#! /usr/bin/env python3 # Author: Martin C. Frith 2008 # SPDX-License-Identifier: GPL-3.0-or-later # Read pair-wise alignments in MAF or LAST tabular format: write an # "Oxford grid", a.k.a. dotplot. # TODO: Currently, pixels with zero aligned nt-pairs are white, and # pixels with one or more aligned nt-pairs are black. This can look # too crowded for large genome alignments. I tried shading each pixel # according to the number of aligned nt-pairs within it, but the # result is too faint. How can this be done better? import collections import functools import gzip from fnmatch import fnmatchcase import logging from operator import itemgetter import subprocess import itertools, optparse, os, re, sys # Try to make PIL/PILLOW work: try: from PIL import Image, ImageDraw, ImageFont, ImageColor except ImportError: import Image, ImageDraw, ImageFont, ImageColor try: from future_builtins import zip except ImportError: pass def myOpen(fileName): # faster than fileinput if fileName is None: return [] if fileName == "-": return sys.stdin if fileName.endswith(".gz"): return gzip.open(fileName, "rt") # xxx dubious for Python2 return open(fileName) def groupByFirstItem(things): for k, v in itertools.groupby(things, itemgetter(0)): yield k, [i[1:] for i in v] def croppedBlocks(blocks, ranges1, ranges2): headBeg1, headBeg2, headSize = blocks[0] for r1 in ranges1: for r2 in ranges2: cropBeg1, cropEnd1 = r1 if headBeg1 < 0: cropBeg1, cropEnd1 = -cropEnd1, -cropBeg1 cropBeg2, cropEnd2 = r2 if headBeg2 < 0: cropBeg2, cropEnd2 = -cropEnd2, -cropBeg2 for beg1, beg2, size in blocks: b1 = max(cropBeg1, beg1) e1 = min(cropEnd1, beg1 + size) if b1 >= e1: continue offset = beg2 - beg1 b2 = max(cropBeg2, b1 + offset) e2 = min(cropEnd2, e1 + offset) if b2 >= e2: continue yield b2 - offset, b2, e2 - b2 def tabBlocks(beg1, beg2, blocks): '''Get the gapless blocks of an alignment, from LAST tabular format.''' for i in blocks.split(","): if ":" in i: x, y = i.split(":") beg1 += int(x) beg2 += int(y) else: size = int(i) yield beg1, beg2, size beg1 += size beg2 += size def mafBlocks(beg1, beg2, seq1, seq2): '''Get the gapless blocks of an alignment, from MAF format.''' size = 0 for x, y in zip(seq1, seq2): if x == "-": if size: yield beg1, beg2, size beg1 += size beg2 += size size = 0 beg2 += 1 elif y == "-": if size: yield beg1, beg2, size beg1 += size beg2 += size size = 0 beg1 += 1 else: size += 1 if size: yield beg1, beg2, size def alignmentFromSegment(qrySeqName, qrySeqLen, segment): refSeqLen = sys.maxsize # XXX refSeqName, refSeqBeg, qrySeqBeg, size = segment block = refSeqBeg, qrySeqBeg, size return refSeqName, refSeqLen, qrySeqName, qrySeqLen, [block] def alignmentInput(lines): '''Get alignments and sequence lengths, from MAF or tabular format.''' mafCount = 0 qrySeqName = "" segments = [] for line in lines: w = line.split() if line[0] == "#": pass elif len(w) == 1: for i in segments: yield alignmentFromSegment(qrySeqName, qrySeqLen, i) qrySeqName = w[0] qrySeqLen = 0 segments = [] elif len(w) == 2 and qrySeqName and w[1].isdigit(): qrySeqLen += int(w[1]) elif len(w) == 4 and qrySeqName and w[1].isdigit() and w[3].isdigit(): refSeqName, refSeqBeg, refSeqEnd = w[0], int(w[1]), int(w[3]) size = abs(refSeqEnd - refSeqBeg) if refSeqBeg > refSeqEnd: refSeqBeg = -refSeqBeg segments.append((refSeqName, refSeqBeg, qrySeqLen, size)) qrySeqLen += size elif line[0].isdigit(): # tabular format chr1, beg1, seqlen1 = w[1], int(w[2]), int(w[5]) if w[4] == "-": beg1 -= seqlen1 chr2, beg2, seqlen2 = w[6], int(w[7]), int(w[10]) if w[9] == "-": beg2 -= seqlen2 blocks = tabBlocks(beg1, beg2, w[11]) yield chr1, seqlen1, chr2, seqlen2, blocks elif line[0] == "s": # MAF format if mafCount == 0: chr1, beg1, seqlen1, seq1 = w[1], int(w[2]), int(w[5]), w[6] if w[4] == "-": beg1 -= seqlen1 mafCount = 1 else: chr2, beg2, seqlen2, seq2 = w[1], int(w[2]), int(w[5]), w[6] if w[4] == "-": beg2 -= seqlen2 blocks = mafBlocks(beg1, beg2, seq1, seq2) yield chr1, seqlen1, chr2, seqlen2, blocks mafCount = 0 for i in segments: yield alignmentFromSegment(qrySeqName, qrySeqLen, i) def seqRequestFromText(text): if ":" in text: pattern, interval = text.rsplit(":", 1) if "-" in interval: beg, end = interval.rsplit("-", 1) return pattern, int(beg), int(end) # beg may be negative return text, 0, sys.maxsize def rangesFromSeqName(seqRequests, name, seqLen): if seqRequests: base = name.split(".")