int main(int argc, char *argv[])
{
char *estName, *targetName, *oocName;
FILE *estFile;
struct dnaSeq *target;
struct dnaSeq *est;
struct patSpace *ps;
struct patClump *clumpList, *clump;
int estIx = 0;
/* Check command line arguments and assign to local variables. */
if (argc != 4)
usage();
estName = argv[1];
estFile = mustOpen(estName, "rb");
targetName = argv[2];
oocName = argv[3];
/* Read in target DNA from fasta files and check not too big. */
fprintf(stderr, "Reading %s\n", targetName);
target = faReadAllDna(targetName);
if (totalSequenceSize(target) > 8000000)
{
errAbort("Can only handle 8000000 bases of genomic sequence at once, %s has %d.",
targetName, totalSequenceSize(target));
}
/* Make a pattern space index structure. */
fprintf(stderr, "Making Pattern Space index\n");
ps = makePatSpace(&target, 1, oocName, 4, 32000);
/* Loop through each EST in query list. */
printf("Searching for hits\n\n");
while (faReadNext(estFile, NULL, TRUE, NULL, &est))
{
boolean isRc; /* Reverse complemented? */
if (++estIx % 5000 == 0)
fprintf(stderr, "Processing EST %d\n", estIx);
if (est->size > 20000)
{
warn("Very large EST sequence %s.\n"
"Maybe you mixed up the EST and genomic parameters?", est->name);
usage();
}
for (isRc = 0; isRc <= 1; ++isRc) /* Search both strands. */
{
if (isRc)
reverseComplement(est->dna, est->size);
clumpList = patSpaceFindOne(ps, est->dna, est->size);
/* For each homology clump patSpace finds, do a fuzzyFinder
* alignment of it and print the results. */
for (clump = clumpList; clump != NULL; clump = clump->next)
{
struct ffAli *ali, *a;
boolean isRc;
int score;
struct dnaSeq *t = clump->seq;
DNA *tStart = t->dna + clump->start;
ali = ffFind(est->dna, est->dna+est->size, tStart, tStart + clump->size, ffCdna);
if (ali != NULL)
{
score = ffScoreCdna(ali);
printf("%s hits %s strand %c score %d\n",
est->name, t->name, (isRc ? '+' : '-'), score);
for (a = ali; a != NULL; a = a->right)
{
printf(" Q %4d - %4d\t T %4d -%4d\n",
a->nStart - est->dna, a->nEnd - est->dna,
a->hStart - t->dna, a->hEnd - t->dna);
}
printf("\n");
ffFreeAli(&ali);
}
else
{
printf("Couldn't align clump at %s %d-%d\n",
t->name, clump->start, clump->start + clump->size);
}
}
slFreeList(&clumpList);
}
freeDnaSeq(&est);
}
/* Clean up time. */
freePatSpace(&ps);
freeSeqList(&target);
return 0;
}
int main(int argc, char *argv[])
{
char *genoListName;
char *cdnaListName;
char *oocFileName;
char *pairFileName;
struct patSpace *patSpace;
long startTime, endTime;
char **genoList;
int genoListSize;
char *genoListBuf;
char **cdnaList;
int cdnaListSize;
char *cdnaListBuf;
char *genoName;
int i;
int estIx = 0;
struct dnaSeq **seqListList = NULL, *seq;
static char hitFileName[512], mergerFileName[512], okFileName[512];
char *outRoot;
struct hash *pairHash;
if (dumpMe)
{
bigHtmlFile = mustOpen("C:\\inetpub\\wwwroot\\test\\patAli.html", "w");
littleHtmlFile = mustOpen("C:\\inetpub\\wwwroot\\test\\patSpace.html", "w");
htmStart(bigHtmlFile, "PatSpace Alignments");
htmStart(littleHtmlFile, "PatSpace Index");
