

   773     beg = rangePixBegs[0]

   774     end = rangePixBegs[-1] + rangePixLens[-1]

   775     margin = max(i[2] for i in labels)

   776     labels = sorted(placedLabels(labels, rangePixBegs, rangePixLens, beg, end))

   777     minPixTweenLabels = 0 if textRot else opts.label_space

   778     labels = nonoverlappingLabels(labels, minPixTweenLabels)

   779     image_size = (margin, end) if textRot else (end, margin)

   780     im = Image.new(image_mode, image_size, opts.margin_color)

   781     draw = ImageDraw.Draw(im)

   782     for sortKey, labelBeg, labelEnd, text, textHeight, strandNum in labels:

   783         base = margin - textHeight if textAln else 0

   784         position = (base, labelBeg) if textRot else (labelBeg, base)

   785         fill = ("black", opts.forwardcolor, opts.reversecolor)[strandNum]

   786         draw.text(position, text, font=font, fill=fill)

   787     return im

   789 def annoTextImage(opts, image_mode, font, margin, length, boxes, isLeftAlign):

   790     image_size = margin, length

   791     im = Image.new(image_mode, image_size, opts.margin_color)

   792     draw = ImageDraw.Draw(im)

   793     for layer, color, isTop, beg, end, name, nameBeg, nameLen in boxes:

   794         xPosition = 0 if isLeftAlign else margin - nameLen

   795         position = xPosition, nameBeg

   796         draw.text(position, name, font=font, fill="black")

   797     return im

   799 def rangesPerSeq(sortedRanges):

   800     for seqName, group in itertools.groupby(sortedRanges, itemgetter(0)):

   801         yield seqName, sorted(group)

   803 def rangesWithOrigins(sortedRanges, rangePixBegs, rangePixLens, bpPerPix):

   804     for i, j, k in zip(sortedRanges, rangePixBegs, rangePixLens):

   805         seqName, rangeBeg, rangeEnd, strandNum = i

   806         isReverseRange = (strandNum > 1)

   807         if isReverseRange:

   808             origin = bpPerPix * (j + k) + rangeBeg

   809         else:

   810             origin = bpPerPix * j - rangeBeg

   811         yield seqName, (rangeBeg, rangeEnd, isReverseRange, origin)

   813 def rangesAndOriginsPerSeq(sortedRanges, rangePixBegs, rangePixLens, bpPerPix):

   814     a = rangesWithOrigins(sortedRanges, rangePixBegs, rangePixLens, bpPerPix)

   815     for seqName, group in itertools.groupby(a, itemgetter(0)):

   816         yield seqName, sorted(i[1] for i in group)

   818 def getFont(opts):

   819     if opts.fontfile:

   820         return ImageFont.truetype(opts.fontfile, opts.fontsize)

   821     fileNames = []

   822     try:

   823         x = ["fc-match", "-f%{file}", "arial"]

   824         p = subprocess.Popen(x, stdout=subprocess.PIPE, stderr=subprocess.PIPE,

   825                              universal_newlines=True)

   826         out, err = p.communicate()

   827         fileNames.append(out)

   828     except OSError as e:

   829         logging.info("fc-match error: " + str(e))

   830     fileNames.append("/Library/Fonts/Arial.ttf")  # for Mac

   831     for i in fileNames:

   832         try:

   833             font = ImageFont.truetype(i, opts.fontsize)

   834             logging.info("font: " + i)

   835             return font

   836         except IOError as e:

   837             logging.info("font load error: " + str(e))

   838     return ImageFont.load_default()

   840 def sequenceSizesAndNames(seqRanges):

   841     for seqName, ranges in itertools.groupby(seqRanges, itemgetter(0)):

   842         size = sum(e - b for n, b, e in ranges)

   843         yield size, seqName

   845 def biggestSequences(seqRanges, maxNumOfSequences):

   846     s = sorted(sequenceSizesAndNames(seqRanges), reverse=True)

   847     if len(s) > maxNumOfSequences:

   848         logging.warning("too many sequences - discarding the smallest ones")

   849         s = s[:maxNumOfSequences]

   850     return set(i[1] for i in s)

   852 def remainingSequenceRanges(seqRanges, alignments, seqIndex):

   853     remainingSequences = set(i[seqIndex] for i in alignments)

   854     return [i for i in seqRanges if i[0] in remainingSequences]

   856 def readAnnotations(opts, font, textDraw, sortedRanges,

   857                     bedFile, repFile, geneFile, gapFile):

