void writeGap(struct dyString *aRes, int aGap, char *aSeq, struct dyString *bRes, int bGap, char *bSeq)
/* Write double - gap. Something like:
* ....123.... or --c
* ...4123.... ag- */
{
char abbrev[16];
int minToAbbreviate = 16;
if (doShort && (aGap >= minToAbbreviate || bGap >= minToAbbreviate))
{
fillShortGapString(abbrev, aGap, '.', 13);
dyStringAppend(aRes, abbrev);
fillShortGapString(abbrev, bGap, '.', 13);
dyStringAppend(bRes, abbrev);
}
else
{
#ifdef OLD
dyStringAppendMultiC(aRes, '-', aGap);
dyStringAppendN(bRes, bSeq, aGap);
dyStringAppendN(aRes, aSeq, bGap);
dyStringAppendMultiC(bRes, '-', bGap);
#endif /* OLD */
dyStringAppendMultiC(aRes, '-', bGap);
dyStringAppendN(bRes, bSeq, bGap);
dyStringAppendN(aRes, aSeq, aGap);
dyStringAppendMultiC(bRes, '-', aGap);
}
}
void writeGap(struct dyString *aRes, int aGap, char *aSeq, struct dyString *bRes, int bGap, char *bSeq)
/* Write double - gap. Something like:
* --c
* ag- */
{
dyStringAppendMultiC(aRes, '-', bGap);
dyStringAppendN(bRes, bSeq, bGap);
dyStringAppendN(aRes, aSeq, aGap);
dyStringAppendMultiC(bRes, '-', aGap);
}
static void fillInMissing(struct oneOrg *nativeOrg, struct oneOrg *orgList,
struct dnaSeq *native, int seqStart, int curPos, int aliStart)
/* Fill in alignment strings in orgList with native sequence
* for first organism, and dots for rest. */
{
int fillSize = aliStart - curPos;
int offset = curPos - seqStart;
struct oneOrg *org;
if (nativeOrg == NULL)
return;
dyStringAppendN(nativeOrg->dy, native->dna + offset, fillSize);
for (org = orgList; org != NULL; org = org->next)
{
if (org != nativeOrg)
dyStringAppendMultiC(org->dy, '.', fillSize);
}
}
void writeInsert(struct dyString *aRes, struct dyString *bRes, char *aSeq, int gapSize)
/* Write out gap, possibly shortened, to aRes, bRes. */
{
int minToAbbreviate = 16;
if (doShort && gapSize >= minToAbbreviate)
{
char abbrevGap[16];
char abbrevSeq[16];
fillSpliceSites(abbrevSeq, gapSize, aSeq, 15);
dyStringAppend(aRes, abbrevSeq);
fillShortGapString(abbrevGap, gapSize, '-', 15);
dyStringAppend(bRes, abbrevGap);
}
else
{
dyStringAppendN(aRes, aSeq, gapSize);
dyStringAppendMultiC(bRes, '-', gapSize);
}
}
static void saveAxtBundle(char *chromName, int chromSize, int chromOffset,
struct ffAli *ali,
struct dnaSeq *tSeq, struct hash *t3Hash, struct dnaSeq *qSeq,
boolean qIsRc, boolean tIsRc,
enum ffStringency stringency, int minMatch, struct gfOutput *out)
/* Save alignment to axtBundle. */
{
struct axtData *ad = out->data;
struct ffAli *sAli, *eAli, *ff, *rt, *eFf = NULL;
struct axt *axt;
struct dyString *q = newDyString(1024), *t = newDyString(1024);
struct axtBundle *gab;
struct trans3 *t3List = NULL;
if (t3Hash != NULL)
t3List = hashMustFindVal(t3Hash, tSeq->name);
AllocVar(gab);
gab->tSize = chromSize;
gab->qSize = qSeq->size;
for (sAli = ali; sAli != NULL; sAli = eAli)
{
eAli = ffNextBreak(sAli, 8, tSeq, t3List);
dyStringClear(q);
dyStringClear(t);
for (ff = sAli; ff != eAli; ff = ff->right)
{
dyStringAppendN(q, ff->nStart, ff->nEnd - ff->nStart);
dyStringAppendN(t, ff->hStart, ff->hEnd - ff->hStart);
rt = ff->right;
if (rt != eAli)
{
int nGap = rt->nStart - ff->nEnd;
int nhStart = trans3GenoPos(rt->hStart, tSeq, t3List, FALSE)
+ chromOffset;
int ohEnd = trans3GenoPos(ff->hEnd, tSeq, t3List, TRUE)
+ chromOffset;
int hGap = nhStart - ohEnd;
int gap = max(nGap, hGap);
if (nGap < 0 || hGap < 0)
{
errAbort("Negative gap size in %s vs %s", tSeq->name, qSeq->name);
}
if (nGap == gap)
{
dyStringAppendN(q, ff->nEnd, gap);
dyStringAppendMultiC(t, '-', gap);
