int main(int argc, char** argv) {
struct mafFile* mf;
struct mafAli* ali;
struct mafComp* mc;
if ( argc != 2) {
printf("remove_self maf-file\n");
return 1;
}
init_scores70();
mafWriteStart(stdout, 0);
mf = mafOpen(argv[1], 0);
while((ali = mafNext(mf)) != NULL) {
mc = ali->components;
if(mc->next->strand == '+' && mc->start > mc->next->start)
continue;
else if(mc->next->strand == '-' && mc->start > (mc->next->srcSize - mc->next->start - mc->next->size))
continue;
else
mafWrite(stdout, ali);
}
mafFileFree(&mf);
mafWriteEnd(stdout);
return 0;
}
void mafMeFirst(char *inMaf, char *meFile, char *outMaf)
/* mafMeFirst - Move component to top if it is one of the named ones. Useful
* in conjunction with mafFrags when you don't want the one with the gene name
* to be in the middle.. */
{
struct hash *meHash = hashWordsInFile(meFile, 18);
struct mafFile *mf = mafOpen(inMaf);
FILE *f = mustOpen(outMaf, "w");
mafWriteStart(f, mf->scoring);
struct mafAli *maf;
while ((maf = mafNext(mf)) != NULL)
{
struct mafComp *comp = compInHash(maf, meHash);
if (comp == NULL)
errAbort("No components in %s in maf ending line %d of %s",
meFile, mf->lf->lineIx, mf->lf->fileName);
slRemoveEl(&maf->components, comp);
slAddHead(&maf->components, comp);
mafWrite(f, maf);
mafAliFree(&maf);
}
mafWriteEnd(f);
carefulClose(&f);
}
int main (int argc, char **argv)
{
int nrow = DEFAULT_NROW,
which_seq = 0,
do_order = 0,
misses_allowed = 0,
output_mode = -1;
float cutoff = DEFAULT_CUTOFF;
char *strand = NULL,
*id = NULL;
struct mafFile *file = NULL;
struct mafAli *ali = NULL;
struct mafComp *comp = NULL;
struct MOTIF *motif = NULL;
struct MATCH *matches = NULL;
id = ckalloc (STRSIZE);
get_args (argc, argv, &output_mode, &file, &id, &motif, &do_order, &nrow, &cutoff, &misses_allowed);
strand = ckalloc (sizeof (char) * (nrow + 1));
while (NULL != (ali = mafNext (file)))
{
for (comp = ali->components, which_seq = 0; comp; comp = comp->next, which_seq++)
strand[which_seq] = comp->strand;
strand[which_seq] = '\0';
/* skip blocks that don't have all seqs in them */
if (which_seq != nrow)
{
mafAliFree (&ali);
continue;
}
/* forward strand */
get_matches (&matches, FORWARD, ali, nrow, motif, do_order, misses_allowed);
/* reverse strand */
for (comp = ali->components; comp; comp = comp->next)
do_revcomp((uchar *)comp->text, ali->textSize );
get_matches (&matches, REVERSE, ali, nrow, motif, do_order, misses_allowed);
/* output matches */
if (matches)
output_matches (matches, strand, id, nrow, motif);
free_match_list (&matches);
mafAliFree (&ali);
}
mafFileFree (&file);
free (strand);
free_motif_list (&motif);
free (id);
return 0;
}
struct mafFile *mafReadAll(char *fileName)
/* Read all elements in a maf file */
{
struct mafFile *mf = mafOpen(fileName);
struct mafAli *ali;
while ((ali = mafNext(mf)) != NULL)
{
slAddHead(&mf->alignments, ali);
}
slReverse(&mf->alignments);
return mf;
}
/* Construct a malnSet from a MAF file. defaultBranchLength is used to
* assign branch lengths when inferring trees from the MAF. */
struct malnSet *malnSet_constructFromMaf(struct Genomes *genomes, char *mafFileName, int maxInputBlkWidth, double defaultBranchLength, struct Genome *treelessRootGenome) {
struct malnSet *malnSet = malnSet_construct(genomes, mafFileName);
struct mafFile *mafFile = mafOpen(mafFileName);
struct mafAli *ali;
