scripts/last-dotplot
author Martin C. Frith
Thu May 18 14:46:45 2017 +0900 (2017-05-18)
changeset 861 a75b439f0cf2
parent 860 6e8ff424ce2b
child 863 6a4915d5b5cb
permissions -rwxr-xr-x
last-dotplot: do e.g. chr7 annotations -> hg19.chr7 only if no ambiguity
     1 #! /usr/bin/env python
     2 
     3 # Read pair-wise alignments in MAF or LAST tabular format: write an
     4 # "Oxford grid", a.k.a. dotplot.
     5 
     6 # TODO: Currently, pixels with zero aligned nt-pairs are white, and
     7 # pixels with one or more aligned nt-pairs are black.  This can look
     8 # too crowded for large genome alignments.  I tried shading each pixel
     9 # according to the number of aligned nt-pairs within it, but the
    10 # result is too faint.  How can this be done better?
    11 
    12 import fnmatch, itertools, optparse, os, re, sys
    13 
    14 # Try to make PIL/PILLOW work:
    15 try: from PIL import Image, ImageDraw, ImageFont, ImageColor
    16 except ImportError: import Image, ImageDraw, ImageFont, ImageColor
    17 
    18 def myOpen(fileName):  # faster than fileinput
    19     if fileName == "-":
    20         return sys.stdin
    21     return open(fileName)
    22 
    23 def warn(message):
    24     prog = os.path.basename(sys.argv[0])
    25     sys.stderr.write(prog + ": " + message + "\n")
    26 
    27 def croppedBlocks(blocks, range1, range2):
    28     cropBeg1, cropEnd1 = range1
    29     cropBeg2, cropEnd2 = range2
    30     if blocks[0][0] < 0: cropBeg1, cropEnd1 = -cropEnd1, -cropBeg1
    31     if blocks[0][1] < 0: cropBeg2, cropEnd2 = -cropEnd2, -cropBeg2
    32     for beg1, beg2, size in blocks:
    33         b1 = max(cropBeg1, beg1)
    34         e1 = min(cropEnd1, beg1 + size)
    35         if b1 >= e1: continue
    36         offset = beg2 - beg1
    37         b2 = max(cropBeg2, b1 + offset)
    38         e2 = min(cropEnd2, e1 + offset)
    39         if b2 >= e2: continue
    40         yield b2 - offset, b2, e2 - b2
    41 
    42 def tabBlocks(beg1, beg2, blocks):
    43     '''Get the gapless blocks of an alignment, from LAST tabular format.'''
    44     for i in blocks.split(","):
    45         if ":" in i:
    46             x, y = i.split(":")
    47             beg1 += int(x)
    48             beg2 += int(y)
    49         else:
    50             size = int(i)
    51             yield beg1, beg2, size
    52             beg1 += size
    53             beg2 += size
    54 
    55 def mafBlocks(beg1, beg2, seq1, seq2):
    56     '''Get the gapless blocks of an alignment, from MAF format.'''
    57     size = 0
    58     for x, y in itertools.izip(seq1, seq2):
    59         if x == "-":
    60             if size:
    61                 yield beg1, beg2, size
    62                 beg1 += size
    63                 beg2 += size
    64                 size = 0
    65             beg2 += 1
    66         elif y == "-":
    67             if size:
    68                 yield beg1, beg2, size
    69                 beg1 += size
    70                 beg2 += size
    71                 size = 0
    72             beg1 += 1
    73         else:
    74             size += 1
    75     if size: yield beg1, beg2, size
    76 
    77 def alignmentInput(lines):
    78     '''Get alignments and sequence lengths, from MAF or tabular format.'''