}
if ((hostName = getenv("HOST")) == NULL)
hostName = "";
if (argc != 6)
usage();
pushWarnHandler(patSpaceWarnHandler);
startTime = clock1000();
dnaUtilOpen();
makePolys();
genoListName = argv[1];
cdnaListName = argv[2];
oocFileName = argv[3];
pairFileName = argv[4];
outRoot = argv[5];
sprintf(hitFileName, "%s.hit", outRoot);
sprintf(mergerFileName, "%s.glu", outRoot);
sprintf(okFileName, "%s.ok", outRoot);
readAllWords(genoListName, &genoList, &genoListSize, &genoListBuf);
readAllWords(cdnaListName, &cdnaList, &cdnaListSize, &cdnaListBuf);
pairHash = makePairHash(pairFileName);
hitOut = mustOpen(hitFileName, "w");
mergerOut = mustOpen(mergerFileName, "w");
dumpOut = mustOpen("dump.out", "w");
seqListList = needMem(genoListSize*sizeof(seqListList[0]) );
fprintf(hitOut, "Pattern space 0.2 cDNA matcher\n");
fprintf(hitOut, "cDNA files: ", cdnaListSize);
for (i=0; i<cdnaListSize; ++i)
fprintf(hitOut, " %s", cdnaList[i]);
fprintf(hitOut, "\n");
fprintf(hitOut, "%d genomic files\n", genoListSize);
for (i=0; i<genoListSize; ++i)
{
genoName = genoList[i];
if (!startsWith("//", genoName) )
{
seqListList[i] = seq = faReadAllDna(genoName);
fprintf(hitOut, "%d els in %s ", slCount(seq), genoList[i]);
for (; seq != NULL; seq = seq->next)
fprintf(hitOut, "%d ", seq->size);
fprintf(hitOut, "\n");
}
}
patSpace = makePatSpace(seqListList, genoListSize, oocFileName);
for (i=0; i<cdnaListSize; ++i)
{
FILE *f;
char *estFileName;
DNA *dna;
char *estName;
int size;
int c;
int maxSizeForFuzzyFind = 20000;
int dotCount = 0;
estFileName = cdnaList[i];
if (startsWith("//", estFileName) )
continue;
f = mustOpen(estFileName, "rb");
while ((c = fgetc(f)) != EOF)
if (c == '>')
break;
printf("%s", cdnaList[i]);
fflush(stdout);
while (fastFaReadNext(f, &dna, &size, &estName))
{
aliSeqName = estName;
if (size < maxSizeForFuzzyFind) /* Some day need to fix this somehow... */
{
struct hashEl *hel;
struct cdnaAliList *calList = NULL;
hel = hashLookup(pairHash, estName);
if (hel != NULL) /* Do pair processing. */
{
struct estPair *ep;
struct seq *thisSeq, *otherSeq;
ep = hel->val;
if (hel->name == ep->name3)
{
thisSeq = &ep->seq3;
otherSeq = &ep->seq5;
}
else
{
thisSeq = &ep->seq5;
otherSeq = &ep->seq3;
}
if (otherSeq->dna == NULL) /* First in pair - need to save sequence. */
{
thisSeq->size = size;
thisSeq->dna = needMem(size);
memcpy(thisSeq->dna, dna, size);
}
else /* Second in pair - do gluing and free partner. */
{
char mergedName[64];
thisSeq->dna = dna;
thisSeq->size = size;
sprintf(mergedName, "%s_AND_%s", ep->name5, ep->name3);
patSpaceFindOne(patSpace, ep->seq5.dna, ep->seq5.size,
'+', '5', ep->name5, &calList);
reverseComplement(ep->seq5.dna, ep->seq5.size);
patSpaceFindOne(patSpace, ep->seq5.dna, ep->seq5.size,
'-', '5', ep->name5, &calList);
patSpaceFindOne(patSpace, ep->seq3.dna, ep->seq3.size,
'+', '3', ep->name3, &calList);
reverseComplement(ep->seq3.dna, ep->seq3.size);
patSpaceFindOne(patSpace, ep->seq3.dna, ep->seq3.size,
'-', '3', ep->name3, &calList);
slReverse(&calList);