   858     rangeDict = expandedSeqDict(dict(rangesPerSeq(sortedRanges)))

   859     annots = sorted(itertools.chain(readBed(bedFile, rangeDict),

   860                                     readRmsk(repFile, rangeDict),

   861                                     readGenePred(opts, geneFile, rangeDict),

   862                                     readGaps(opts, gapFile, rangeDict)))

   863     names = set(i[5] for i in annots)

   864     textSizes = dict(sizesPerText(names, font, textDraw))

   865     margin = max(i[0] for i in textSizes.values()) if textSizes else 0

   866     return annots, textSizes, margin

   868 def lastDotplot(opts, args):

   869     logLevel = logging.INFO if opts.verbose else logging.WARNING

   870     logging.basicConfig(format="%(filename)s: %(message)s", level=logLevel)

   872     font = getFont(opts)

   873     image_mode = 'RGB'

   874     forward_color = ImageColor.getcolor(opts.forwardcolor, image_mode)

   875     reverse_color = ImageColor.getcolor(opts.reversecolor, image_mode)

   876     zipped_colors = zip(forward_color, reverse_color)

   877     overlap_color = tuple([(i + j) // 2 for i, j in zipped_colors])

   879     maxGap1, maxGapB1 = twoValuesFromOption(opts.max_gap1, ":")

   880     maxGap2, maxGapB2 = twoValuesFromOption(opts.max_gap2, ":")

   882     logging.info("reading alignments...")

   883     alnData = readAlignments(args[0], opts)

   884     alignments, seqRanges1, coverDict1, seqRanges2, coverDict2 = alnData

   885     if not alignments: raise Exception("there are no alignments")

   886     logging.info("cutting...")

   887     coverDict1 = dict(mergedRangesPerSeq(coverDict1))

   888     coverDict2 = dict(mergedRangesPerSeq(coverDict2))

   889     minAlignedBases = min(coveredLength(coverDict1), coveredLength(coverDict2))

   890     pad = int(opts.pad * minAlignedBases)

   891     cutRanges1 = list(trimmed(seqRanges1, coverDict1, minAlignedBases,

   892                               maxGap1, pad, pad))

   893     cutRanges2 = list(trimmed(seqRanges2, coverDict2, minAlignedBases,

   894                               maxGap2, pad, pad))

   896     biggestSeqs1 = biggestSequences(cutRanges1, opts.maxseqs)

   897     cutRanges1 = [i for i in cutRanges1 if i[0] in biggestSeqs1]

   898     alignments = [i for i in alignments if i[0] in biggestSeqs1]

   899     cutRanges2 = remainingSequenceRanges(cutRanges2, alignments, 1)

   901     biggestSeqs2 = biggestSequences(cutRanges2, opts.maxseqs)

   902     cutRanges2 = [i for i in cutRanges2 if i[0] in biggestSeqs2]

   903     alignments = [i for i in alignments if i[1] in biggestSeqs2]

   904     cutRanges1 = remainingSequenceRanges(cutRanges1, alignments, 0)

   906     logging.info("reading secondary alignments...")

   907     alnDataB = readSecondaryAlignments(opts, cutRanges1, cutRanges2)

   908     alignmentsB, seqRangesB1, coverDictB1, seqRangesB2, coverDictB2 = alnDataB

   909     logging.info("cutting...")

   910     coverDictB1 = dict(mergedRangesPerSeq(coverDictB1))

   911     coverDictB2 = dict(mergedRangesPerSeq(coverDictB2))

   912     cutRangesB1 = trimmed(seqRangesB1, coverDictB1, minAlignedBases,

   913                           maxGapB1, 0, 0)

   914     cutRangesB2 = trimmed(seqRangesB2, coverDictB2, minAlignedBases,

   915                           maxGapB2, 0, 0)

   917     logging.info("sorting...")

   918     sortOut = allSortedRanges(opts, alignments, alignmentsB,

   919                               cutRanges1, cutRangesB1, cutRanges2, cutRangesB2)

   920     sortedRanges1, sortedRanges2 = sortOut

   922     textDraw = None

   923     if opts.fontsize:

   924         textDraw = ImageDraw.Draw(Image.new(image_mode, (1, 1)))

   926     textRot1 = "vertical".startswith(opts.rot1)

   927     i1 = dataFromRanges(sortedRanges1, font, textDraw, opts.labels1, textRot1)

   928     rangeSizes1, labelData1, tMargin = i1

   930     textRot2 = "horizontal".startswith(opts.rot2)

   931     i2 = dataFromRanges(sortedRanges2, font, textDraw, opts.labels2, textRot2)

   932     rangeSizes2, labelData2, lMargin = i2

   934     logging.info("reading annotations...")

