4586cbe789d7ff340473682da13625025263131d larrym Mon Nov 15 10:52:07 2010 -0800 support codon numbering in genePred tracks diff --git src/hg/hgTracks/cds.c src/hg/hgTracks/cds.c index 11d4fc2..05b6088 100644 --- src/hg/hgTracks/cds.c +++ src/hg/hgTracks/cds.c @@ -46,58 +46,66 @@ /* Array of colors used in drawing codons/bases/differences: */ Color cdsColor[CDS_NUM_COLORS]; boolean cdsColorsMade = FALSE; Color getCdsColor(int index) { assert(index < CDS_NUM_COLORS); return cdsColor[index]; } static void drawScaledBoxSampleWithText(struct hvGfx *hvg, int chromStart, int chromEnd, double scale, int xOff, int y, int height, Color color, int score, MgFont *font, char *text, bool zoomed, - int winStart, int maxPixels, boolean isCoding) + int winStart, int maxPixels, boolean isCoding, boolean justifyString) /* Draw a box scaled from chromosome to window coordinates with a codon or set of 3 or less bases drawn in the box. */ { /*first draw the box itself*/ drawScaledBoxSample(hvg, chromStart, chromEnd, scale, xOff, y, height, color, score); /*draw text in box if space, and align properly for codons or DNA*/ if (zoomed) { int i; int x1, x2, w; x1 = round((double)(chromStart-winStart)*scale) + xOff; x2 = round((double)(chromEnd-winStart)*scale) + xOff; if (x2 >= maxPixels) x2 = maxPixels - 1; w = x2-x1; if (w < 1) w = 1; if (chromEnd - chromStart == 3 && isCoding) - spreadBasesString(hvg,x1,y,w,height,whiteIndex(), - font,text,strlen(text), TRUE); + { + if(justifyString) + spreadBasesString(hvg, x1, y, w, height, whiteIndex(), font, text, strlen(text), TRUE); + else + hvGfxTextCentered(hvg, x1, y, w, height, whiteIndex(), font, text); + } else if (chromEnd - chromStart < 3 && isCoding) - spreadBasesString(hvg,x1,y,w,height,cdsColor[CDS_PARTIAL_CODON],font, - text,strlen(text), TRUE); + { + if(justifyString) + spreadBasesString(hvg, x1, y, w, height, cdsColor[CDS_PARTIAL_CODON], font, text, strlen(text), TRUE); + else + hvGfxTextCentered(hvg, x1, y, w, height, cdsColor[CDS_PARTIAL_CODON], font, text); + } else { int thisX,thisX2; char c[2]; c[1] = '\0'; int iMin = max(0, (winStart-chromStart)); int iMax = min((chromEnd-chromStart), (winEnd-chromStart)); for (i=iMin; i<iMax; i++) { if (text[i] == ' ') continue; c[0] = text[i]; thisX = round((double)(chromStart+i-winStart)*scale) + xOff; thisX2 = round((double)(chromStart+1+i-winStart)*scale) + xOff; hvGfxTextCentered(hvg, thisX, y, thisX2-thisX, height, @@ -644,31 +652,31 @@ /* Previous code didn't use exon frames for baseColorDrawGenomicCodons. * This meant simply counting off aligned bases to define frames. It * didn't work very well for TransMap alignments and not clear that its * the right thing to do for any alignment. By using exonFrames for * genomic codons, this is letting the query sequence define the frame. */ struct genbankCds cds; getPslCds(psl, tg, &cds); int insertMergeSize = -1; unsigned opts = genePredCdsStatFld|genePredExonFramesFld; struct genePred *gp = genePredFromPsl2(psl, opts, &cds, insertMergeSize); lf->start = gp->txStart; lf->end = gp->txEnd; lf->tallStart = gp->cdsStart; lf->tallEnd = gp->cdsEnd; - sfList = baseColorCodonsFromGenePred(lf, gp, colorStopStart); + sfList = baseColorCodonsFromGenePred(lf, gp, colorStopStart, FALSE); genePredFree(&gp); } return(sfList); } struct