while ((ali = mafNext(mafFile)) != NULL) {
checkMafAli(ali);
addMafAli(malnSet, ali, maxInputBlkWidth, defaultBranchLength, treelessRootGenome);
mafAliFree(&ali);
}
malnSet_assert(malnSet);
mafFileFree(&mafFile);
return malnSet;
}
void mafStats(char *twoBitFile, char *mafDir, char *outFile)
/* mafStats - Calculate basic stats on maf file including species-by-species
* coverage and percent ID. */
{
struct twoBitFile *tbf = twoBitOpen(twoBitFile);
FILE *f = mustOpen(outFile, "w");
struct twoBitIndex *ix;
long genomeSize = 0;
struct hash *speciesHash = hashNew(0);
struct speciesAcc *speciesList = NULL, *species;
for (ix = tbf->indexList; ix != NULL; ix = ix->next)
{
unsigned chromSize = twoBitSeqSizeNoNs(tbf, ix->name);
genomeSize += chromSize;
char mafFileName[PATH_LEN];
safef(mafFileName, sizeof(mafFileName), "%s/%s.maf", mafDir, ix->name);
struct mafFile *mf = mafMayOpen(mafFileName);
verbose(1, "processing %s\n", ix->name);
if (mf == NULL)
{
warn("%s doesn't exist", mafFileName);
continue;
}
struct mafAli *maf;
while ((maf = mafNext(mf)) != NULL)
{
struct mafComp *mc;
for (mc = maf->components; mc != NULL; mc = mc->next)
{
if (mc->text != NULL)
toUpperN(mc->text, maf->textSize);
}
addCounts(maf, speciesHash, &speciesList);
mafAliFree(&maf);
}
mafFileFree(&mf);
}
slReverse(&speciesList);
for (species = speciesList; species != NULL; species = species->next)
{
fprintf(f, "counts: %s\t%ld\t%ld\t%ld\n", species->name, species->covCount, species->aliCount, species->idCount);
fprintf(f, "precents: %s\t%4.2f%%\t%4.2f%%\t%4.2f%%\n",
species->name, 100.0 * species->covCount/genomeSize,
100.0 * species->aliCount/genomeSize,
100.0 * species->idCount/species->aliCount);
}
carefulClose(&f);
}
void mafToPsl(char *querySrc, char *targetSrc, char *inName, char *outName)
/* mafToPsl - Convert maf to psl format. */
{
struct mafFile *mf = mafOpen(inName);
FILE *pslFh = mustOpen(outName, "w");
struct mafAli *maf;
while ((maf = mafNext(mf)) != NULL)
{
mafAliToPsl(querySrc, targetSrc, maf, pslFh);
mafAliFree(&maf);
}
carefulClose(&pslFh);
mafFileFree(&mf);
}
void mafToFa(char *inName, char *outName)
/* mafToFa - convert maf file to fasta. */
{
struct mafFile *mf = mafOpen(inName);
FILE *faFh = mustOpen(outName, "w");
struct mafAli *maf;
while ((maf = mafNext(mf)) != NULL)
{
mafAliToFa(maf, faFh);
mafAliFree(&maf);
}
carefulClose(&faFh);
mafFileFree(&mf);
}
struct mafAli *mafLoadInRegion2(struct sqlConnection *conn,
struct sqlConnection *conn2, char *table, char *chrom,
int start, int end, char *file)
/* Return list of alignments in region. */
{
char **row;
unsigned int extFileId = 0;
struct mafAli *maf, *mafList = NULL;
struct mafFile *mf = NULL;
int rowOffset;
if (file != NULL)
mf = mafOpen(file);
struct sqlResult *sr = hRangeQuery(conn, table, chrom,
start, end, NULL, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
struct scoredRef ref;
scoredRefStaticLoad(row + rowOffset, &ref);
if ((file != NULL) && (ref.extFile != 0))
errAbort("expect extFile to be zero if file specified\n");
if ((file == NULL) && (ref.extFile == 0))
errAbort("expect extFile to be not zero or file specified\n");
if (ref.extFile != extFileId)
{
char *path = hExtFileNameC(conn2, "extFile", ref.extFile);
mafFileFree(&mf);
mf = mafOpen(path);
extFileId = ref.extFile;
}
lineFileSeek(mf->lf, ref.offset, SEEK_SET);
maf = mafNext(mf);
if (maf == NULL)