   936     a1 = readAnnotations(opts, font, textDraw, sortedRanges1,

   937                          opts.bed1, opts.rmsk1, opts.genePred1, opts.gap1)

   938     annots1, annoTextSizes1, bMargin = a1

   940     a2 = readAnnotations(opts, font, textDraw, sortedRanges2,

   941                          opts.bed2, opts.rmsk2, opts.genePred2, opts.gap2)

   942     annots2, annoTextSizes2, rMargin = a2

   944     maxPixels1 = opts.width  - lMargin - rMargin

   945     maxPixels2 = opts.height - tMargin - bMargin

   946     bpPerPix1 = get_bp_per_pix(rangeSizes1, opts.border_pixels, maxPixels1)

   947     bpPerPix2 = get_bp_per_pix(rangeSizes2, opts.border_pixels, maxPixels2)

   948     bpPerPix = max(bpPerPix1, bpPerPix2)

   949     logging.info("bp per pixel = " + str(bpPerPix))

   951     p1 = pixelData(rangeSizes1, bpPerPix, opts.border_pixels, lMargin)

   952     rangePixBegs1, rangePixLens1, rMarginBeg = p1

   953     width = rMarginBeg + rMargin

   954     rangeDict1 = dict(rangesAndOriginsPerSeq(sortedRanges1, rangePixBegs1,

   955                                              rangePixLens1, bpPerPix))

   957     p2 = pixelData(rangeSizes2, bpPerPix, opts.border_pixels, tMargin)

   958     rangePixBegs2, rangePixLens2, bMarginBeg = p2

   959     height = bMarginBeg + bMargin

   960     rangeDict2 = dict(rangesAndOriginsPerSeq(sortedRanges2, rangePixBegs2,

   961                                              rangePixLens2, bpPerPix))

   963     logging.info("width:  " + str(width))

   964     logging.info("height: " + str(height))

   966     logging.info("processing alignments...")

   967     allAlignments = alignments + alignmentsB

   968     hits = alignmentPixels(width, height, allAlignments, bpPerPix,

   969                            rangeDict1, rangeDict2)

   971     rangeDict1 = expandedSeqDict(rangeDict1)

   972     rangeDict2 = expandedSeqDict(rangeDict2)

   974     boxes1 = list(bedBoxes(annots1, rangeDict1, rMarginBeg, True, bpPerPix,

   975                            annoTextSizes1))

   976     boxes2 = list(bedBoxes(annots2, rangeDict2, bMarginBeg, False, bpPerPix,

   977                            annoTextSizes2))

   978     boxes = sorted(itertools.chain(boxes1, boxes2))

   980     logging.info("drawing...")

   982     image_size = width, height

   983     im = Image.new(image_mode, image_size, opts.background_color)

   985     drawAnnotations(im, boxes, tMargin, bMarginBeg, lMargin, rMarginBeg)

   987     joinA, joinB = twoValuesFromOption(opts.join, ":")

   988     if joinA in "13":

   989         drawJoins(im, alignments, bpPerPix, 0, rangeDict1, rangeDict2)

   990     if joinB in "13":

   991         drawJoins(im, alignmentsB, bpPerPix, 0, rangeDict1, rangeDict2)

   992     if joinA in "23":

   993         drawJoins(im, alignments, bpPerPix, 1, rangeDict2, rangeDict1)

   994     if joinB in "23":

   995         drawJoins(im, alignmentsB, bpPerPix, 1, rangeDict2, rangeDict1)

   997     for i in range(height):

   998         for j in range(width):

   999             store_value = hits[i * width + j]

  1000             xy = j, i

  1001             if   store_value == 1: im.putpixel(xy, forward_color)

  1002             elif store_value == 2: im.putpixel(xy, reverse_color)

  1003             elif store_value == 3: im.putpixel(xy, overlap_color)

  1005     if opts.fontsize != 0:

  1006         axis1 = axisImage(labelData1, rangePixBegs1, rangePixLens1,

  1007                           textRot1, False, font, image_mode, opts)

  1008         if textRot1:

  1009             axis1 = axis1.transpose(Image.ROTATE_90)