}
else
{
dyStringAppendN(t, ff->hEnd, gap);
dyStringAppendMultiC(q, '-', gap);
}
}
eFf = ff; /* Keep track of last block in bunch */
}
assert(t->stringSize == q->stringSize);
AllocVar(axt);
axt->qName = cloneString(qSeq->name);
axt->qStart = sAli->nStart - qSeq->dna;
axt->qEnd = eFf->nEnd - qSeq->dna;
axt->qStrand = (qIsRc ? '-' : '+');
axt->tName = cloneString(chromName);
axt->tStart = trans3GenoPos(sAli->hStart, tSeq, t3List, FALSE) + chromOffset;
axt->tEnd = trans3GenoPos(eFf->hEnd, tSeq, t3List, TRUE) + chromOffset;
axt->tStrand = (tIsRc ? '-' : '+');
axt->symCount = t->stringSize;
axt->qSym = cloneString(q->string);
axt->tSym = cloneString(t->string);
axt->frame = trans3Frame(sAli->hStart, t3List);
if (out->qIsProt)
axt->score = axtScoreProteinDefault(axt);
else
axt->score = axtScoreDnaDefault(axt);
slAddHead(&gab->axtList, axt);
}
slReverse(&gab->axtList);
dyStringFree(&q);
dyStringFree(&t);
slAddHead(&ad->bundleList, gab);
}
struct mafAli *hgMafFrag(
char *database, /* Database, must already have hSetDb to this */
char *track, /* Name of MAF track */
char *chrom, /* Chromosome (in database genome) */
int start, int end, /* start/end in chromosome */
char strand, /* Chromosome strand. */
char *outName, /* Optional name to use in first component */
struct slName *orderList /* Optional order of organisms. */
)
/* mafFrag- Extract maf sequences for a region from database.
* This creates a somewhat unusual MAF that extends from start
* to end whether or not there are actually alignments. Where
* there are no alignments (or alignments missing a species)
* a . character fills in. The score is always zero, and
* the sources just indicate the species. You can mafFree this
* as normal. */
{
int chromSize = hChromSize(database, chrom);
struct sqlConnection *conn = hAllocConn(database);
struct dnaSeq *native = hChromSeq(database, chrom, start, end);
struct mafAli *maf, *mafList = mafLoadInRegion(conn, track, chrom, start, end);
char masterSrc[128];
struct hash *orgHash = newHash(10);
struct oneOrg *orgList = NULL, *org, *nativeOrg = NULL;
int curPos = start, symCount = 0;
struct slName *name;
int order = 0;
/* Check that the mafs are really copacetic, the particular
* subtype we think is in the database that this (relatively)
* simple code can handle. */
safef(masterSrc, sizeof(masterSrc), "%s.%s", database, chrom);
mafCheckFirstComponentSrc(mafList, masterSrc);
mafCheckFirstComponentStrand(mafList, '+');
slSort(&mafList, mafCmp);
/* Prebuild organisms if possible from input orderList. */
for (name = orderList; name != NULL; name = name->next)
{
AllocVar(org);
slAddHead(&orgList, org);
hashAddSaveName(orgHash, name->name, org, &org->name);
org->dy = dyStringNew(native->size*1.5);
org->order = order++;
if (nativeOrg == NULL)
nativeOrg = org;
}
if (orderList == NULL)
{
AllocVar(org);
slAddHead(&orgList, org);
hashAddSaveName(orgHash, database, org, &org->name);
org->dy = dyStringNew(native->size*1.5);
if (nativeOrg == NULL)
nativeOrg = org;
}
/* Go through all mafs in window, mostly building up
* org->dy strings. */
for (maf = mafList; maf != NULL; maf = maf->next)
{
struct mafComp *mc, *mcMaster = maf->components;
struct mafAli *subMaf = NULL;
order = 0;
if (curPos < mcMaster->start)
{
fillInMissing(nativeOrg, orgList, native, start,
curPos, mcMaster->start);
symCount += mcMaster->start - curPos;
}
if (curPos < mcMaster->start + mcMaster->size) /* Prevent worst
* backtracking */
{
if (mafNeedSubset(maf, masterSrc, curPos, end))
{
subMaf = mafSubset(maf, masterSrc, curPos, end);