internalErr();
slAddHead(&mafList, maf);
}
sqlFreeResult(&sr);
mafFileFree(&mf);
slReverse(&mafList);
/* hRangeQuery may return items out-of-order when bin is used in the query,
* so sort here in order to avoid trouble at base-level view: */
slSort(&mafList, mafCmp);
return mafList;
}
struct mafAli *readMafs(struct mafFile *mf)
{
struct mafAli *maf;
char buffer[2048];
char buffer2[2048];
struct strandHead *strandHead;
struct mafAli *mafList = NULL;
while((maf = mafNext(mf)) != NULL)
{
struct mafComp *mc, *masterMc = maf->components;
char *species = buffer;
char *chrom;
strcpy(species, masterMc->src);
chrom = strchr(species,'.');
if (chrom)
*chrom++ = 0;
if (masterSpecies == NULL)
{
masterSpecies = cloneString(species);
masterChrom = cloneString(chrom);
//printf("master %s %s\n",masterSpecies,masterChrom);
}
else
{
if (!sameString(masterSpecies, species))
errAbort("first species (%s) not master species (%s)\n",species,masterSpecies);
}
for(mc= masterMc->next; mc; mc = mc->next)
{
struct linkBlock *linkBlock;
struct subSpecies *subSpecies = NULL;
strcpy(species, mc->src);
chrom = strchr(species,'.');
*chrom++ = 0;
if ((subSpecies = hashFindVal(speciesHash, species)) == NULL)
{
//printf("new species %s\n",species);
AllocVar(subSpecies);
subSpecies->name = cloneString(species);
subSpecies->hash = newHash(6);
subSpecies->blockStatus.strand = '+';
subSpecies->blockStatus.masterStart = masterMc->start;
slAddHead(&speciesList, subSpecies);
hashAdd(speciesHash, species, subSpecies);
}
subSpecies->blockStatus.masterEnd = masterMc->start + masterMc->size ;
sprintf(buffer2, "%s%c%s", masterChrom,mc->strand,chrom);
if ((strandHead = hashFindVal(subSpecies->hash, buffer2)) == NULL)
{
//printf("new strand %s for species %s\n",buffer2, species);
AllocVar(strandHead);
hashAdd(subSpecies->hash, buffer2, strandHead);
strandHead->name = cloneString(buffer2);
strandHead->species = cloneString(species);
strandHead->qName = cloneString(chrom);
strandHead->qSize = mc->srcSize;
strandHead->strand = mc->strand;
slAddHead(&strandHeads, strandHead);
}
AllocVar(linkBlock);
linkBlock->mc = mc;
linkBlock->cb.qStart = mc->start;
linkBlock->cb.qEnd = mc->start + mc->size;
linkBlock->cb.tStart = masterMc->start;
linkBlock->cb.tEnd = masterMc->start + masterMc->size;
slAddHead(&strandHead->links, linkBlock);
}
slAddHead(&mafList, maf);
}
slReverse(&mafList);
return mafList;
}
void scanMaf(char *database, char *fileName, struct hash *chromHash, boolean covRestrict, int spCount)
/* Scan through maf file (which must be sorted by
* chromosome) and fill in coverage histograms on
* each chromosome. */
{
struct mafFile *mf = mafOpen(fileName);
struct mafAli *ali = NULL;
struct mafComp *comp = NULL;
struct chromSizes *lastCs = NULL, *cs = NULL;
char *chrom = NULL;
int start = 0, end = 0, size = 0, j, k;
int idStart = 0, idEnd = 0, idSize = 0;
UBYTE *cov = NULL;
UBYTE *align = NULL;
UBYTE *id = NULL;
char *tPtr[MAXALIGN];
bool hit = FALSE;
while ((ali = mafNext(mf)) != NULL)
{
int cCount = slCount(ali->components);
int i = 1;
int nextStart, idNextStart;
comp = ali->components;
tPtr[0] = comp->text;
chrom = strchr(comp->src,'.')+1;
if (chrom == NULL)
chrom = comp->src;
start = comp->start;
idStart = comp->start;
nextStart = idNextStart = start;
cs = hashMustFindVal(chromHash, chrom);
if (cs != lastCs)
{
if (lastCs != NULL)