[-1] # allow for names like hg19.chr7 for pat, beg, end in seqRequests: if fnmatchcase(name, pat) or fnmatchcase(base, pat): yield max(beg, 0), min(end, seqLen) else: yield 0, seqLen def updateSeqs(coverDict, seqRanges, seqName, ranges, coveredRange): beg, end = coveredRange if beg < 0: coveredRange = -end, -beg if seqName in coverDict: coverDict[seqName].append(coveredRange) else: coverDict[seqName] = [coveredRange] for beg, end in ranges: r = seqName, beg, end seqRanges.append(r) def readAlignments(fileName, opts): '''Get alignments and sequence limits, from MAF or tabular format.''' seqRequests1 = [seqRequestFromText(i) for i in opts.seq1] seqRequests2 = [seqRequestFromText(i) for i in opts.seq2] alignments = [] seqRanges1 = [] seqRanges2 = [] coverDict1 = {} coverDict2 = {} lines = myOpen(fileName) for seqName1, seqLen1, seqName2, seqLen2, blocks in alignmentInput(lines): ranges1 = sorted(rangesFromSeqName(seqRequests1, seqName1, seqLen1)) if not ranges1: continue ranges2 = sorted(rangesFromSeqName(seqRequests2, seqName2, seqLen2)) if not ranges2: continue b = list(croppedBlocks(list(blocks), ranges1, ranges2)) if not b: continue aln = seqName1, seqName2, b alignments.append(aln) coveredRange1 = b[0][0], b[-1][0] + b[-1][2] updateSeqs(coverDict1, seqRanges1, seqName1, ranges1, coveredRange1) coveredRange2 = b[0][1], b[-1][1] + b[-1][2] updateSeqs(coverDict2, seqRanges2, seqName2, ranges2, coveredRange2) return alignments, seqRanges1, coverDict1, seqRanges2, coverDict2 def nameAndRangesFromDict(cropDict, seqName): if seqName in cropDict: return seqName, cropDict[seqName] n = seqName.split(".")[-1] if n in cropDict: return n, cropDict[n] return seqName, [] def rangesForSecondaryAlignments(primaryRanges, seqLen): if primaryRanges: return primaryRanges return [(0, seqLen)] def readSecondaryAlignments(opts, cropRanges1, cropRanges2): cropDict1 = dict(groupByFirstItem(cropRanges1)) cropDict2 = dict(groupByFirstItem(cropRanges2)) alignments = [] seqRanges1 = [] seqRanges2 = [] coverDict1 = {} coverDict2 = {} lines = myOpen(opts.alignments) for seqName1, seqLen1, seqName2, seqLen2, blocks in alignmentInput(lines): seqName1, ranges1 = nameAndRangesFromDict(cropDict1, seqName1) seqName2, ranges2 = nameAndRangesFromDict(cropDict2, seqName2) if not ranges1 and not ranges2: continue r1 = rangesForSecondaryAlignments(ranges1, seqLen1) r2 = rangesForSecondaryAlignments(ranges2, seqLen2) b = list(croppedBlocks(list(blocks), r1, r2)) if not b: continue aln = seqName1, seqName2, b alignments.append(aln) if not ranges1: coveredRange1 = b[0][0], b[-1][0] + b[-1][2] updateSeqs(coverDict1, seqRanges1, seqName1, r1, coveredRange1) if not ranges2: coveredRange2 = b[0][1], b[-1][1] + b[-1][2] updateSeqs(coverDict2, seqRanges2, seqName2, r2, coveredRange2) return alignments, seqRanges1, coverDict1, seqRanges2, coverDict2 def twoValuesFromOption(text, separator): if separator in text: return text.split(separator) return text, text def mergedRanges(ranges): oldBeg, maxEnd = ranges[0] for beg, end in ranges: if beg > maxEnd: yield oldBeg, maxEnd oldBeg = beg maxEnd = end elif end > maxEnd: maxEnd = end yield oldBeg, maxEnd def mergedRangesPerSeq(coverDict): for k, v in coverDict.items(): v.sort() yield k, list(mergedRanges(v)) def coveredLength(mergedCoverDict): return sum(sum(e - b for b, e in v) for v in mergedCoverDict.values()) def trimmed(seqRanges, coverDict, minAlignedBases, maxGapFrac, endPad, midPad): maxEndGapFrac, maxMidGapFrac = twoValuesFromOption(maxGapFrac, ",") maxEndGap = max(float(maxEndGapFrac) * minAlignedBases, endPad * 1.0) maxMidGap = max(float(maxMidGapFrac) * minAlignedBases, midPad * 2.0) for seqName, rangeBeg, rangeEnd in seqRanges: seqBlocks = coverDict[seqName] blocks = [i for i in seqBlocks if i[0] < rangeEnd and i[1] > rangeBeg] if blocks[0][0] - rangeBeg > maxEndGap: rangeBeg = blocks[0][0] - endPad for j, y in enumerate(blocks): if j: x = blocks[j - 1] if y[0] - x[1] > maxMidGap: yield seqName, rangeBeg, x[1] + midPad rangeBeg = y[0] - midPad if rangeEnd - blocks[-1][1] > maxEndGap: rangeEnd = blocks[-1][1] + endPad yield seqName, rangeBeg, rangeEnd def rangesWithStrandInfo(seqRanges, strandOpt, alignments, seqIndex): if strandOpt == "1": forwardMinusReverse = collections.defaultdict(int) for i in alignments: blocks = i[2] beg1, beg2, size = blocks[0] numOfAlignedLetterPairs = sum(i[2] for i in blocks) if (beg1 < 0) != (beg2 < 0): # opposite-strand alignment numOfAlignedLetterPairs *= -1 forwardMinusReverse[i[seqIndex]] += numOfAlignedLetterPairs strandNum = 0 for seqName, beg, end in seqRanges: if strandOpt == "1": strandNum = 1 if forwardMinusReverse[seqName] >= 0 else 2 yield seqName, beg, end, strandNum def natural_sort_key(my_string): '''Return a sort key for "natural" ordering, e.g. chr9 < chr10.''' parts = re.split(r'(\d+)', my_string) parts[1::2] = map(int, parts[1::2]) return parts def nameKey(oneSeqRanges): return natural_sort_key(oneSeqRanges[0][0]) def sizeKey(oneSeqRanges): return sum(b - e for n, b, e, s in oneSeqRanges), nameKey(oneSeqRanges) def alignmentKey(seqNamesToLists, oneSeqRanges): seqName = oneSeqRanges[0][0] alignmentsOfThisSequence = seqNamesToLists[seqName] numOfAlignedLetterPairs = sum(i[3] for i in alignmentsOfThisSequence) toMiddle = numOfAlignedLetterPairs // 2 for i in alignmentsOfThisSequence: toMiddle -= i[3] if toMiddle < 0: return i[1:3] # sequence-rank and "position" of this alignment def rankAndFlipPerSeq(seqRanges): rangesGroupedBySeqName = itertools.groupby(seqRanges, itemgetter(0)) for rank, group in