  1010         axis2 = axisImage(labelData2, rangePixBegs2, rangePixLens2,

  1011                           textRot2, textRot2, font, image_mode, opts)

  1012         if not textRot2:

  1013             axis2 = axis2.transpose(Image.ROTATE_270)

  1014         im.paste(axis1, (0, 0))

  1015         im.paste(axis2, (0, 0))

  1017         annoImage1 = annoTextImage(opts, image_mode, font, bMargin, width,

  1018                                    boxes1, False)

  1019         annoImage1 = annoImage1.transpose(Image.ROTATE_90)

  1020         annoImage2 = annoTextImage(opts, image_mode, font, rMargin, height,

  1021                                    boxes2, True)

  1022         im.paste(annoImage1, (0, bMarginBeg))

  1023         im.paste(annoImage2, (rMarginBeg, 0))

  1025     for i in rangePixBegs1[1:]:

  1026         box = i - opts.border_pixels, tMargin, i, bMarginBeg

  1027         im.paste(opts.border_color, box)

  1029     for i in rangePixBegs2[1:]:

  1030         box = lMargin, i - opts.border_pixels, rMarginBeg, i

  1031         im.paste(opts.border_color, box)

  1033     im.save(args[1])

  1035 if __name__ == "__main__":

  1036     usage = """%prog --help

  1037    or: %prog [options] maf-or-tab-alignments dotplot.png

  1038    or: %prog [options] maf-or-tab-alignments dotplot.gif

  1039    or: ..."""

  1040     description = "Draw a dotplot of pair-wise sequence alignments in MAF or tabular format."

  1041     op = optparse.OptionParser(usage=usage, description=description)

  1042     op.add_option("-v", "--verbose", action="count",

  1043                   help="show progress messages & data about the plot")

  1044     # Replace "width" & "height" with a single "length" option?

  1045     op.add_option("-x", "--width", metavar="INT", type="int", default=1000,

  1046                   help="maximum width in pixels (default: %default)")

  1047     op.add_option("-y", "--height", metavar="INT", type="int", default=1000,

  1048                   help="maximum height in pixels (default: %default)")

  1049     op.add_option("-m", "--maxseqs", type="int", default=100, metavar="M",

  1050                   help="maximum number of horizontal or vertical sequences "

  1051                   "(default=%default)")

  1052     op.add_option("-1", "--seq1", metavar="PATTERN", action="append",

  1053                   default=[],

  1054                   help="which sequences to show from the 1st genome")

  1055     op.add_option("-2", "--seq2", metavar="PATTERN", action="append",

  1056                   default=[],

  1057                   help="which sequences to show from the 2nd genome")

  1058     op.add_option("-c", "--forwardcolor", metavar="COLOR", default="red",

  1059                   help="color for forward alignments (default: %default)")

  1060     op.add_option("-r", "--reversecolor", metavar="COLOR", default="blue",

  1061                   help="color for reverse alignments (default: %default)")

  1062     op.add_option("--alignments", metavar="FILE", help="secondary alignments")

  1063     op.add_option("--sort1", default="1", metavar="N",

  1064                   help="genome1 sequence order: 0=input order, 1=name order, "

  1065                   "2=length order, 3=alignment order (default=%default)")

  1066     op.add_option("--sort2", default="1", metavar="N",

  1067                   help="genome2 sequence order: 0=input order, 1=name order, "

  1068                   "2=length order, 3=alignment order (default=%default)")

  1069     op.add_option("--strands1", default="0", metavar="N", help=

  1070                   "genome1 sequence orientation: 0=forward orientation, "

  1071                   "1=alignment orientation (default=%default)")

  1072     op.add_option("--strands2", default="0", metavar="N", help=

  1073                   "genome2 sequence orientation: 0=forward orientation, "

  1074                   "1=alignment orientation (default=%default)")

  1075     op.add_option("--max-gap1", metavar="FRAC", default="0.5,2", help=

  1076                   "maximum unaligned (end,mid) gap in genome1: "

  1077                   "fraction of aligned length (default=%default)")

  1078     op.add_option("--max-gap2", metavar="FRAC", default="0.5,2", help=

  1079                   "maximum unaligned (end,mid) gap in genome2: "

  1080                   "fraction of aligned length (default=%default)")

  1081     op.add_option("--pad", metavar="FRAC", type="float", default=0.04, help=

  1082                   "pad length when cutting unaligned gaps: "

  1083                   "fraction of aligned length (default=%default)")

  1084     op.add_option("-j", "--join", default="0", metavar="N", help=

  1085                   "join: 0=nothing, 1=alignments adjacent in genome1, "

  1086                   "2=alignments adjacent in genome2 (default=%default)")

  1087     op.add_option("--border-pixels", metavar="INT", type="int", default=1,

  1088                   help="number of pixels between sequences (default=%default)")

  1089     op.add_option("--border-color", metavar="COLOR", default="black",

  1090                   help="color for pixels between sequences (default=%default)")

  1091     # --break-color and/or --break-pixels for intra-sequence breaks?