closeChromCov(fileName, lastCs, &cov, &align, &id);
AllocArray(cov, cs->totalSize);
AllocArray(align, cs->totalSize);
AllocArray(id, cs->totalSize);
if (covRestrict)
{
restrictCov(cov, cs->totalSize, cs->restrictList);
restrictCov(align, cs->totalSize, cs->restrictList);
restrictCov(id, cs->totalSize, cs->restrictList);
}
restrictGaps(database, cov, cs->totalSize, chrom);
restrictGaps(database, align, cs->totalSize, chrom);
restrictGaps(database, id, cs->totalSize, chrom);
cs->unrestrictedSize = calcUnrestrictedSize(cov, cs->totalSize);
lastCs = cs;
}
/* don't count if few alignments than spCount */
if ((ali->components->next == NULL) || (cCount < spCount))
{
mafAliFree(&ali);
continue;
}
//printf("coverage %d, size %d\n", start, comp->size);
incNoOverflow(cov+start, comp->size);
for (comp = ali->components->next; comp != NULL; comp = comp->next)
{
if (comp->size > 0) // do not process e lines
{
tPtr[i] = comp->text;
i++;
assert (i < MAXALIGN-1);
}
else
--cCount;
}
size = 0;
assert(cs != NULL);
/* count gapless columns */
for (j = 0 ; j<ali->textSize ; j++)
{
hit = TRUE;
/* look for aligning bases in query seqs , abort if any is a gap */
for (i = 1 ; i < cCount ; i++)
{
if (tPtr[i][j] == '-' || tPtr[0][j] == '-')
{
// printf("align %d, size %d\n", start, size);
incNoOverflow(align+start, size);
cs->totalDepth += size;
start = nextStart;
size = 0;
hit = FALSE;
break;
}
}
if (hit)
size++;
/* if there is a gap in the target, start a new alignment block*/
if (tPtr[0][j] != '-')
nextStart++;
}
assert(cs!=NULL);
end = start+size;
if (end > cs->totalSize)
{
if (cs->name != NULL)
errAbort("End %d past end %ld of %f\n", end, (long)cs->totalSize, ali->score);
else
{
if (ali!=NULL)
errAbort("End %d past end %ld %f\n", end, (long)cs->totalSize, ali->score );
else
errAbort("End %d past end %ld \n", end, (long)cs->totalSize);
}
}
incNoOverflow(align+start, size-1);
cs->totalDepth += size-1;
/* count percent id */
idSize = 0;
assert(cs != NULL);
for (k = 0 ; k<ali->textSize ; k++)
{
hit = TRUE;
char tc = toupper(tPtr[0][k]);
for (i = 1 ; i < cCount ; i++)
{
if (toupper(tPtr[i][k]) != tc || tc == '-' || tc == 'N')
{
incNoOverflow(id+idStart, idSize);
idStart = idNextStart;
idSize = 0;
hit = FALSE;
break;
}
}
if (hit)
idSize++;
/* skip over gaps */
if (tc != '-')
idNextStart++;
}
assert(cs!=NULL);
idEnd = idStart+idSize;
if (idEnd > cs->totalSize)
{
if (cs->name != NULL)
errAbort("End %d past end %ld of %f\n", idEnd, (long)cs->totalSize, ali->score);
else
{
if (ali!=NULL)
errAbort("End %d past end %ld %f\n", idEnd, (long)cs->totalSize, ali->score );
else
errAbort("End %d past end %ld \n", idEnd, (long)cs->totalSize);
}
}
incNoOverflow(id+idStart, idSize-1);
mafAliFree(&ali);
}
closeChromCov(fileName, cs, &cov, &align, &id);
}
void extractMafs(char *file, FILE *f, struct hash *regionHash)
/* extract MAFs in a file from regions specified in hash */
{
char *chrom = NULL;
struct bed *bed = NULL;
struct mafFile *mf = mafOpen(file);
struct mafAli *maf = NULL;
struct mafComp *mc;
char path[256];
verbose(1, "extracting from %s\n", file);
maf = mafNext(mf);
while (maf)
{
mc = maf->components;
if (!chrom || differentString(chrom, chromFromSrc(mc->src)))
chrom = cloneString(chromFromSrc(mc->src)); /* new chrom */
bed = (struct bed *)hashFindVal(regionHash, chrom);
if (!bed)
{
/* no regions on this chrom -- skip to next chrom */