    79     mafCount = 0
    80     for line in lines:
    81         w = line.split()
    82         if line[0].isdigit():  # tabular format
    83             chr1, beg1, seqlen1 = w[1], int(w[2]), int(w[5])
    84             if w[4] == "-": beg1 -= seqlen1
    85             chr2, beg2, seqlen2 = w[6], int(w[7]), int(w[10])
    86             if w[9] == "-": beg2 -= seqlen2
    87             blocks = tabBlocks(beg1, beg2, w[11])
    88             yield chr1, seqlen1, chr2, seqlen2, blocks
    89         elif line[0] == "s":  # MAF format
    90             if mafCount == 0:
    91                 chr1, beg1, seqlen1, seq1 = w[1], int(w[2]), int(w[5]), w[6]
    92                 if w[4] == "-": beg1 -= seqlen1
    93                 mafCount = 1
    94             else:
    95                 chr2, beg2, seqlen2, seq2 = w[1], int(w[2]), int(w[5]), w[6]
    96                 if w[4] == "-": beg2 -= seqlen2
    97                 blocks = mafBlocks(beg1, beg2, seq1, seq2)
    98                 yield chr1, seqlen1, chr2, seqlen2, blocks
    99                 mafCount = 0
   100 
   101 def seqRangeFromText(text):
   102     if ":" in text:
   103         pattern, interval = text.rsplit(":", 1)
   104         if "-" in interval:
   105             beg, end = interval.rsplit("-", 1)
   106             return pattern, int(beg), int(end)  # beg may be negative
   107     return text, 0, sys.maxsize
   108 
   109 def rangeFromSeqName(seqRanges, name, seqLen):
   110     if not seqRanges: return 0, seqLen
   111     base = name.split(".")[-1]  # allow for names like hg19.chr7
   112     for pat, beg, end in seqRanges:
   113         if fnmatch.fnmatchcase(name, pat) or fnmatch.fnmatchcase(base, pat):
   114             return max(beg, 0), min(end, seqLen)
   115     return None
   116 
   117 def updateSeqs(isTrim, seqNames, seqLimits, seqName, seqRange, blocks, index):
   118     if seqName not in seqLimits:
   119         seqNames.append(seqName)
   120     if isTrim:
   121         beg = blocks[0][index]
   122         end = blocks[-1][index] + blocks[-1][2]
   123         if beg < 0: beg, end = -end, -beg
   124         if seqName in seqLimits:
   125             b, e = seqLimits[seqName]
   126             seqLimits[seqName] = min(b, beg), max(e, end)
   127         else:
   128             seqLimits[seqName] = beg, end
   129     else:
   130         seqLimits[seqName] = seqRange
   131 
   132 def readAlignments(fileName, opts):
   133     '''Get alignments and sequence limits, from MAF or tabular format.'''
   134     seqRanges1 = map(seqRangeFromText, opts.seq1)
   135     seqRanges2 = map(seqRangeFromText, opts.seq2)
   136 
   137     alignments = []
   138     seqNames1 = []
   139     seqNames2 = []
   140     seqLimits1 = {}
   141     seqLimits2 = {}
   142     lines = myOpen(fileName)
   143     for seqName1, seqLen1, seqName2, seqLen2, blocks in alignmentInput(lines):
   144         range1 = rangeFromSeqName(seqRanges1, seqName1, seqLen1)
   145         if not range1: continue
   146         range2 = rangeFromSeqName(seqRanges2, seqName2, seqLen2)
   147         if not range2: continue
   148         b = list(croppedBlocks(list(blocks), range1, range2))
   149         if not b: continue
   150         aln = seqName1, seqName2, b
   151         alignments.append(aln)
   152         updateSeqs(opts.trim1, seqNames1, seqLimits1, seqName1, range1, b, 0)
   153         updateSeqs(opts.trim2, seqNames2, seqLimits2, seqName2, range2, b, 1)
   154     return alignments, seqNames1, seqNames2, seqLimits1, seqLimits2
   155 
   156 def natural_sort_key(my_string):
   157     '''Return a sort key for "natural" ordering, e.g. chr9 < chr10.'''
   158     parts = re.split(r'(\d+)', my_string)
   159     parts[1::2] = map(int, parts[1::2])
   160     return parts
   161 
   162 def get_text_sizes(my_strings, font, fontsize, image_mode):
   163     '''Get widths & heights, in pixels, of some strings.'''