  1092     op.add_option("--margin-color", metavar="COLOR", default="#dcdcdc",

  1093                   help="margin color")

  1095     og = optparse.OptionGroup(op, "Text options")

  1096     og.add_option("-f", "--fontfile", metavar="FILE",

  1097                   help="TrueType or OpenType font file")

  1098     og.add_option("-s", "--fontsize", metavar="SIZE", type="int", default=14,

  1099                   help="TrueType or OpenType font size (default: %default)")

  1100     og.add_option("--labels1", type="int", default=0, metavar="N", help=

  1101                   "genome1 labels: 0=name, 1=name:length, "

  1102                   "2=name:start:length, 3=name:start-end (default=%default)")

  1103     og.add_option("--labels2", type="int", default=0, metavar="N", help=

  1104                   "genome2 labels: 0=name, 1=name:length, "

  1105                   "2=name:start:length, 3=name:start-end (default=%default)")

  1106     og.add_option("--rot1", metavar="ROT", default="h",

  1107                   help="text rotation for the 1st genome (default=%default)")

  1108     og.add_option("--rot2", metavar="ROT", default="v",

  1109                   help="text rotation for the 2nd genome (default=%default)")

  1110     op.add_option_group(og)

  1112     og = optparse.OptionGroup(op, "Annotation options")

  1113     og.add_option("--bed1", metavar="FILE",

  1114                   help="read genome1 annotations from BED file")

  1115     og.add_option("--bed2", metavar="FILE",

  1116                   help="read genome2 annotations from BED file")

  1117     og.add_option("--rmsk1", metavar="FILE", help="read genome1 repeats from "

  1118                   "RepeatMasker .out or rmsk.txt file")

  1119     og.add_option("--rmsk2", metavar="FILE", help="read genome2 repeats from "

  1120                   "RepeatMasker .out or rmsk.txt file")

  1121     op.add_option_group(og)

  1123     og = optparse.OptionGroup(op, "Gene options")

  1124     og.add_option("--genePred1", metavar="FILE",

  1125                   help="read genome1 genes from genePred file")

  1126     og.add_option("--genePred2", metavar="FILE",

  1127                   help="read genome2 genes from genePred file")

  1128     og.add_option("--exon-color", metavar="COLOR", default="PaleGreen",

  1129                   help="color for exons (default=%default)")

  1130     og.add_option("--cds-color", metavar="COLOR", default="LimeGreen",

  1131                   help="color for protein-coding regions (default=%default)")

  1132     op.add_option_group(og)

  1134     og = optparse.OptionGroup(op, "Unsequenced gap options")

  1135     og.add_option("--gap1", metavar="FILE",

  1136                   help="read genome1 unsequenced gaps from agp or gap file")

  1137     og.add_option("--gap2", metavar="FILE",

  1138                   help="read genome2 unsequenced gaps from agp or gap file")

  1139     og.add_option("--bridged-color", metavar="COLOR", default="yellow",

  1140                   help="color for bridged gaps (default: %default)")

  1141     og.add_option("--unbridged-color", metavar="COLOR", default="orange",

  1142                   help="color for unbridged gaps (default: %default)")

  1143     op.add_option_group(og)

  1144     (opts, args) = op.parse_args()

  1145     if len(args) != 2: op.error("2 arguments needed")

  1147     opts.background_color = "white"

  1148     opts.label_space = 5     # minimum number of pixels between axis labels

  1150     try: lastDotplot(opts, args)

  1151     except KeyboardInterrupt: pass  # avoid silly error message

  1152     except Exception as e:

  1153         prog = os.path.basename(sys.argv[0])

  1154         sys.exit(prog + ": error: " + str(e))

author	Martin C. Frith
	Fri Jul 31 15:53:49 2020 +0900 (21 months ago)
changeset 1079	5aaae537a930
parent 1078	579d74b5066f
child 1132	2b38a62095d1
permissions	-rwxr-xr-x