if (subMaf == NULL)
continue;
}
else
subMaf = maf;
for (mc = subMaf->components; mc != NULL; mc = mc->next, ++order)
{
/* Extract name up to dot into 'orgName' */
char buf[128], *e, *orgName;
if ((mc->size == 0) || (mc->srcSize == 0)) /* skip over components without sequence */
continue;
mc->leftStatus = mc->rightStatus = 0; /* squash annotation */
e = strchr(mc->src, '.');
if (e == NULL)
orgName = mc->src;
else
{
int len = e - mc->src;
if (len >= sizeof(buf))
errAbort("organism/database name %s too long", mc->src);
memcpy(buf, mc->src, len);
buf[len] = 0;
orgName = buf;
}
/* Look up dyString corresponding to org, and create a
* new one if necessary. */
org = hashFindVal(orgHash, orgName);
if (org == NULL)
{
if (orderList != NULL)
errAbort("%s is not in orderList", orgName);
AllocVar(org);
slAddHead(&orgList, org);
hashAddSaveName(orgHash, orgName, org, &org->name);
org->dy = dyStringNew(native->size*1.5);
dyStringAppendMultiC(org->dy, '.', symCount);
if (nativeOrg == NULL)
nativeOrg = org;
}
if (orderList == NULL && order > org->order)
org->order = order;
org->hit = TRUE;
/* Fill it up with alignment. */
dyStringAppendN(org->dy, mc->text, subMaf->textSize);
}
for (org = orgList; org != NULL; org = org->next)
{
if (!org->hit)
dyStringAppendMultiC(org->dy, '.', subMaf->textSize);
org->hit = FALSE;
}
symCount += subMaf->textSize;
curPos = mcMaster->start + mcMaster->size;
if (subMaf != maf)
mafAliFree(&subMaf);
}
}
if (curPos < end)
{
fillInMissing(nativeOrg, orgList, native, start, curPos, end);
symCount += end - curPos;
}
mafAliFreeList(&mafList);
slSort(&orgList, oneOrgCmp);
if (strand == '-')
{
for (org = orgList; org != NULL; org = org->next)
reverseComplement(org->dy->string, org->dy->stringSize);
}
/* Construct our maf */
AllocVar(maf);
maf->textSize = symCount;
for (org = orgList; org != NULL; org = org->next)
{
struct mafComp *mc;
AllocVar(mc);
if (org == orgList)
{
if (outName != NULL)
{
mc->src = cloneString(outName);
mc->srcSize = native->size;
mc->strand = '+';
mc->start = 0;
mc->size = native->size;
}
else
{
mc->src = cloneString(masterSrc);
mc->srcSize = chromSize;
mc->strand = strand;
if (strand == '-')
reverseIntRange(&start, &end, chromSize);
mc->start = start;
mc->size = end-start;
}
}
else
{
int size = countAlpha(org->dy->string);
mc->src = cloneString(org->name);
mc->srcSize = size;
mc->strand = '+';
mc->start = 0;
mc->size = size;
}
mc->text = cloneString(org->dy->string);
dyStringFree(&org->dy);
slAddHead(&maf->components, mc);
}
slReverse(&maf->components);
slFreeList(&orgList);
freeHash(&orgHash);
hFreeConn(&conn);
return maf;
}
void writeInsert(struct dyString *aRes, struct dyString *bRes, char *aSeq, int gapSize)
/* Write out gap, possibly shortened, to aRes, bRes. */
{
dyStringAppendN(aRes, aSeq, gapSize);
dyStringAppendMultiC(bRes, '-', gapSize);
}
void outputBlocks(struct lineFile *lf,
struct block *blockList, int score, FILE *f, boolean isRc,
char *qName, int qSize, char *qNibDir, struct dlList *qCache,
char *tName, int tSize, char *tNibDir, struct dlList *tCache,
boolean rescore)
/* Output block list as an axt to file f. */
{
int qStart = BIGNUM, qEnd = 0, tStart = BIGNUM, tEnd = 0;
struct block *lastBlock = NULL;
struct block *block;
struct dyString *qSym = newDyString(16*1024);
struct dyString *tSym = newDyString(16*1024);
struct dnaSeq *qSeq = NULL, *tSeq = NULL, *seq = NULL;
struct axt axt;
boolean qIsTwoBit = twoBitIsFile(qNibDir);
boolean tIsTwoBit = twoBitIsFile(tNibDir);
if (blockList == NULL)
return;
/* Figure overall dimensions. */