enumerate(rangesGroupedBySeqName): seqName, ranges = group strandNum = next(ranges)[3] flip = 1 if strandNum < 2 else -1 yield seqName, (rank, flip) def alignmentSortData(alignments, seqIndex, otherNamesToRanksAndFlips): otherIndex = 1 - seqIndex for i in alignments: blocks = i[2] otherRank, otherFlip = otherNamesToRanksAndFlips[i[otherIndex]] otherPos = otherFlip * abs(blocks[0][otherIndex] + blocks[-1][otherIndex] + blocks[-1][2]) numOfAlignedLetterPairs = sum(i[2] for i in blocks) yield i[seqIndex], otherRank, otherPos, numOfAlignedLetterPairs def mySortedRanges(seqRanges, sortOpt, seqIndex, alignments, otherRanges): rangesGroupedBySeqName = itertools.groupby(seqRanges, itemgetter(0)) g = [list(ranges) for seqName, ranges in rangesGroupedBySeqName] for i in g: if i[0][3] > 1: i.reverse() if sortOpt == "1": g.sort(key=nameKey) if sortOpt == "2": g.sort(key=sizeKey) if sortOpt == "3": otherNamesToRanksAndFlips = dict(rankAndFlipPerSeq(otherRanges)) alns = sorted(alignmentSortData(alignments, seqIndex, otherNamesToRanksAndFlips)) alnsGroupedBySeqName = itertools.groupby(alns, itemgetter(0)) seqNamesToLists = dict((k, list(v)) for k, v in alnsGroupedBySeqName) g.sort(key=functools.partial(alignmentKey, seqNamesToLists)) return [j for i in g for j in i] def allSortedRanges(opts, alignments, alignmentsB, seqRanges1, seqRangesB1, seqRanges2, seqRangesB2): o1, oB1 = twoValuesFromOption(opts.strands1, ":") o2, oB2 = twoValuesFromOption(opts.strands2, ":") if o1 == "1" and o2 == "1": raise RuntimeError("the strand options have circular dependency") seqRanges1 = list(rangesWithStrandInfo(seqRanges1, o1, alignments, 0)) seqRanges2 = list(rangesWithStrandInfo(seqRanges2, o2, alignments, 1)) seqRangesB1 = list(rangesWithStrandInfo(seqRangesB1, oB1, alignmentsB, 0)) seqRangesB2 = list(rangesWithStrandInfo(seqRangesB2, oB2, alignmentsB, 1)) o1, oB1 = twoValuesFromOption(opts.sort1, ":") o2, oB2 = twoValuesFromOption(opts.sort2, ":") if o1 == "3" and o2 == "3": raise RuntimeError("the sort options have circular dependency") if o1 != "3": s1 = mySortedRanges(seqRanges1, o1, None, None, None) if o2 != "3": s2 = mySortedRanges(seqRanges2, o2, None, None, None) if o1 == "3": s1 = mySortedRanges(seqRanges1, o1, 0, alignments, s2) if o2 == "3": s2 = mySortedRanges(seqRanges2, o2, 1, alignments, s1) t1 = mySortedRanges(seqRangesB1, oB1, 0, alignmentsB, s2) t2 = mySortedRanges(seqRangesB2, oB2, 1, alignmentsB, s1) return s1 + t1, s2 + t2 def myTextsize(textDraw, font, text): try: out = textDraw.textsize(text, font=font) except AttributeError: a, b, c, d = textDraw.textbbox((0, 0), text, font=font) out = c, d return out def sizesPerText(texts, font, textDraw): sizes = 0, 0 for t in texts: if textDraw is not None: sizes = myTextsize(textDraw, font, t) yield t, sizes def prettyNum(n): t = str(n) groups = [] while t: groups.append(t[-3:]) t = t[:-3] return ",".join(reversed(groups)) def sizeText(size): suffixes = "bp", "kb", "Mb", "Gb" for i, x in enumerate(suffixes): j = 10 ** (i * 3) if size < j * 10: return "%.2g" % (1.0 * size / j) + x if size < j * 1000 or i == len(suffixes) - 1: return "%.0f" % (1.0 * size / j) + x def labelText(seqRange, labelOpt): seqName, beg, end, strandNum = seqRange if labelOpt == 1: return seqName + ": " + sizeText(end - beg) if labelOpt == 2: return seqName + ":" + prettyNum(beg) + ": " + sizeText(end - beg) if labelOpt == 3: return seqName + ":" + prettyNum(beg) + "-" + prettyNum(end) return seqName def rangeLabels(seqRanges, labelOpt, font, textDraw, textRot): x = y = 0 for r in seqRanges: text = labelText(r, labelOpt) if textDraw is not None: x, y = myTextsize(textDraw, font, text) if textRot: x, y = y, x yield text, x, y, r[3] def dataFromRanges(sortedRanges, font, textDraw, labelOpt, textRot): for seqName, rangeBeg, rangeEnd, strandNum in sortedRanges: out = [seqName, str(rangeBeg), str(rangeEnd)] if strandNum > 0: out.append(".+-"[strandNum]) logging.info("\t".join(out)) logging.info("") rangeSizes = [e - b for n, b, e, s in sortedRanges] labs = list(rangeLabels(sortedRanges, labelOpt, font, textDraw, textRot)) margin = max(i[2] for i in labs) # xxx the margin may be too big, because some labels may get omitted return rangeSizes, labs, margin def div_ceil(x, y): '''Return x / y rounded up.''' q, r = divmod(x, y) return q + (r != 0) def get_bp_per_pix(rangeSizes, pixTweenRanges, maxPixels): '''Get the minimum bp-per-pixel that fits in the size limit.''' logging.info("choosing bp per pixel...") numOfRanges = len(rangeSizes) maxPixelsInRanges = maxPixels - pixTweenRanges * (numOfRanges - 1) if maxPixelsInRanges < numOfRanges: raise RuntimeError("can't fit the image: too many sequences?") negLimit = -maxPixelsInRanges negBpPerPix = sum(rangeSizes) // negLimit while True: if sum(i // negBpPerPix for i in rangeSizes) >= negLimit: return -negBpPerPix negBpPerPix -= 1 def getRangePixBegs(rangePixLens, pixTweenRanges, margin): '''Get the start pixel for each range.''' rangePixBegs = [] pix_tot = margin - pixTweenRanges for i in rangePixLens: pix_tot += pixTweenRanges rangePixBegs.append(pix_tot) pix_tot += i return rangePixBegs def pixelData(rangeSizes, bp_per_pix, pixTweenRanges, margin): '''Return pixel information about the ranges.''' rangePixLens = [div_ceil(i, bp_per_pix) for i in rangeSizes] rangePixBegs = getRangePixBegs(rangePixLens, pixTweenRanges, margin) tot_pix = rangePixBegs[-1] + rangePixLens[-1] return rangePixBegs, rangePixLens, tot_pix def drawLineForward(hits, width, bp_per_pix, beg1, beg2, size): while True: q1, r1 = divmod(beg1, bp_per_pix) q2, r2 = divmod(beg2, bp_per_pix) hits[q2 * width + q1] |= 1 next_pix = min(bp_per_pix - r1, bp_per_pix - r2) if next_pix >= size: break beg1 += next_pix beg2 += next_pix size -= next_pix def drawLineReverse(hits, width, bp_per_pix, beg1, beg2, size): while True: q1, r1 = divmod(beg1, bp_per_pix) q2, r2 = divmod(beg2, bp_per_pix) hits[q2 * width + q1] |= 2 next_pix = min(bp_per_pix - r1, r2 + 1) if next_pix >= size: break beg1 += next_pix beg2 -= next_pix size -= next_pix def strandAndOrigin(ranges, beg, size): isReverseStrand = (beg < 0) if isReverseStrand: beg = -(beg + size) for rangeBeg, rangeEnd, isReverseRange, origin in ranges: if rangeEnd > beg: # assumes the ranges are sorted return (isReverseStrand != isReverseRange), origin def alignmentPixels(width, height, alignments, bp_per_pix, rangeDict1, rangeDict2): hits = [0] * (width * height) # the image data for seq1, seq2, blocks in alignments: beg1, beg2, size = blocks[0] isReverse1, ori1 = strandAndOrigin(rangeDict1[seq1], beg1, size) isReverse2, ori2 = strandAndOrigin(rangeDict2[seq2], beg2, size) for beg1, beg2, size in blocks: if isReverse1: beg1 = -(beg1 + size) beg2 = -(beg2 + size) if isReverse1 == isReverse2: drawLineForward(hits, width, bp_per_pix, ori1 + beg1, ori2 + beg2, size) else: drawLineReverse(hits, width, bp_per_pix, ori1 + beg1, ori2 - beg2 - 1, size) return hits def orientedBlocks(alignments, seqIndex): otherIndex = 1 - seqIndex for a in alignments: seq1, seq2, blocks = a for b in blocks: beg1, beg2, size = b if b[seqIndex] < 0: b = -(beg1 + size), -(beg2 + size), size yield a[seqIndex], b[seqIndex], a[otherIndex], b[otherIndex], size def drawJoins(im, alignments, bpPerPix, seqIndex, rangeDict1, rangeDict2): blocks = orientedBlocks(alignments, seqIndex) oldSeq1 = "" for seq1, beg1, seq2, beg2, size in sorted(blocks): isReverse1, ori1 = strandAndOrigin(rangeDict1[seq1], beg1, size) isReverse2, ori2 = strandAndOrigin(rangeDict2[seq2], beg2, size) end1 = beg1 + size - 1 end2 = beg2 + size - 1 if isReverse1: beg1 = -(beg1 + 1) end1 = -(end1 + 1) if isReverse2: beg2 = -(beg2 + 1) end2 = -(end2 + 1) newPix1 = (ori1 + beg1) // bpPerPix newPix2 = (ori2 + beg2) // bpPerPix if seq1 == oldSeq1: lowerPix2 = min(oldPix2, newPix2) upperPix2 = max(oldPix2, newPix2) midPix1 = (oldPix1 + newPix1) // 2 if isReverse1: midPix1 = (oldPix1 + newPix1 + 1) // 2 oldPix1, newPix1 = newPix1, oldPix1 if upperPix2 - lowerPix2 > 1 and oldPix1 <= newPix1 <= oldPix1 + 1: if seqIndex == 0: box = midPix1, lowerPix2, midPix1 + 1, upperPix2 + 1 else: box = lowerPix2, midPix1, upperPix2 + 1, midPix1 + 1 im.paste("lightgray", box) oldPix1 = (ori1 + end1) // bpPerPix oldPix2 = (ori2 + end2) // bpPerPix oldSeq1 = seq1 def expandedSeqDict(seqDict): '''Allow lookup by short sequence names, e.g. chr7 as well as hg19.chr7.''' newDict = seqDict.copy() for name, x in seqDict.items(): if "." in name: base = name.split(".")[-1] if base in newDict: # an ambiguous case was found: return seqDict # so give up completely newDict[base] = x return newDict def commaSeparatedInts(text): return map(int, text.rstrip(",").split(",")) def annotsFromBedOrAgp(opts, rangeDict, fields): seqName = fields[0] if seqName not in rangeDict: return end = int(fields[2]) if len(fields) > 7 and fields[4].isalpha(): # agp format, or gap.txt if fields[4] in "NU" and fields[5].isdigit(): beg = end - int(fields[5]) # zero-based coordinate if fields[7] == "yes": yield 30000, opts.bridged_color, seqName, beg, end, "" else: yield 20000, opts.unbridged_color, seqName, beg, end, "" else: # BED format beg = int(fields[1]) itemName = fields[3] if len(fields) > 3 and fields[3] != "." else "" layer = 900 color = "#fbf" if len(fields) > 4: if fields[4] != ".": layer = float(fields[4]) if len(fields) > 5: if len(fields) > 8 and fields[8].count(",") == 2: color = "rgb(" + fields[8] + ")" else: strand = fields[5] isRev = (rangeDict[seqName][0][3] > 1) if strand == "+" and not isRev or strand == "-" and isRev: color = "#ffe8e8" if strand == "-" and not isRev or strand == "+" and isRev: color = "#e8e8ff" yield layer, color, seqName, beg, end, itemName def annotsFromGpd(opts, rangeDict, fields, geneName): # xxx split on tabs? seqName = fields[1] if seqName in rangeDict: cdsBeg = int(fields[5]) cdsEnd = int(fields[6]) exonBegs = commaSeparatedInts(fields[8]) exonEnds = commaSeparatedInts(fields[9]) for beg, end in zip(exonBegs, exonEnds): yield 