   164     if fontsize == 0: return [(0, 0) for i in my_strings]
   165     image_size = 1, 1
   166     im = Image.new(image_mode, image_size)
   167     draw = ImageDraw.Draw(im)
   168     return [draw.textsize(i, font=font) for i in my_strings]
   169 
   170 def sizeText(size):
   171     suffixes = "bp", "kb", "Mb", "Gb"
   172     for i, x in enumerate(suffixes):
   173         j = 10 ** (i * 3)
   174         if size < j * 10:
   175             return "%.2g" % (1.0 * size / j) + x
   176         if size < j * 1000 or i == len(suffixes) - 1:
   177             return "%.0f" % (1.0 * size / j) + x
   178 
   179 def seqNameAndSizeText(seqName, seqSize):
   180     return seqName + ": " + sizeText(seqSize)
   181 
   182 def getSeqInfo(sortOpt, seqNames, seqLimits,
   183                font, fontsize, image_mode, isShowSize):
   184     '''Return miscellaneous information about the sequences.'''
   185     if sortOpt == 1:
   186         seqNames.sort(key=natural_sort_key)
   187     seqSizes = [seqLimits[i][1] - seqLimits[i][0] for i in seqNames]
   188     if sortOpt == 2:
   189         seqRecords = sorted(zip(seqSizes, seqNames), reverse=True)
   190         seqSizes = [i[0] for i in seqRecords]
   191         seqNames = [i[1] for i in seqRecords]
   192     if isShowSize:
   193         seqLabels = map(seqNameAndSizeText, seqNames, seqSizes)
   194     else:
   195         seqLabels = seqNames
   196     labelSizes = get_text_sizes(seqLabels, font, fontsize, image_mode)
   197     margin = max(zip(*labelSizes)[1])  # maximum text height
   198     return seqNames, seqSizes, seqLabels, labelSizes, margin
   199 
   200 def div_ceil(x, y):
   201     '''Return x / y rounded up.'''
   202     q, r = divmod(x, y)
   203     return q + (r != 0)
   204 
   205 def tot_seq_pix(seq_sizes, bp_per_pix):
   206     '''Return the total pixels needed for sequences of the given sizes.'''
   207     return sum([div_ceil(i, bp_per_pix) for i in seq_sizes])
   208 
   209 def get_bp_per_pix(seq_sizes, pix_tween_seqs, pix_limit):
   210     '''Get the minimum bp-per-pixel that fits in the size limit.'''
   211     seq_num = len(seq_sizes)
   212     seq_pix_limit = pix_limit - pix_tween_seqs * (seq_num - 1)
   213     if seq_pix_limit < seq_num:
   214         raise Exception("can't fit the image: too many sequences?")
   215     lower_bound = div_ceil(sum(seq_sizes), seq_pix_limit)
   216     for bp_per_pix in itertools.count(lower_bound):  # slow linear search
   217         if tot_seq_pix(seq_sizes, bp_per_pix) <= seq_pix_limit: break
   218     return bp_per_pix
   219 
   220 def get_seq_starts(seq_pix, pix_tween_seqs, margin):
   221     '''Get the start pixel for each sequence.'''
   222     seq_starts = []
   223     pix_tot = margin - pix_tween_seqs
   224     for i in seq_pix:
   225         pix_tot += pix_tween_seqs
   226         seq_starts.append(pix_tot)
   227         pix_tot += i
   228     return seq_starts
   229 
   230 def get_pix_info(seq_sizes, bp_per_pix, pix_tween_seqs, margin):
   231     '''Return pixel information about the sequences.'''