writeMergers(calList, mergedName, genoList);
freez(&otherSeq->dna);
thisSeq->dna = NULL;
thisSeq->size =otherSeq->size = 0;
}
}
else
{
patSpaceFindOne(patSpace, dna, size, '+', '5', estName, &calList);
reverseComplement(dna, size);
patSpaceFindOne(patSpace, dna, size, '-', '5', estName, &calList);
slReverse(&calList);
writeMergers(calList, estName, genoList);
}
++estIx;
if ((estIx & 0xfff) == 0)
{
printf(".");
++dotCount;
fflush(stdout);
}
}
}
printf("\n");
}
aliSeqName = "";
printf("raw %4d ffSubmitted %3d ffAccepted %3d ffOkScore %3d ffSolidMatch %2d\n",
grandTotalHits, ffSubmitted, ffAccepted, ffOkScore, ffSolidMatch);
endTime = clock1000();
printf("Total time is %4.2f\n", 0.001*(endTime-startTime));
/* Write out file who's presense say's we succeeded */
{
FILE *f = mustOpen(okFileName, "w");
fputs("ok", f);
fclose(f);
}
if (dumpMe)
{
htmEnd(bigHtmlFile);
htmEnd(littleHtmlFile);
}
return 0;
}
void fakeFinContigs(char *agpName, char *faName, char *finDir, char *rootName, char *finFaDir, char *ooVer)
/* fakeFinContigs - Fake up contigs for a finished chromosome. */
{
struct contig *contigList = NULL, *contig = NULL;
struct agpFrag *agp;
struct lineFile *lf = lineFileOpen(agpName, TRUE);
char *line, *words[16];
int lineSize, wordCount;
int contigIx = 0;
char liftDir[512], contigDir[512], path[512];
char chrom[128];
FILE *f;
struct dnaSeq *seq;
int fragIx;
/* Build up contig list by scanning agp file. */
printf("Reading %s\n", lf->fileName);
while (lineFileNext(lf, &line, &lineSize))
{
if (line[0] == '#' || line[0] == 0)
continue;
wordCount = chopLine(line, words);
if (wordCount < 5)
errAbort("Expecting at least 5 words line %d of %s", lf->lineIx, lf->fileName);
if (words[4][0] == 'N' || words[4][0] == 'U')
{
contig = NULL;
continue;
}
lineFileExpectWords(lf, 9, wordCount);
agp = agpFragLoad(words);
// file is 1-based but agpFragLoad() now assumes 0-based:
agp->chromStart -= 1;
agp->fragStart -= 1;
if (contig == NULL)
{
AllocVar(contig);
sprintf(contig->name, "%s%d", rootName, ++contigIx);
contig->startOffset = agp->chromStart;
slAddHead(&contigList, contig);
}
else
{
if (contig->agpList != NULL && contig->agpList->chromEnd != agp->chromStart)
errAbort("Start doesn't match previous end line %d of %s",
lf->lineIx, lf->fileName);
}
if (agp->chromEnd - agp->chromStart != agp->fragEnd - agp->fragStart)
errAbort("Chrom and frag size mismatch line %d of %s", lf->lineIx, lf->fileName);
slAddHead(&contig->agpList, agp);
contig->endOffset = agp->chromEnd;
}
slReverse(&contigList);
for (contig = contigList; contig != NULL; contig = contig->next)
slReverse(&contig->agpList);
lineFileClose(&lf);
/* Load up chromosome sequence and make sure it is in one piece. */
printf("Reading %s\n", faName);
seq = faReadAllDna(faName);
if (slCount(seq) != 1)
errAbort("Got %d sequences in %s, can only handle one.", slCount(seq), faName);
/* Fix up agp coordinates. Make a directory for each contig. Fill it with
* .fa .agp barge.NN files for that contig. */
printf("Writing contig dirs\n");