simpleFeature *baseColorCodonsFromPsl(struct linkedFeatures *lf, struct psl *psl, int sizeMul, boolean isXeno, int maxShade, enum baseColorDrawOpt drawOpt, struct track *tg) /* Given an lf and the psl from which the lf was constructed, * return a list of simpleFeature elements, one per codon (or partial * codon if the codon falls on a gap boundary. sizeMul, isXeno and maxShade * are for defaulting to one-simpleFeature-per-exon if cds is not found. */ { @@ -999,31 +1007,31 @@ { sf->start = winEnd - sf->start + winStart - 3; sf->end = sf->start + 3; } // Base offsets mod 6 for alternating colors: 0,1,2 --> first codon, 3,4,5 --> second codon. bool codonFirstColor = (sf->start % 6 < 3); sf->grayIx = codonToGrayIx(codon, codonFirstColor, NULL, FALSE, TRUE); zeroBytes(codon, 4); slAddHead(&sfList, sf); } slReverse(&sfList); return sfList; } struct simpleFeature *baseColorCodonsFromGenePred(struct linkedFeatures *lf, - struct genePred *gp, boolean colorStopStart) + struct genePred *gp, boolean colorStopStart, boolean codonNumbering) /* Given an lf and the genePred from which the lf was constructed, * return a list of simpleFeature elements, one per codon (or partial * codon if the codon falls on a gap boundary. */ { unsigned *starts = gp->exonStarts; unsigned *ends = gp->exonEnds; int blockCount = gp->exonCount; unsigned cdsStart = gp->cdsStart; unsigned cdsEnd = gp->cdsEnd; int *exonFrames = gp->exonFrames; boolean useExonFrames = (gp->optFields >= genePredExonFramesFld); int frame = 0; int currentStart = 0, currentEnd = 0; char partialCodonSeq[4]; int currentSize; @@ -1037,30 +1045,33 @@ partialCodonSeq[0] = '\0'; if (lf->orientation > 0) //positive strand { i0 = 0; iN = blockCount; iInc = 1; posStrand = TRUE; } else { i0 = blockCount-1; iN=-1; iInc = -1; posStrand = FALSE; } bool altColor = FALSE; unsigned cds5Prime = posStrand ? cdsStart : cdsEnd; + int width = winEnd - winStart; + // width cutoff really should be based on (a) how many codons this gene has, (2) the current font and (3) image width. + int codonIndex = !codonNumbering || width > 60 ? 0 : 1; for (i=i0; (iInc*i)<(iInc*iN); i=i+iInc) { int exonStart = starts[i]; int exonEnd = ends[i]; if (useExonFrames) { if(exonFrames[i] > 0) frame = 3 - exonFrames[i]; else frame = 0; } if(frame == 0) strcpy(partialCodonSeq,""); @@ -1163,30 +1174,31 @@ safef(tempCodonSeq, sizeof(tempCodonSeq), "%s%s", partialCodonSeq, theRestOfCodon); else safef(tempCodonSeq, sizeof(tempCodonSeq), "%s%s", theRestOfCodon, partialCodonSeq ); tempCodonSeq[4] = '\0'; // no more than 3 bases AllocVar(sf); sf->start = currentStart; sf->end = currentEnd; sf->grayIx = ((posStrand && currentEnd <= cdsEnd) || (!posStrand && currentStart >= cdsStart)) ? codonToGrayIx(tempCodonSeq, altColor, &foundStart, !posStrand, colorStopStart) : GRAYIX_CDS_ERROR; + sf->codonIndex = codonIndex; slAddHead(&sfList, sf); } break; } // end if we've gone off the end of the current exon currentSize = currentEnd - currentStart; /*inside a coding block (with 3 bases)*/ if (currentSize == 3) { AllocVar(sf); sf->start = currentStart; sf->end = currentEnd; if ((posStrand && currentEnd <= cdsEnd) || (!posStrand && currentStart >= cdsStart)) { @@ -1209,30 +1221,32 @@ sf->end = currentEnd; if (strlen(partialCodonSeq) == 3) sf->grayIx = codonToGrayIx(partialCodonSeq, altColor, &foundStart, !posStrand, colorStopStart); else sf->grayIx = GRAYIX_CDS_ERROR; strcpy(partialCodonSeq,"" ); /*update frame based on bases appended*/ frame -= currentSize; } else errAbort("%s: Too much dna (%d - %d = %d)<br>\n", lf->name, currentEnd, currentStart, currentSize); + if(codonIndex) + sf->codonIndex = codonIndex++; slAddHead(&sfList, sf); if(posStrand) currentStart = currentEnd; else currentEnd = currentStart; } // end loop on codons within exon /* coding + UTR exon */ if (posStrand && (exonEnd < ends[i]) && exonEnd < winEnd && ends[i] > winStart) { AllocVar(sf); sf->start = exonEnd; sf->end = ends[i]; slAddHead(&sfList, sf); @@ -1344,88 +1358,88 @@ mrnaBases[0] = '\0'; if (psl && isCoding) getMrnaBases(psl, mrnaSeq, mrnaS, s, e, (lf->orientation == -1), mrnaBases, &queryInsertion); if (queryInsertion && isCoding) color = cdsColor[CDS_QUERY_INSERTION]; dyStringAppendN(dyMrnaSeq, (char*)&mrnaSeq->dna[mrnaS], e-s); if (drawOpt == baseColorDrawItemBases) { if (cartUsualBooleanDb(cart, database, COMPLEMENT_BASES_VAR, FALSE)) complement(dyMrnaSeq->string, dyMrnaSeq->stringSize); drawScaledBoxSampleWithText(hvg, s, e, scale, xOff, y, heightPer, color, lf->score, font, dyMrnaSeq->string, - zoomedToBaseLevel, winStart, maxPixels, isCoding); + zoomedToBaseLevel, winStart, maxPixels, isCoding, TRUE); } else if (drawOpt == baseColorDrawItemCodons) { if (e <= lf->tallEnd) { boolean startColor = FALSE; /* re-set color of this block based on mrna codons rather than * genomic, but keep the odd/even cycle of dark/light shades. */ int mrnaGrayIx = codonToGrayIx(mrnaBases, (grayIx > 26), NULL, FALSE, TRUE); if (color == cdsColor[CDS_START]) startColor = TRUE; color = colorAndCodonFromGrayIx(hvg, mrnaCodon, mrnaGrayIx, ixColor); if (startColor && sameString(mrnaCodon,"M")) color = cdsColor[CDS_START]; drawScaledBoxSampleWithText(hvg, s, e, scale, xOff, y, heightPer, color, lf->score, font, mrnaCodon, zoomedToCodonLevel, winStart, - maxPixels, isCoding); + maxPixels, isCoding, TRUE); } else drawScaledBox(hvg, s, e, scale, xOff, y, heightPer, color); } else if (drawOpt == baseColorDrawDiffBases) { char *diffStr = NULL; char *genoDna = getCachedDna(s, e); diffStr = needMem(sizeof(char) * (e - s + 1)); maskDiffString(diffStr, dyMrnaSeq->string, genoDna, ' '); if (cartUsualBooleanDb(cart, database, COMPLEMENT_BASES_VAR, FALSE)) complement(diffStr, strlen(diffStr)); drawScaledBoxSampleWithText(hvg, s, e, scale, xOff, y, heightPer, color, lf->score, font, diffStr, - zoomedToBaseLevel, winStart, maxPixels, isCoding); + zoomedToBaseLevel, winStart, maxPixels, isCoding, TRUE); freeMem(diffStr); } else if (drawOpt == baseColorDrawDiffCodons) { if (e <= lf->tallEnd) { /* Color codons red wherever mrna differs from genomic; * keep the odd/even cycle of dark/light shades. */ colorAndCodonFromGrayIx(hvg, genomicCodon, grayIx, ixColor); int mrnaGrayIx = mrnaCodonToGrayIx(mrnaBases, genomicCodon[0], (grayIx > 26)); color = colorAndCodonFromGrayIx(hvg, mrnaCodon, mrnaGrayIx, ixColor); // Look up mrnaCodon again because if mrnaGrayIx is GRAYIX_SYN_PROT, // codon value is lost: safef(mrnaCodon, sizeof(mrnaCodon), "%c", baseColorLookupCodon(mrnaBases)); if (mrnaCodon[0] != genomicCodon[0]) { drawScaledBoxSampleWithText(hvg, s, e, scale, xOff, y, heightPer, color, lf->score, font, mrnaCodon, zoomedToCodonLevel, - winStart, maxPixels, isCoding); + winStart, maxPixels, isCoding, TRUE); } else drawScaledBox(hvg, s, e, scale, xOff, y, heightPer, color); } else drawScaledBox(hvg, s, e, scale, xOff, y, heightPer, color); } else if (drawOpt != baseColorDrawCds) errAbort("Unknown drawOpt: %d<br>\n", drawOpt); dyStringFree(&dyMrnaSeq); } else { /*show we have an error by coloring entire exon block yellow*/ @@ -1433,44 +1447,46 @@ // FIXME: this shouldn't ever happen, should be an errAbort warn("Bug: drawDiffTextBox: convertCoordUsingPsl failed<br>\n"); } } void baseColorDrawItem(struct track *tg, struct linkedFeatures *lf, int grayIx, struct hvGfx *hvg, int xOff, int y, double scale, MgFont *font, int s, int e, int heightPer, boolean zoomedToCodonLevel, struct dnaSeq *mrnaSeq, struct simpleFeature *sf, struct psl *psl, enum baseColorDrawOpt drawOpt, int maxPixels, int winStart, Color originalColor) /* Draw codon/base-colored item. */ { -char codon[2] = " "; +char codon[64] = " "; Color color = colorAndCodonFromGrayIx(hvg, codon, grayIx, originalColor); +if(sf->codonIndex) + safef(codon, sizeof(codon), "%c %d", codon[0], sf->codonIndex); /* When we are zoomed out far enough so that multiple bases/codons share the * same pixel, we have to draw differences in a separate pass (baseColorOverdrawDiff) * so don't waste time drawing the differences here: */ boolean zoomedOutToPostProcessing = ((drawOpt == baseColorDrawDiffBases && !zoomedToBaseLevel) || (drawOpt == baseColorDrawDiffCodons && !zoomedToCdsColorLevel)); if (drawOpt == baseColorDrawGenomicCodons && (e-s <= 3)) { drawScaledBoxSampleWithText(hvg, s, e, scale, xOff, y, heightPer, color, lf->score, font, codon, - zoomedToCodonLevel, winStart, maxPixels, TRUE); + zoomedToCodonLevel, winStart, maxPixels, TRUE, !sf->codonIndex); } else if (mrnaSeq != NULL && (psl != NULL || sf != NULL) && !zoomedOutToPostProcessing && drawOpt != baseColorDrawGenomicCodons && drawOpt != baseColorDrawOff) { drawDiffTextBox(hvg, xOff, y, scale, heightPer, font, color, chromName, s, e, sf, psl, mrnaSeq, lf, grayIx, drawOpt, maxPixels, tg->colorShades, originalColor); } else { /* revert to normal coloring */ drawScaledBoxSample(hvg, s, e, scale, xOff, y, heightPer, color, lf->score ); } @@ -1806,31 +1822,31 @@ struct simpleFeature *sf; if (!cdsColorsMade) { makeCdsShades(hvg, cdsColor); cdsColorsMade = TRUE; } for (sf = sfList; sf != NULL; sf = sf->next) { char codon[4]; Color color = colorAndCodonFromGrayIx(hvg, codon, sf->grayIx, MG_GRAY); if (zoomedToText) drawScaledBoxSampleWithText(hvg, sf->start, sf->end, scale, insideX, y, height, color, 1.0, font, codon, TRUE, - winStart, maxPixels, TRUE); + winStart, maxPixels, TRUE, TRUE); else /* zoomed in just enough to see colored boxes */ drawScaledBox(hvg, sf->start, sf->end, scale, xOff, y, height, color); } } void baseColorDrawCleanup(struct linkedFeatures *lf, struct dnaSeq **pMrnaSeq, struct psl **pPsl) /* Free structures allocated just for base/cds coloring. */ { // We could free lf->original here (either genePredFree or pslFree, depending // on the type -- but save time by skipping that. Maybe we should save time // by skipping this free too: if (pMrnaSeq != NULL)