   232     seq_pix = [div_ceil(i, bp_per_pix) for i in seq_sizes]
   233     seq_starts = get_seq_starts(seq_pix, pix_tween_seqs, margin)
   234     tot_pix = seq_starts[-1] + seq_pix[-1]
   235     return seq_pix, seq_starts, tot_pix
   236 
   237 def drawLineForward(hits, width, bp_per_pix, beg1, beg2, size):
   238     while True:
   239         q1, r1 = divmod(beg1, bp_per_pix)
   240         q2, r2 = divmod(beg2, bp_per_pix)
   241         hits[q2 * width + q1] |= 1
   242         next_pix = min(bp_per_pix - r1, bp_per_pix - r2)
   243         if next_pix >= size: break
   244         beg1 += next_pix
   245         beg2 += next_pix
   246         size -= next_pix
   247 
   248 def drawLineReverse(hits, width, bp_per_pix, beg1, beg2, size):
   249     beg2 = -1 - beg2
   250     while True:
   251         q1, r1 = divmod(beg1, bp_per_pix)
   252         q2, r2 = divmod(beg2, bp_per_pix)
   253         hits[q2 * width + q1] |= 2
   254         next_pix = min(bp_per_pix - r1, r2 + 1)
   255         if next_pix >= size: break
   256         beg1 += next_pix
   257         beg2 -= next_pix
   258         size -= next_pix
   259 
   260 def alignmentPixels(width, height, alignments, bp_per_pix, origins1, origins2):
   261     hits = [0] * (width * height)  # the image data
   262     for seq1, seq2, blocks in alignments:
   263         ori1 = origins1[seq1]
   264         ori2 = origins2[seq2]
   265         for beg1, beg2, size in blocks:
   266             if beg1 < 0:
   267                 beg1 = -(beg1 + size)
   268                 beg2 = -(beg2 + size)
   269             if beg2 >= 0:
   270                 drawLineForward(hits, width, bp_per_pix,
   271                                 beg1 + ori1, beg2 + ori2, size)
   272             else:
   273                 drawLineReverse(hits, width, bp_per_pix,
   274                                 beg1 + ori1, beg2 - ori2, size)
   275     return hits
   276 
   277 def expandedSeqDict(seqDict):
   278     '''Allow lookup by short sequence names, e.g. chr7 as well as hg19.chr7.'''
   279     newDict = seqDict.copy()
   280     for name, x in seqDict.items():
   281         if "." in name:
   282             base = name.split(".")[-1]
   283             if base in newDict:  # an ambiguous case was found:
   284                 return seqDict   # so give up completely
   285             newDict[base] = x
   286     return newDict
   287 
   288 def readBed(fileName, seqLimits):
   289     if not fileName: return
   290     for line in myOpen(fileName):
   291         w = line.split()
   292         if not w: continue
   293         seqName = w[0]
   294         if seqName not in seqLimits: continue
   295         beg = int(w[1])
   296         end = int(w[2])
   297         layer = 900
   298         color = "#ffe4ff"
   299         if len(w) > 4:
   300             if w[4] != ".":
   301                 layer = float(w[4])
   302             if len(w) > 5:
   303                 if len(w) > 8 and w[8].count(",") == 2:
   304                     color = "rgb(" + w[8] + ")"
   305                 elif w[5] == "+":
   306                     color = "#fff4f4"
   307                 elif w[5] == "-":
   308                     color = "#f4f4ff"
   309         yield layer, color, seqName, beg, end
   310 
   311 def commaSeparatedInts(text):
   312     return map(int, text.rstrip(",").split(","))
   313 
   314 def readGenePred(opts, fileName, seqLimits):
   315     if not fileName: return
   316     for line in myOpen(fileName):
   317         fields = line.split()
   318         if not fields: continue
   319         if fields[2] not in "+-": fields = fields[1:]
   320         seqName = fields[1]
   321         if seqName not in seqLimits: continue
   322         #strand = fields[2]
   323         cdsBeg = int(fields[5])
   324         cdsEnd = int(fields[6])
   325         exonBegs = commaSeparatedInts(fields[8])
   326         exonEnds = commaSeparatedInts(fields[9])