for (contig = contigList; contig != NULL; contig = contig->next)
{
/* Make Contig dir. */
sprintf(contigDir, "%s/%s", finDir, contig->name);
makeDir(contigDir);
/* Make contig.agp file. */
sprintf(path, "%s/%s.agp", contigDir, contig->name);
f = mustOpen(path, "w");
fragIx = 0;
for (agp = contig->agpList; agp != NULL; agp = agp->next)
{
char buf[128];
sprintf(buf, "%s/%s", skipChr(agp->chrom), contig->name);
freez(&agp->chrom);
agp->chrom = cloneString(buf);
agp->chromStart -= contig->startOffset;
agp->chromEnd -= contig->startOffset;
agp->ix = ++fragIx;
agpFragTabOut(agp, f);
}
carefulClose(&f);
/* Make ooGreedy.NN.gl file */
sprintf(path, "%s/%s.%s.gl", contigDir, "ooGreedy", ooVer);
f = mustOpen(path, "w");
for (agp = contig->agpList; agp != NULL; agp = agp->next)
{
if (agp->type[0] != 'N' && agp->type[0] != 'U')
{
fprintf(f, "%s_1\t%d\t%d\t%s\n", agp->frag,
agp->chromStart,
agp->chromEnd,
agp->strand);
}
}
carefulClose(&f);
/* Make contig.fa file. */
sprintf(path, "%s/%s.fa", contigDir, contig->name);
faWrite(path, contig->name, seq->dna + contig->startOffset,
contig->endOffset - contig->startOffset);
/* Make contig/barge file. */
sprintf(path, "%s/barge.%s", contigDir, ooVer);
f = mustOpen(path, "w");
fprintf(f, "Barge (Connected Clone) File ooGreedy Version %s\n", ooVer);
fprintf(f, "\n");
fprintf(f, "start accession size overlap maxClone maxOverlap\n");
fprintf(f, "------------------------------------------------------------\n");
for (agp = contig->agpList; agp != NULL; agp = agp->next)
{
char clone[128];
strcpy(clone, agp->frag);
chopSuffix(clone);
fprintf(f, "%d\t%s\t%d\t100\tn/a\t0\n", agp->chromStart,
clone, agp->chromEnd);
}
carefulClose(&f);
/* Make contig/gold file. */
sprintf(path, "%s/gold.%s", contigDir, ooVer);
f = mustOpen(path, "w");
fragIx = 0;
for (agp = contig->agpList; agp != NULL; agp = agp->next)
{
char fragName[128];
struct agpFrag frag = *agp;
sprintf(fragName, "%s_1", agp->frag);
frag.frag = fragName;
frag.type[0] = '0';
agpFragTabOut(&frag, f);
}
carefulClose(&f);
}
/* Create lift subdirectory. */
printf("Creating lift files\n");
sprintf(liftDir, "%s/lift", finDir);
makeDir(liftDir);
/* Create lift/oOut.lst file (just a list of contigs). */
sprintf(path, "%s/oOut.lst", liftDir);
f = mustOpen(path, "w");
for (contig = contigList; contig != NULL; contig = contig->next)
fprintf(f, "%s/%s.fa.out\n", contig->name, contig->name);
carefulClose(&f);
/* Create lift/ordered.lst file (just a list of contigs). */
sprintf(path, "%s/ordered.lst", liftDir);
f = mustOpen(path, "w");
for (contig = contigList; contig != NULL; contig = contig->next)
fprintf(f, "%s\n", contig->name);
carefulClose(&f);
/* Create lift/ordered.lft file. */
sprintf(path, "%s/ordered.lft", liftDir);
f = mustOpen(path, "w");
splitPath(faName, NULL, chrom, NULL);
for (contig = contigList; contig != NULL; contig = contig->next)
fprintf(f, "%d\t%s/%s\t%d\t%s\t%d\n",
contig->startOffset, skipChr(chrom), contig->name,
contig->endOffset - contig->startOffset,
chrom, seq->size);
carefulClose(&f);
}