for (block = blockList; block != NULL; block = block->next)
{
if (qStart > block->qStart) qStart = block->qStart;
if (qEnd < block->qEnd) qEnd = block->qEnd;
if (tStart > block->tStart) tStart = block->tStart;
if (tEnd < block->tEnd) tEnd = block->tEnd;
}
/* Load sequence covering alignment from nib files. */
if (isRc)
{
reverseIntRange(&qStart, &qEnd, qSize);
if (qIsFa)
{
for (seq = qFaList ; seq != NULL ; seq = seq->next)
if (sameString(qName, seq->name))
break;
if (seq != NULL)
{
AllocVar(qSeq);
qSeq->size = qEnd - qStart;
qSeq->name = cloneString(qName);
qSeq->dna = cloneMem((seq->dna)+qStart, qSeq->size);
}
else
errAbort("sequence not found %s\n",qName);
}
else
qSeq = readFromCache(qCache, qNibDir, qName, qStart, qEnd - qStart, qSize, qIsTwoBit);
reverseIntRange(&qStart, &qEnd, qSize);
reverseComplement(qSeq->dna, qSeq->size);
}
else
{
if (qIsFa)
{
for (seq = qFaList ; seq != NULL ; seq = seq->next)
{
if (sameString(qName, seq->name))
break;
}
if (seq != NULL)
{
AllocVar(qSeq);
qSeq->size = qEnd - qStart;
qSeq->name = cloneString(qName);
qSeq->dna = (seq->dna)+qStart;
}
else
errAbort("sequence not found %s\n",qName);
}
else
qSeq = readFromCache(qCache, qNibDir, qName, qStart, qEnd - qStart, qSize, qIsTwoBit);
}
if (tIsFa)
{
for (seq = tFaList ; seq != NULL ; seq = seq->next)
if (sameString(tName, seq->name))
break;
if (seq != NULL)
{
AllocVar(tSeq);
tSeq->size = tEnd - tStart;
tSeq->name = cloneString(tName);
tSeq->dna = cloneMem((seq->dna)+tStart, tSeq->size);
}
else
errAbort("sequence not found %s\n",tName);
}
else
tSeq = readFromCache(tCache, tNibDir, tName, tStart, tEnd - tStart, tSize, tIsTwoBit);
/* Loop through blocks copying sequence into dynamic strings. */
for (block = blockList; block != NULL; block = block->next)
{
if (lastBlock != NULL)
{
int qGap = block->qStart - lastBlock->qEnd;
int tGap = block->tStart - lastBlock->tEnd;
if (qGap != 0 && tGap != 0)
{
errAbort("Gaps in both strand on alignment ending line %d of %s",
lf->lineIx, lf->fileName);
}
if (qGap > 0)
{
dyStringAppendMultiC(tSym, '-', qGap);
dyStringAppendN(qSym, qSeq->dna + lastBlock->qEnd - qStart, qGap);
}
if (tGap > 0)
{
dyStringAppendMultiC(qSym, '-', tGap);
dyStringAppendN(tSym, tSeq->dna + lastBlock->tEnd - tStart, tGap);
}
}
if (qSeq->size < block->qStart - qStart)
{
errAbort("read past end of sequence %s size =%d block->qStart-qstart=%d block->qStart=%d qEnd=%d \n", qName, qSeq->size, block->qStart-qStart,block->qStart, block->qEnd );
}
dyStringAppendN(qSym, qSeq->dna + block->qStart - qStart,
block->qEnd - block->qStart);
if (tSeq->size < block->tStart - tStart)
{
errAbort("read past end of sequence %s size =%d block->tStart-tstart=%d\n", tName, tSeq->size, block->tStart-tStart);
}
dyStringAppendN(tSym, tSeq->dna + block->tStart - tStart,
block->tEnd - block->tStart);
lastBlock = block;
}
if (qSym->stringSize != tSym->stringSize)
errAbort("qSize and tSize don't agree in alignment ending line %d of %s",
lf->lineIx, lf->fileName);
if (rescore)
score = axtScoreSym(scoreScheme, qSym->stringSize,
qSym->string, tSym->string);
/* Fill in an axt and write it to output. */
ZeroVar(&axt);
axt.qName = qName;
axt.qStart = qStart;
axt.qEnd = qEnd;
axt.qStrand = (isRc ? '-' : '+');
axt.tName = tName;
axt.tStart = tStart;
axt.tEnd = tEnd;
axt.tStrand = '+';
axt.score = score;
axt.symCount = qSym->stringSize;
axt.qSym = qSym->string;
axt.tSym = tSym->string;
axtWrite(&axt, f);
/* Clean up. */
if (!qIsFa)
freeDnaSeq(&qSeq);
freeDnaSeq(&tSeq);
dyStringFree(&qSym);
dyStringFree(&tSym);
}