   327         for beg, end in zip(exonBegs, exonEnds):
   328             yield 300, opts.exon_color, seqName, beg, end
   329             b = max(beg, cdsBeg)
   330             e = min(end, cdsEnd)
   331             if b < e: yield 400, opts.cds_color, seqName, b, e
   332 
   333 def readRmsk(fileName, seqLimits):
   334     if not fileName: return
   335     for line in myOpen(fileName):
   336         fields = line.split()
   337         if len(fields) == 17:  # rmsk.txt
   338             seqName = fields[5]
   339             if seqName not in seqLimits: continue  # do this ASAP for speed
   340             beg = int(fields[6])
   341             end = int(fields[7])
   342             strand = fields[9]
   343             repeatClass = fields[11]
   344         elif len(fields) == 15:  # .out
   345             seqName = fields[4]
   346             if seqName not in seqLimits: continue
   347             beg = int(fields[5]) - 1
   348             end = int(fields[6])
   349             strand = fields[8]
   350             repeatClass = fields[10]
   351         else:
   352             continue
   353         if repeatClass in ("Low_complexity", "Simple_repeat"):
   354             yield 200, "#ffe4ff", seqName, beg, end
   355         elif strand == "+":
   356             yield 100, "#fff4f4", seqName, beg, end
   357         else:
   358             yield 100, "#f4f4ff", seqName, beg, end
   359 
   360 def isExtraFirstGapField(fields):
   361     return fields[4].isdigit()
   362 
   363 def readGaps(opts, fileName, seqLimits):
   364     '''Read locations of unsequenced gaps, from an agp or gap file.'''
   365     if not fileName: return
   366     for line in myOpen(fileName):
   367         w = line.split()
   368         if not w or w[0][0] == "#": continue
   369         if isExtraFirstGapField(w): w = w[1:]
   370         if w[4] not in "NU": continue
   371         seqName = w[0]
   372         if seqName not in seqLimits: continue
   373         end = int(w[2])
   374         beg = end - int(w[5])  # zero-based coordinate
   375         if w[7] == "yes":
   376             yield 3000, opts.bridged_color, seqName, beg, end
   377         else:
   378             yield 2000, opts.unbridged_color, seqName, beg, end
   379 
   380 def bedBoxes(beds, seqLimits, origins, margin, edge, isTop, bpPerPix):
   381     for layer, color, seqName, beg, end in beds:
   382         cropBeg, cropEnd = seqLimits[seqName]
   383         beg = max(beg, cropBeg)
   384         end = min(end, cropEnd)
   385         if beg >= end: continue
   386         ori = origins[seqName]
   387         if layer <= 1000:
   388             # include partly-covered pixels
   389             b = (ori + beg) // bpPerPix
   390             e = div_ceil(ori + end, bpPerPix)
   391         else:
   392             # exclude partly-covered pixels
   393             b = div_ceil(ori + beg, bpPerPix)
   394             e = (ori + end) // bpPerPix
   395             if e <= b: continue
   396         if isTop:
   397             box = b, margin, e, edge
   398         else:
   399             box = margin, b, edge, e
   400         yield layer, color, box
   401 
   402 def drawAnnotations(im, boxes):
   403     # xxx use partial transparency for different-color overlaps?
   404     for layer, color, box in boxes:
   405         im.paste(color, box)
   406 
   407 def make_label(text, text_size, range_start, range_size):
   408     '''Return an axis label with endpoint & sort-order information.'''
   409     text_width  = text_size[0]
   410     label_start = range_start + (range_size - text_width) // 2
   411     label_end   = label_start + text_width
   412     sort_key    = text_width - range_size
   413     return sort_key, label_start, label_end, text
   414 
   415 def get_nonoverlapping_labels(labels, label_space):
   416     '''Get a subset of non-overlapping axis labels, greedily.'''