do
mafAliFree(&maf);
while (((maf = mafNext(mf)) != NULL) && sameString(chromFromSrc(maf->components->src), chrom));
continue; // start over with this maf
}
verbose(2, "region: %s:%d-%d\n",
bed->chrom, bed->chromStart+1, bed->chromEnd);
if (outDir)
{
if (f)
endOutFile(f);
safef(path, sizeof (path), "%s/%s.maf", dir, bed->name);
f = startOutFile(path);
}
/* skip mafs before region, stopping if chrom changes */
while (maf && (mc = maf->components) && sameString(chrom, chromFromSrc(mc->src)) &&
(mc->start + mc->size) <= bed->chromStart)
{
mafAliFree(&maf);
maf = mafNext(mf);
}
/* extract all mafs and pieces of mafs in region */
while (maf && (mc = maf->components) && sameString(chrom, chromFromSrc(mc->src)) &&
(bed->chromStart < mc->start + mc->size && bed->chromEnd > mc->start))
{
int mafStart = mc->start;
int mafEnd = mc->start + mc->size;
struct mafAli *full = maf;
if (mafStart < bed->chromStart || mafEnd > bed->chromEnd)
{
full = maf;
maf = mafSubsetE(full, mc->src, bed->chromStart, bed->chromEnd, keepInitialGaps);
mc = maf->components;
}
verbose(2, " %s:%d-%d\n", chrom, mc->start+1, mc->start + mc->size);
mafWrite(f, maf);
struct mafAli *nextMaf = (mafEnd > bed->chromEnd+1)
? mafSubset(full, mc->src, bed->chromEnd+1, mafEnd) : mafNext(mf);
if (maf != full)
mafAliFree(&maf);
mafAliFree(&full);
maf = nextMaf;
}
/* get next region */
hashRemove(regionHash, bed->chrom);
if (bed->next)
hashAdd(regionHash, bed->chrom, bed->next);
}
mafFileFree(&mf);
}
void hgLoadMafSummary(char *db, char *table, char *fileName)
/* hgLoadMafSummary - Load a summary table of pairs in a maf into a database. */
{
long mafCount = 0, allMafCount = 0;
struct mafComp *mcMaster = NULL;
struct mafAli *maf;
struct mafFile *mf = mafOpen(fileName);
struct sqlConnection *conn;
FILE *f = hgCreateTabFile(".", table);
long componentCount = 0;
struct hash *componentHash = newHash(0);
if (!test)
{
conn = sqlConnect(database);
mafSummaryTableCreate(conn, table, hGetMinIndexLength(db));
}
verbose(1, "Indexing and tabulating %s\n", fileName);
/* process mafs */
while ((maf = mafNext(mf)) != NULL)
{
mcMaster = mafMaster(maf, mf, fileName);
allMafCount++;
if (mcMaster->srcSize < minSeqSize)
continue;
while (mcMaster->size > maxSize)
{
/* break maf into maxSize pieces */
int end = mcMaster->start + maxSize;
struct mafAli *subMaf =
mafSubset(maf, mcMaster->src, mcMaster->start, end);
verbose(3, "Splitting maf %s:%d len %d\n", mcMaster->src,
mcMaster->start, mcMaster->size);
componentCount +=
processMaf(subMaf, componentHash, f, mf, fileName);
mafAliFree(&subMaf);
subMaf = mafSubset(maf, mcMaster->src,
end, end + (mcMaster->size - maxSize));
mafAliFree(&maf);
maf = subMaf;
mcMaster = mafMaster(maf, mf, fileName);
}
if (mcMaster->size != 0)
{
/* remainder of maf after splitting off maxSize submafs */
componentCount +=
processMaf(maf, componentHash, f, mf, fileName);
}
mafAliFree(&maf);
mafCount++;
}
mafFileFree(&mf);
flushSummaryBlocks(componentHash, f);
verbose(1,
"Created %ld summary blocks from %ld components and %ld mafs from %s\n",
summaryCount, componentCount, allMafCount, fileName);
if (test)
return;
verbose(1, "Loading into %s table %s...\n", database, table);
hgLoadTabFile(conn, ".", table, &f);
verbose(1, "Loading complete");
hgEndUpdate(&conn, "Add %ld maf summary blocks from %s\n",
summaryCount, fileName);
}