300, opts.exon_color, seqName, beg, end, geneName b = max(beg, cdsBeg) e = min(end, cdsEnd) if b < e: yield 400, opts.cds_color, seqName, b, e, "" def annotsFromGff(opts, line, seqName): fields = line.rstrip().split("\t") feature = fields[2] beg = int(fields[3]) - 1 end = int(fields[4]) if feature == "exon": geneName = fields[8] if ";" in geneName or "=" in geneName: parts = geneName.rstrip(";").split(";") attrs = dict(re.split('[= ]', i.strip(), 1) for i in parts) if "gene" in attrs: geneName = attrs["gene"] # seems good for NCBI gff elif "Name" in attrs: geneName = attrs["Name"] else: geneName = "" yield 300, opts.exon_color, seqName, beg, end, geneName elif feature == "CDS": yield 400, opts.cds_color, seqName, beg, end, "" def annotsFromRep(rangeDict, seqName, beg, end, strand, repName, repClass): simple = "Low_complexity", "Simple_repeat", "Satellite" if repClass.startswith(simple): yield 200, "#fbf", seqName, beg, end, repName elif (strand == "+") != (rangeDict[seqName][0][3] > 1): yield 100, "#ffe8e8", seqName, beg, end, repName else: yield 100, "#e8e8ff", seqName, beg, end, repName def annotsFromFiles(opts, fileNames, rangeDict): isDig = str.isdigit for fileName in fileNames: for line in myOpen(fileName): w = line.split() n = len(w) if n > 10 and w[8] in "+C-" and isDig(w[5]) and isDig(w[6]): seq = w[4] # RepeatMasker .out if seq not in rangeDict: continue # do this ASAP for speed beg = int(w[5]) - 1 end = int(w[6]) g = annotsFromRep(rangeDict, seq, beg, end, w[8], w[9], w[10]) elif n > 11 and w[9] in "+-" and isDig(w[6]) and isDig(w[7]): seq = w[5] # rmsk.txt if seq not in rangeDict: continue beg = int(w[6]) end = int(w[7]) g = annotsFromRep(rangeDict, seq, beg, end, w[9], w[10], w[11]) elif n > 8 and w[6] in "+-." and isDig(w[3]) and isDig(w[4]): seqName = w[0] if seqName not in rangeDict: continue g = annotsFromGff(opts, line, seqName) elif n > 9 and w[2] in "+-" and isDig(w[4] + w[5] + w[6]): geneName = w[12 if n > 12 else 0] # XXX ??? g = annotsFromGpd(opts, rangeDict, w, geneName) elif n > 10 and w[3] in "+-" and isDig(w[5] + w[6] + w[7]): geneName = w[12 if n > 12 else 0] # XXX ??? g = annotsFromGpd(opts, rangeDict, w[1:], geneName) elif n > 2 and isDig(w[1]) and isDig(w[2]): g = annotsFromBedOrAgp(opts, rangeDict, w) elif n > 3 and isDig(w[2]) and isDig(w[3]): g = annotsFromBedOrAgp(opts, rangeDict, w[1:]) else: continue if line[0] == "#": continue for i in g: layer, color, seqName, beg, end, name = i if any(beg < r[2] and end > r[1] for r in rangeDict[seqName]): yield i def bedBoxes(annots, rangeDict, limit, isTop, bpPerPix): beds, textSizes, margin = annots cover = [(limit, limit)] for layer, color, seqName, bedBeg, bedEnd, name in reversed(beds): textWidth, textHeight = textSizes[name] for rangeBeg, rangeEnd, isReverseRange, origin in rangeDict[seqName]: beg = max(bedBeg, rangeBeg) end = min(bedEnd, rangeEnd) if beg >= end: continue if isReverseRange: beg, end = -end, -beg if layer <= 10000: # include partly-covered pixels pixBeg = (origin + beg) // bpPerPix pixEnd = div_ceil(origin + end, bpPerPix) else: # exclude partly-covered pixels pixBeg = div_ceil(origin + beg, bpPerPix) pixEnd = (origin + end) // bpPerPix if pixEnd <= pixBeg: continue if bedEnd >= rangeEnd: # include partly-covered end pixels if isReverseRange: pixBeg = (origin + beg) // bpPerPix else: pixEnd = div_ceil(origin + end, bpPerPix) nameBeg = (pixBeg + pixEnd - textHeight) // 2 nameEnd = nameBeg + textHeight n = "" if name and all(e <= nameBeg or b >= nameEnd for b, e in cover): if textWidth <= margin: cover.append((nameBeg, nameEnd)) n = name yield layer, color, isTop, pixBeg, pixEnd, n, nameBeg, textWidth def drawAnnotations(im, boxes, tMargin, bMarginBeg, lMargin, rMarginBeg): # xxx use partial transparency for different-color overlaps? for layer, color, isTop, beg, end, name, nameBeg, nameLen in boxes: if isTop: box = beg, tMargin, end, bMarginBeg else: box = lMargin, beg, rMarginBeg, end im.paste(color, box) def placedLabels(labels, rangePixBegs, rangePixLens, beg, end): '''Return axis labels with endpoint & sort-order information.''' maxWidth = end - beg for i, j, k in zip(labels, rangePixBegs, rangePixLens): text, textWidth, textHeight, strandNum = i if textWidth > maxWidth: continue labelBeg = j + (k - textWidth) // 2 labelEnd = labelBeg + textWidth sortKey = textWidth - k if labelBeg < beg: sortKey += maxWidth * (beg - labelBeg) labelBeg = beg labelEnd = beg + textWidth if labelEnd > end: sortKey += maxWidth * (labelEnd - end) labelEnd = end labelBeg = end - textWidth yield sortKey, labelBeg, labelEnd, text, textHeight, strandNum def nonoverlappingLabels(labels, minPixTweenLabels): '''Get a subset of non-overlapping axis labels, greedily.''' out = [] for i in labels: beg = i[1] - minPixTweenLabels end = i[2] + minPixTweenLabels if all(j[2] <= beg or j[1] >= end for j in out): out.append(i) return out def axisImage(labels, rangePixBegs, rangePixLens, textRot, textAln, font, image_mode, opts): '''Make