   417     nonoverlapping_labels = []
   418     for i in labels:
   419         if True not in [i[1] < j[2] + label_space and j[1] < i[2] + label_space
   420                         for j in nonoverlapping_labels]:
   421             nonoverlapping_labels.append(i)
   422     return nonoverlapping_labels
   423 
   424 def get_axis_image(seqNames, name_sizes, seq_starts, seq_pix,
   425                    font, image_mode, opts):
   426     '''Make an image of axis labels.'''
   427     min_pos = seq_starts[0]
   428     max_pos = seq_starts[-1] + seq_pix[-1]
   429     height = max(zip(*name_sizes)[1])
   430     labels = map(make_label, seqNames, name_sizes, seq_starts, seq_pix)
   431     labels = [i for i in labels if i[1] >= min_pos and i[2] <= max_pos]
   432     labels.sort()
   433     labels = get_nonoverlapping_labels(labels, opts.label_space)
   434     image_size = max_pos, height
   435     im = Image.new(image_mode, image_size, opts.border_color)
   436     draw = ImageDraw.Draw(im)
   437     for i in labels:
   438         position = i[1], 0
   439         draw.text(position, i[3], font=font, fill=opts.text_color)
   440     return im
   441 
   442 def seqOrigins(seqNames, seq_starts, seqLimits, bp_per_pix):
   443     for i, j in zip(seqNames, seq_starts):
   444         yield i, bp_per_pix * j - seqLimits[i][0]
   445 
   446 def lastDotplot(opts, args):
   447     if opts.fontfile:  font = ImageFont.truetype(opts.fontfile, opts.fontsize)
   448     else:              font = ImageFont.load_default()
   449 
   450     image_mode = 'RGB'
   451     forward_color = ImageColor.getcolor(opts.forwardcolor, image_mode)
   452     reverse_color = ImageColor.getcolor(opts.reversecolor, image_mode)
   453     zipped_colors = zip(forward_color, reverse_color)
   454     overlap_color = tuple([(i + j) // 2 for i, j in zipped_colors])
   455 
   456     warn("reading alignments...")
   457     alignmentInfo = readAlignments(args[0], opts)
   458     alignments, seqNames1, seqNames2, seqLimits1, seqLimits2 = alignmentInfo
   459     warn("done")
   460     if not alignments: raise Exception("there are no alignments")
   461 
   462     i1 = getSeqInfo(opts.sort1, seqNames1, seqLimits1,
   463                     font, opts.fontsize, image_mode, opts.lengths1)
   464     seqNames1, seqSizes1, seqLabels1, labelSizes1, margin1 = i1
   465 
   466     i2 = getSeqInfo(opts.sort2, seqNames2, seqLimits2,
   467                     font, opts.fontsize, image_mode, opts.lengths2)
   468     seqNames2, seqSizes2, seqLabels2, labelSizes2, margin2 = i2
   469 
   470     warn("choosing bp per pixel...")
   471     pix_limit1 = opts.width  - margin1
   472     pix_limit2 = opts.height - margin2
   473     bpPerPix1 = get_bp_per_pix(seqSizes1, opts.border_pixels, pix_limit1)
   474     bpPerPix2 = get_bp_per_pix(seqSizes2, opts.border_pixels, pix_limit2)
   475     bpPerPix = max(bpPerPix1, bpPerPix2)
   476     warn("bp per pixel = " + str(bpPerPix))
   477 
   478     seq_pix1, seq_starts1, width  = get_pix_info(seqSizes1, bpPerPix,
   479                                                  opts.border_pixels, margin1)
   480     seq_pix2, seq_starts2, height = get_pix_info(seqSizes2, bpPerPix,
   481                                                  opts.border_pixels, margin2)
   482     warn("width:  " + str(width))
   483     warn("height: " + str(height))
   484 
   485     origins1 = dict(seqOrigins(seqNames1, seq_starts1, seqLimits1, bpPerPix))
   486     origins2 = dict(seqOrigins(seqNames2, seq_starts2, seqLimits2, bpPerPix))
   487 
   488     warn("processing alignments...")
