struct ssBundle *ffSeedExtInMem(struct genoFind *gf, struct dnaSeq *qSeq, Bits *qMaskBits,
int qOffset, struct lm *lm, int minScore, boolean isRc)
/* Do seed and extend type alignment */
{
struct ssBundle *bunList = NULL, *bun;
int hitCount;
struct gfClump *clumpList, *clump;
struct gfRange *rangeList = NULL, *range;
struct dnaSeq *tSeq;
clumpList = gfFindClumpsWithQmask(gf, qSeq, qMaskBits, qOffset, lm, &hitCount);
for (clump = clumpList; clump != NULL; clump = clump->next)
clumpToExactRange(clump, qSeq, gf->tileSize, 0, NULL, &rangeList);
slSort(&rangeList, gfRangeCmpTarget);
rangeList = gfRangesBundle(rangeList, ffIntronMax);
for (range = rangeList; range != NULL; range = range->next)
{
range->qStart += qOffset;
range->qEnd += qOffset;
tSeq = range->tSeq;
AllocVar(bun);
bun->qSeq = qSeq;
bun->genoSeq = tSeq;
bun->ffList = gfRangesToFfItem(range->components, qSeq);
bun->isProt = FALSE;
bun->avoidFuzzyFindKludge = TRUE;
ssStitch(bun, ffCdna, 16, 10);
refineBundle(gf, qSeq, qMaskBits, qOffset, tSeq, lm, bun, isRc);
slAddHead(&bunList, bun);
}
gfRangeFreeList(&rangeList);
gfClumpFreeList(&clumpList);
return bunList;
}
void gfLongDnaInMem(struct dnaSeq *query, struct genoFind *gf,
boolean isRc, int minScore, Bits *qMaskBits,
struct gfOutput *out, boolean fastMap, boolean band)
/* Chop up query into pieces, align each, and stitch back
* together again. */
{
int hitCount;
int maxSize = MAXSINGLEPIECESIZE;
int preferredSize = 4500;
int overlapSize = 250;
struct dnaSeq subQuery = *query;
struct lm *lm = lmInit(0);
int subOffset, subSize, nextOffset;
DNA saveEnd, *endPos;
struct ssBundle *oneBunList = NULL, *bigBunList = NULL, *bun;
struct hash *bunHash = newHash(8);
for (subOffset = 0; subOffset<query->size; subOffset = nextOffset)
{
struct gfClump *clumpList;
struct gfRange *rangeList = NULL;
/* Figure out size of this piece. If query is
* maxSize or less do it all. Otherwise just
* do prefered size, and set it up to overlap
* with surrounding pieces by overlapSize. */
if (subOffset == 0 && query->size <= maxSize)
nextOffset = subSize = query->size;
else
{
subSize = preferredSize;
if (subSize + subOffset >= query->size)
{
subSize = query->size - subOffset;
nextOffset = query->size;
}
else
{
nextOffset = subOffset + preferredSize - overlapSize;
}
}
subQuery.dna = query->dna + subOffset;
subQuery.size = subSize;
endPos = &subQuery.dna[subSize];
saveEnd = *endPos;
*endPos = 0;
if (band)
{
oneBunList = ffSeedExtInMem(gf, &subQuery, qMaskBits, subOffset, lm, minScore, isRc);
}
else
{
clumpList = gfFindClumpsWithQmask(gf, &subQuery, qMaskBits, subOffset, lm, &hitCount);
if (fastMap)
{
oneBunList = fastMapClumpsToBundles(gf, clumpList, &subQuery);
}
else
{
oneBunList = gfClumpsToBundles(clumpList, isRc, &subQuery, minScore, &rangeList);
gfRangeFreeList(&rangeList);
}
gfClumpFreeList(&clumpList);
}
addToBigBundleList(&oneBunList, bunHash, &bigBunList, query);
*endPos = saveEnd;
}
#ifdef DEBUG
dumpBunList(bigBunList);
#endif /* DEBUG */
for (bun = bigBunList; bun != NULL; bun = bun->next)
{
ssStitch(bun, ffCdna, minScore, ssAliCount);
if (!fastMap && !band)
refineSmallExonsInBundle(bun);
saveAlignments(bun->genoSeq->name, bun->genoSeq->size, 0,
bun, NULL, isRc, FALSE, ffCdna, minScore, out);
}
ssBundleFreeList(&bigBunList);
freeHash(&bunHash);
lmCleanup(&lm);
}