an image of axis labels.''' beg = rangePixBegs[0] end = rangePixBegs[-1] + rangePixLens[-1] margin = max(i[2] for i in labels) labels = sorted(placedLabels(labels, rangePixBegs, rangePixLens, beg, end)) minPixTweenLabels = 0 if textRot else opts.label_space labels = nonoverlappingLabels(labels, minPixTweenLabels) image_size = (margin, end) if textRot else (end, margin) im = Image.new(image_mode, image_size, opts.margin_color) draw = ImageDraw.Draw(im) for sortKey, labelBeg, labelEnd, text, textHeight, strandNum in labels: base = margin - textHeight if textAln else 0 position = (base, labelBeg) if textRot else (labelBeg, base) fill = ("black", opts.forwardcolor, opts.reversecolor)[strandNum] draw.text(position, text, font=font, fill=fill) return im def annoTextImage(opts, image_mode, font, margin, length, boxes, isLeftAlign): image_size = margin, length im = Image.new(image_mode, image_size, opts.margin_color) draw = ImageDraw.Draw(im) for layer, color, isTop, beg, end, name, nameBeg, nameLen in boxes: xPosition = 0 if isLeftAlign else margin - nameLen position = xPosition, nameBeg draw.text(position, name, font=font, fill="black") return im def rangesPerSeq(sortedRanges): for seqName, group in itertools.groupby(sortedRanges, itemgetter(0)): yield seqName, sorted(group) def rangesWithOrigins(sortedRanges, rangePixBegs, rangePixLens, bpPerPix): for i, j, k in zip(sortedRanges, rangePixBegs, rangePixLens): seqName, rangeBeg, rangeEnd, strandNum = i isReverseRange = (strandNum > 1) if isReverseRange: origin = bpPerPix * (j + k) + rangeBeg else: origin = bpPerPix * j - rangeBeg yield seqName, (rangeBeg, rangeEnd, isReverseRange, origin) def rangesAndOriginsPerSeq(sortedRanges, rangePixBegs, rangePixLens, bpPerPix): a = rangesWithOrigins(sortedRanges, rangePixBegs, rangePixLens, bpPerPix) for seqName, group in itertools.groupby(a, itemgetter(0)): yield seqName, sorted(i[1] for i in group) def getFont(opts): if opts.fontfile: return ImageFont.truetype(opts.fontfile, opts.fontsize) fileNames = [] try: x = ["fc-match", "-f%{file}", "arial"] p = subprocess.Popen(x, stdout=subprocess.PIPE, stderr=subprocess.PIPE, universal_newlines=True) out, err = p.communicate() fileNames.append(out) except OSError as e: logging.info("fc-match error: " + str(e)) fileNames.append("/Library/Fonts/Arial.ttf") # for Mac for i in fileNames: try: font = ImageFont.truetype(i, opts.fontsize) logging.info("font: " + i) return font except IOError as e: logging.info("font load error: " + str(e)) return ImageFont.load_default() def sequenceSizesAndNames(seqRanges): for seqName, ranges in itertools.groupby(seqRanges, itemgetter(0)): size = sum(e - b for n, b, e in ranges) yield size, seqName def biggestSequences(seqRanges, maxNumOfSequences): s = sorted(sequenceSizesAndNames(seqRanges), reverse=True) if len(s) > maxNumOfSequences: logging.warning("too many sequences - discarding the smallest ones") s = s[:maxNumOfSequences] return set(i[1] for i in s) def remainingSequenceRanges(seqRanges, alignments, seqIndex): remainingSequences = set(i[seqIndex] for i in alignments) return [i for i in seqRanges if i[0] in remainingSequences] def readAnnots(opts, font, textDraw, sortedRanges, totalLength, fileNames): rangeDict = expandedSeqDict(dict(rangesPerSeq(sortedRanges))) annots = sorted(annotsFromFiles(opts, fileNames, rangeDict)) names = set(i[5] for i in annots) textSizes = dict(sizesPerText(names, font, textDraw)) maxTextLength = totalLength // 2 okLengths = [i[0] for i in textSizes.values() if i[0] <= maxTextLength] margin = max(okLengths) if okLengths else 0 return annots, textSizes, margin def lastDotplot(opts, args): logLevel = logging.INFO if opts.verbose else logging.WARNING logging.basicConfig(format="%(filename)s: %(message)s", level=logLevel) font = getFont(opts) image_mode = 'RGB' forward_color = ImageColor.getcolor(opts.forwardcolor, image_mode) reverse_color = ImageColor.getcolor(opts.reversecolor, image_mode) zipped_colors = zip(forward_color, reverse_color) overlap_color = tuple([(i + j) // 2 for i, j in zipped_colors]) maxGap1, maxGapB1 = twoValuesFromOption(opts.max_gap1, ":") maxGap2, maxGapB2 = twoValuesFromOption(opts.max_gap2, ":") logging.info("reading alignments...") alnData = readAlignments(args[0], opts) alignments, seqRanges1, coverDict1, seqRanges2, coverDict2 = alnData if not alignments: raise RuntimeError("there are no alignments") logging.info("cutting...") coverDict1 = dict(mergedRangesPerSeq(coverDict1)) coverDict2 = dict(mergedRangesPerSeq(coverDict2)) minAlignedBases = min(coveredLength(coverDict1), coveredLength(coverDict2)) pad = int(opts.pad * minAlignedBases) cutRanges1 = list(trimmed(seqRanges1, coverDict1, minAlignedBases, maxGap1, pad, pad)) cutRanges2 = list(trimmed(seqRanges2, coverDict2, minAlignedBases, maxGap2, pad, pad)) biggestSeqs1 = biggestSequences(cutRanges1, opts.maxseqs) cutRanges1 = [i for i in cutRanges1 if i[0] in biggestSeqs1] alignments = [i for i in alignments if i[0] in