   489     hits = alignmentPixels(width, height, alignments, bpPerPix,
   490                            origins1, origins2)
   491     warn("done")
   492 
   493     image_size = width, height
   494     im = Image.new(image_mode, image_size, opts.background_color)
   495 
   496     seqLimits1 = expandedSeqDict(seqLimits1)
   497     seqLimits2 = expandedSeqDict(seqLimits2)
   498     origins1 = expandedSeqDict(origins1)
   499     origins2 = expandedSeqDict(origins2)
   500 
   501     beds1 = itertools.chain(readBed(opts.bed1, seqLimits1),
   502                             readRmsk(opts.rmsk1, seqLimits1),
   503                             readGenePred(opts, opts.genePred1, seqLimits1),
   504                             readGaps(opts, opts.gap1, seqLimits1))
   505     b1 = bedBoxes(beds1, seqLimits1, origins1, margin2, height, True, bpPerPix)
   506 
   507     beds2 = itertools.chain(readBed(opts.bed2, seqLimits2),
   508                             readRmsk(opts.rmsk2, seqLimits2),
   509                             readGenePred(opts, opts.genePred2, seqLimits2),
   510                             readGaps(opts, opts.gap2, seqLimits2))
   511     b2 = bedBoxes(beds2, seqLimits2, origins2, margin1, width, False, bpPerPix)
   512 
   513     boxes = sorted(itertools.chain(b1, b2))
   514     drawAnnotations(im, boxes)
   515 
   516     for i in range(height):
   517         for j in range(width):
   518             store_value = hits[i * width + j]
   519             xy = j, i
   520             if   store_value == 1: im.putpixel(xy, forward_color)
   521             elif store_value == 2: im.putpixel(xy, reverse_color)
   522             elif store_value == 3: im.putpixel(xy, overlap_color)
   523 
   524     if opts.fontsize != 0:
   525         axis1 = get_axis_image(seqLabels1, labelSizes1, seq_starts1, seq_pix1,
   526                                font, image_mode, opts)
   527         axis2 = get_axis_image(seqLabels2, labelSizes2, seq_starts2, seq_pix2,
   528                                font, image_mode, opts)
   529         axis2 = axis2.transpose(Image.ROTATE_270)  # !!! bug hotspot
   530         im.paste(axis1, (0, 0))
   531         im.paste(axis2, (0, 0))
   532 
   533     for i in seq_starts1[1:]:
   534         box = i - opts.border_pixels, margin2, i, height
   535         im.paste(opts.border_color, box)
   536 
   537     for i in seq_starts2[1:]:
   538         box = margin1, i - opts.border_pixels, width, i
   539         im.paste(opts.border_color, box)
   540 
   541     im.save(args[1])
   542 
   543 if __name__ == "__main__":
   544     usage = """%prog --help
   545    or: %prog [options] maf-or-tab-alignments dotplot.png
   546    or: %prog [options] maf-or-tab-alignments dotplot.gif
   547    or: ..."""
   548     description = "Draw a dotplot of pair-wise sequence alignments in MAF or tabular format."
   549     op = optparse.OptionParser(usage=usage, description=description)
   550     op.add_option("-1", "--seq1", metavar="PATTERN", action="append",
   551                   default=[],
   552                   help="which sequences to show from the 1st genome")
   553     op.add_option("-2", "--seq2", metavar="PATTERN", action="append",
   554                   default=[],
   555                   help="which sequences to show from the 2nd genome")
   556     # Replace "width" & "height" with a single "length" option?
   557     op.add_option("-x", "--width", type="int", default=1000,
   558                   help="maximum width in pixels (default: %default)")
   559     op.add_option("-y", "--height", type="int", default=1000,
   560                   help="maximum height in pixels (default: %default)")
   561     op.add_option("-c", "--forwardcolor", metavar="COLOR", default="red",
   562                   help="color for forward alignments (default: %default)")
   563     op.add_option("-r", "--reversecolor", metavar="COLOR", default="blue",
   564                   help="color for reverse alignments (default: %default)")
   565     op.add_option("--sort1", type="int", default=1, metavar="N",
   566                   help="genome1 sequence order: 0=input order, 1=name order, "
   567                   "2=length order (default=%default)")
   568     op.add_option("--sort2", type="int", default=1, metavar="N",
   569                   help="genome2 sequence order: 0=input order, 1=name order, "
   570                   "2=length order (default=%default)")
   571     op.add_option("--trim1", action="store_true",
   572                   help="trim unaligned sequence flanks from the 1st genome")
   573     op.add_option("--trim2", action="store_true",
   574                   help="trim unaligned sequence flanks from the 2nd genome")
   575     op.add_option("--border-pixels", metavar="INT", type="int", default=1,
   576                   help="number of pixels between sequences (default=%default)")
   577     op.add_option("--border-color", metavar="COLOR", default="#dcdcdc",
   578                   help="color for pixels between sequences (default=%default)")
   579     # xxx --margin-color?
   580 
   581     og = optparse.OptionGroup(op, "Text options")
   582     og.add_option("-f", "--fontfile", metavar="FILE",
   583                   help="TrueType or OpenType font file")
   584     og.add_option("-s", "--fontsize", metavar="SIZE", type="int", default=11,
   585                   help="TrueType or OpenType font size (default: %default)")
   586     og.add_option("--lengths1", action="store_true",
   587                   help="show sequence lengths for the 1st (horizontal) genome")
   588     og.add_option("--lengths2", action="store_true",
   589                   help="show sequence lengths for the 2nd (vertical) genome")
   590     op.add_option_group(og)
   591 
   592     og = optparse.OptionGroup(op, "Annotation options")
   593     og.add_option("--bed1", metavar="FILE",
   594                   help="read genome1 annotations from BED file")
   595     og.add_option("--bed2", metavar="FILE",
   596                   help="read genome2 annotations from BED file")
   597     og.add_option("--rmsk1", metavar="FILE", help="read genome1 repeats from "
   598                   "RepeatMasker .out or rmsk.txt file")
   599     og.add_option("--rmsk2", metavar="FILE", help="read genome2 repeats from "
   600                   "RepeatMasker .out or rmsk.txt file")
   601     op.add_option_group(og)
   602 
   603     og = optparse.OptionGroup(op, "Gene options")
   604     og.add_option("--genePred1", metavar="FILE",
   605                   help="read genome1 genes from genePred file")
   606     og.add_option("--genePred2", metavar="FILE",
   607                   help="read genome2 genes from genePred file")
   608     og.add_option("--exon-color", metavar="COLOR", default="#dfd",
   609                   help="color for exons (default=%default)")
   610     og.add_option("--cds-color", metavar="COLOR", default="#bdb",
   611                   help="color for protein-coding regions (default=%default)")
   612     op.add_option_group(og)
   613 
   614     og = optparse.OptionGroup(op, "Unsequenced gap options")
   615     og.add_option("--gap1", metavar="FILE",
   616                   help="read genome1 unsequenced gaps from agp or gap file")
   617     og.add_option("--gap2", metavar="FILE",
   618                   help="read genome2 unsequenced gaps from agp or gap file")
   619     og.add_option("--bridged-color", metavar="COLOR", default="yellow",
   620                   help="color for bridged gaps (default: %default)")
   621     og.add_option("--unbridged-color", metavar="COLOR", default="pink",
   622                   help="color for unbridged gaps (default: %default)")
   623     op.add_option_group(og)
   624     (opts, args) = op.parse_args()
   625     if len(args) != 2: op.error("2 arguments needed")
   626 
   627     opts.text_color = "black"
   628     opts.background_color = "white"
   629     opts.label_space = 5     # minimum number of pixels between axis labels
   630 
   631     try: lastDotplot(opts, args)
   632     except KeyboardInterrupt: pass  # avoid silly error message
   633     except Exception, e:
   634         prog = os.path.basename(sys.argv[0])
   635         sys.exit(prog + ": error: " + str(e))