biggestSeqs1] cutRanges2 = remainingSequenceRanges(cutRanges2, alignments, 1) biggestSeqs2 = biggestSequences(cutRanges2, opts.maxseqs) cutRanges2 = [i for i in cutRanges2 if i[0] in biggestSeqs2] alignments = [i for i in alignments if i[1] in biggestSeqs2] cutRanges1 = remainingSequenceRanges(cutRanges1, alignments, 0) logging.info("reading secondary alignments...") alnDataB = readSecondaryAlignments(opts, cutRanges1, cutRanges2) alignmentsB, seqRangesB1, coverDictB1, seqRangesB2, coverDictB2 = alnDataB logging.info("cutting...") coverDictB1 = dict(mergedRangesPerSeq(coverDictB1)) coverDictB2 = dict(mergedRangesPerSeq(coverDictB2)) cutRangesB1 = trimmed(seqRangesB1, coverDictB1, minAlignedBases, maxGapB1, 0, 0) cutRangesB2 = trimmed(seqRangesB2, coverDictB2, minAlignedBases, maxGapB2, 0, 0) logging.info("sorting...") sortOut = allSortedRanges(opts, alignments, alignmentsB, cutRanges1, cutRangesB1, cutRanges2, cutRangesB2) sortedRanges1, sortedRanges2 = sortOut textDraw = None if opts.fontsize: textDraw = ImageDraw.Draw(Image.new(image_mode, (1, 1))) textRot1 = "vertical".startswith(opts.rot1) i1 = dataFromRanges(sortedRanges1, font, textDraw, opts.labels1, textRot1) rangeSizes1, labelData1, tMargin = i1 textRot2 = "horizontal".startswith(opts.rot2) i2 = dataFromRanges(sortedRanges2, font, textDraw, opts.labels2, textRot2) rangeSizes2, labelData2, lMargin = i2 logging.info("reading annotations...") annots1 = readAnnots(opts, font, textDraw, sortedRanges1, opts.height, opts.bed1) bMargin = annots1[-1] annots2 = readAnnots(opts, font, textDraw, sortedRanges2, opts.width, opts.bed2) rMargin = annots2[-1] maxPixels1 = opts.width - lMargin - rMargin maxPixels2 = opts.height - tMargin - bMargin bpPerPix1 = get_bp_per_pix(rangeSizes1, opts.border_pixels, maxPixels1) bpPerPix2 = get_bp_per_pix(rangeSizes2, opts.border_pixels, maxPixels2) bpPerPix = max(bpPerPix1, bpPerPix2) logging.info("bp per pixel = " + str(bpPerPix)) p1 = pixelData(rangeSizes1, bpPerPix, opts.border_pixels, lMargin) rangePixBegs1, rangePixLens1, rMarginBeg = p1 width = rMarginBeg + rMargin rangeDict1 = dict(rangesAndOriginsPerSeq(sortedRanges1, rangePixBegs1, rangePixLens1, bpPerPix)) p2 = pixelData(rangeSizes2, bpPerPix, opts.border_pixels, tMargin) rangePixBegs2, rangePixLens2, bMarginBeg = p2 height = bMarginBeg + bMargin rangeDict2 = dict(rangesAndOriginsPerSeq(sortedRanges2, rangePixBegs2, rangePixLens2, bpPerPix)) logging.info("width: " + str(width)) logging.info("height: " + str(height)) logging.info("processing alignments...") allAlignments = alignments + alignmentsB hits = alignmentPixels(width, height, allAlignments, bpPerPix, rangeDict1, rangeDict2) rangeDict1 = expandedSeqDict(rangeDict1) rangeDict2 = expandedSeqDict(rangeDict2) boxes1 = list(bedBoxes(annots1, rangeDict1, rMarginBeg, True, bpPerPix)) boxes2 = list(bedBoxes(annots2, rangeDict2, bMarginBeg, False, bpPerPix)) boxes = sorted(itertools.chain(boxes1, boxes2)) logging.info("drawing...") image_size = width, height im = Image.new(image_mode, image_size, opts.background_color) drawAnnotations(im, boxes, tMargin, bMarginBeg, lMargin, rMarginBeg) joinA, joinB = twoValuesFromOption(opts.join, ":") if joinA in "13": drawJoins(im, alignments, bpPerPix, 0, rangeDict1, rangeDict2) if joinB in "13": drawJoins(im, alignmentsB, bpPerPix, 0, rangeDict1, rangeDict2) if joinA in "23": drawJoins(im, alignments, bpPerPix, 1, rangeDict2, rangeDict1) if joinB in "23": drawJoins(im, alignmentsB, bpPerPix, 1, rangeDict2, rangeDict1) for i in range(height): for j in range(width): store_value = hits[i * width + j] xy = j, i if store_value == 1: im.putpixel(xy, forward_color) elif store_value == 2: im.putpixel(xy, reverse_color) elif store_value == 3: im.putpixel(xy, overlap_color) if opts.fontsize != 0: axis1 = axisImage(labelData1, rangePixBegs1, rangePixLens1, textRot1, False, font, image_mode, opts) if textRot1: axis1 = axis1.transpose(Image.ROTATE_90) axis2 = axisImage(labelData2, rangePixBegs2, rangePixLens2, textRot2, textRot2, font, image_mode, opts) if not textRot2: axis2 = axis2.transpose(Image.ROTATE_270) im.paste(axis1, (0, 0)) im.paste(axis2, (0, 0)) annoImage1 = annoTextImage(opts, image_mode, font, bMargin, width, boxes1, False) annoImage1 = annoImage1.transpose(Image.ROTATE_90) annoImage2 = annoTextImage(opts, image_mode, font, rMargin, height, boxes2, True) im.paste(annoImage1, (0, bMarginBeg)) im.paste(annoImage2, (rMarginBeg, 0)) for i in rangePixBegs1[1:]: box = i - opts.border_pixels, tMargin, i, bMarginBeg im.paste(opts.border_color, box) for i in rangePixBegs2[1:]: box = lMargin, i - opts.border_pixels, rMarginBeg, i im.paste(opts.border_color, box) im.save(args[1]) if __name__ == "__main__": usage = """%prog --help or: %prog [options] maf-or-tab-alignments dotplot.png or: %prog [options] maf-or-tab-alignments dotplot.gif or: ...""" description = "Draw a dotplot of pair-wise sequence alignments in MAF or tabular format." op = optparse.OptionParser(usage=usage, description=description) op.add_option("-v", "--verbose", action="count", help="show progress messages & data about the plot") # Replace "width" & "height" with a single "length" option? op.add_option("-x", "--width", metavar="INT", type="int", default=1000, help="maximum width in pixels (default: %default)") op.add_option("-y", "--height", metavar="INT", type="int", default=1000, help="maximum height in pixels (default: %default)") op.add_option("-m", "--maxseqs", type="int", default=100, metavar="M", help="maximum number of horizontal or vertical sequences " "(default=%default)") op.add_option("-1", "--seq1", metavar="PATTERN", action="append", default=[], help="which sequences to show from the 1st genome") op.add_option("-2", "--seq2", metavar="PATTERN", action="append", default=[], help="which sequences to show from the 2nd genome") op.add_option("--alignments", metavar="FILE", help="secondary alignments") op.add_option("--sort1", default="1", metavar="N", help="genome1 sequence order: 0=input order, 1=name order, " "2=length order, 3=alignment order (default=%default)") op.add_option("--sort2", default="1", metavar="N", help="genome2 sequence order: 0=input order, 1=name order, " "2=length order, 3=alignment order (default=%default)") op.add_option("--strands1", default="0", metavar="N", help= "genome1 sequence orientation: 0=forward orientation, " "1=alignment orientation (default=%default)") op.add_option("--strands2", default="0", metavar="N", help= "genome2 sequence orientation: 0=forward orientation, " "1=alignment orientation (default=%default)") op.add_option("--max-gap1", metavar="FRAC", default="0.5,2", help= "maximum unaligned (end,mid) gap in genome1: " "fraction of aligned length (default=%default)") op.add_option("--max-gap2", metavar="FRAC", default="0.5,2", help= "maximum unaligned (end,mid) gap in genome2: " "fraction of aligned length (default=%default)") op.add_option("--pad", metavar="FRAC", type="float", default=0.04, help= "pad length when cutting unaligned gaps: " "fraction of aligned length (default=%default)") op.add_option("-j", "--join", default="0", metavar="N", help= "join: 0=nothing, 1=alignments adjacent in genome1, " "2=alignments adjacent in genome2 (default=%default)") op.add_option("--border-pixels", metavar="INT", type="int", default=1, help="number of pixels between sequences (default=%default)") op.add_option("-a", "--bed1", "--rmsk1", "--genePred1", "--gap1", action="append", default=[], metavar="FILE", help="read genome1 annotations") op.add_option("-b", "--bed2", "--rmsk2", "--genePred2", "--gap2", action="append", default=[], metavar="FILE", help="read genome2 annotations") og = optparse.OptionGroup(op, "Text options") og.add_option("-f", "--fontfile", metavar="FILE", help="TrueType or OpenType font file") og.add_option("-s", "--fontsize", metavar="SIZE", type="int", default=14, help="TrueType or OpenType font size (default: %default)") og.add_option("--labels1", type="int", default=0, metavar="N", help= "genome1 labels: 0=name, 1=name:length, " "2=name:start:length, 3=name:start-end (default=%default)") og.add_option("--labels2", type="int", default=0, metavar="N", help= "genome2 labels: 0=name, 1=name:length, " "2=name:start:length, 3=name:start-end (default=%default)") og.add_option("--rot1", metavar="ROT", default="h", help="text rotation for the 1st genome (default=%default)") og.add_option("--rot2", metavar="ROT", default="v", help="text rotation for the 2nd genome (default=%default)") op.add_option_group(og) og = optparse.OptionGroup(op, "Color options") og.add_option("-c", "--forwardcolor", metavar="COLOR", default="red", help="color for forward alignments (default: %default)") og.add_option("-r", "--reversecolor", metavar="COLOR", default="blue", help="color for reverse alignments (default: %default)") og.add_option("--border-color", metavar="COLOR", default="black", help="color for pixels between sequences (default=%default)") # --break-color and/or --break-pixels for intra-sequence breaks? og.add_option("--margin-color", metavar="COLOR", default="#dcdcdc", help="margin color") og.add_option("--exon-color", metavar="COLOR", default="PaleGreen", help="color for exons (default=%default)") og.add_option("--cds-color", metavar="COLOR", default="LimeGreen", help="color for protein-coding regions (default=%default)") og.add_option("--bridged-color", metavar="COLOR", default="yellow", help="color for bridged gaps (default: %default)") og.add_option("--unbridged-color", metavar="COLOR", default="orange", help="color for unbridged gaps (default: %default)") op.add_option_group(og) (opts, args) = op.parse_args() if len(args) != 2: op.error("2 arguments needed") opts.background_color = "white" opts.label_space = 5 # minimum number of pixels between axis labels try: lastDotplot(opts, args) except KeyboardInterrupt: pass # avoid silly error message except RuntimeError as e: prog = os.path.basename(sys.argv[0]) sys.exit(prog + ": error: " + str(e))