void doOneChrom(char *database, char *chrom, char *rnaTable, char *expTable, FILE *f)
/* Process one chromosome. */
{
int chromSize = hChromSize(database, chrom);
struct binKeeper *bk = binKeeperNew(0, chromSize);
struct sqlConnection *conn = hAllocConn(database);
struct sqlResult *sr;
char **row;
struct bed *exp, *rna;
int rowOffset;
struct binElement *be, *beList;
int oneCount;
/* Load up expTable into bin-keeper. */
sr = hChromQuery(conn, expTable, chrom, NULL, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
exp = bedLoadN(row + rowOffset, 12);
binKeeperAdd(bk, exp->chromStart, exp->chromEnd, exp);
}
sqlFreeResult(&sr);
/* Loop through rnaTable and look at intersections. */
sr = hChromQuery(conn, rnaTable, chrom, NULL, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
rna = bedLoadN(row + rowOffset, 12);
beList = binKeeperFind(bk, rna->chromStart, rna->chromEnd);
oneCount = 0;
for (be = beList; be != NULL; be = be->next)
{
exp = be->val;
if (exp->strand[0] == rna->strand[0])
{
++oneCount;
++hitCount;
// fprintf(f, "%s:%d-%d\t%s\t%s\n",
// rna->chrom, rna->chromStart, rna->chromEnd, rna->name, exp->name);
}
}
slFreeList(&beList);
if (oneCount == 0)
{
++missCount;
fprintf(f, "miss %s:%d-%d %c %s\n", rna->chrom, rna->chromStart, rna->chromEnd, rna->strand[0], rna->name);
}
else if (oneCount == 1)
{
fprintf(f, "uniq %s:%d-%d %c %s\n", rna->chrom, rna->chromStart, rna->chromEnd, rna->strand[0], rna->name);
++uniqCount;
}
else
{
fprintf(f, "dupe %s:%d-%d %c %s\n", rna->chrom, rna->chromStart, rna->chromEnd, rna->strand[0], rna->name);
++dupeCount;
}
}
sqlFreeResult(&sr);
hFreeConn(&conn);
}
void checkInputOpenFiles(struct inInfo *array, int count)
/* Make sure all of the input is there and of right format before going forward. Since
* this is going to take a while we want to fail fast. */
{
int i;
for (i=0; i<count; ++i)
{
struct inInfo *in = &array[i];
switch (in->type)
{
case itBigWig:
{
/* Just open and close, it will abort if any problem. */
in->bbi = bigWigFileOpen(in->fileName);
break;
}
case itPromoterBed:
case itUnstrandedBed:
case itBlockedBed:
{
struct lineFile *lf = in->lf = lineFileOpen(in->fileName, TRUE);
char *line;
lineFileNeedNext(lf, &line, NULL);
char *dupe = cloneString(line);
char *row[256];
int wordCount = chopLine(dupe, row);
struct bed *bed = NULL;
switch (in->type)
{
case itPromoterBed:
lineFileExpectAtLeast(lf, 6, wordCount);
bed = bedLoadN(row, 6);
char strand = bed->strand[0];
if (strand != '+' && strand != '-')
errAbort("%s must be stranded, got %s in that field", lf->fileName, row[6]);
break;
case itUnstrandedBed:
lineFileExpectAtLeast(lf, 4, wordCount);
bed = bedLoadN(row, 4);
break;
case itBlockedBed:
lineFileExpectAtLeast(lf, 4, wordCount);
bed = bedLoadN(row, 12);
break;
default:
internalErr();
break;
}
bedFree(&bed);
freez(&dupe);
lineFileReuse(lf);
break;
}
default:
internalErr();
break;
}
}
}
static void mafFrags(char *database, char *track, char *bedFile, char *mafFile)
/* mafFrags - Collect MAFs from regions specified in a 6 column bed file. */
{
struct slName *orgList = NULL;
struct lineFile *lf = lineFileOpen(bedFile, TRUE);
FILE *f = mustOpen(mafFile, "w");
if (optionExists("orgs"))
{
char *orgFile = optionVal("orgs", NULL);
char *buf;
readInGulp(orgFile, &buf, NULL);
orgList = stringToSlNames(buf);
/* Ensure that org list starts with database. */
struct slName *me = slNameFind(orgList, database);
if (me == NULL)
errAbort("Need to have reference database '%s' in %s", database, orgFile);
if (me != orgList)
{
slRemoveEl(&orgList, me);
slAddHead(&orgList, me);
}
}
mafWriteStart(f, "zero");
if (bed12)
{
char *row[12];
while (lineFileRow(lf, row))
{
struct bed *bed = bedLoadN(row, ArraySize(row));
struct mafAli *maf = mafFromBed12(database, track, bed, orgList);
if (meFirst)
moveMeToFirst(maf, bed->name);
mafWrite(f, maf);
mafAliFree(&maf);
bedFree(&bed);
}
}
else
{
char *row[6];
while (lineFileRow(lf, row))
{
struct bed *bed = bedLoadN(row, ArraySize(row));
processBed6(database, track, f, bed, orgList);
bedFree(&bed);
}
}
mafWriteEnd(f);
carefulClose(&f);
}
void doTriangle(struct trackDb *tdb, char *item, char *motifTable)
/* Display detailed info on a regulatory triangle item. */
{
int start = cartInt(cart, "o");
struct dnaSeq *seq = NULL;
struct dnaMotif *motif = loadDnaMotif(item, motifTable);
char *table = tdb->table;
int rowOffset = hOffsetPastBin(database, seqName, table);
char query[256];
struct sqlResult *sr;
char **row;
struct bed *hit = NULL;
struct sqlConnection *conn = hAllocConn(database);
cartWebStart(cart, database, "Regulatory Motif Info");
genericBedClick(conn, tdb, item, start, 6);
sqlSafef(query, sizeof query,
"select * from %s where name = '%s' and chrom = '%s' and chromStart = %d",
table, item, seqName, start);
sr = sqlGetResult(conn, query);
row = sqlNextRow(sr);
if (row != NULL)
hit = bedLoadN(row + rowOffset, 6);
sqlFreeResult(&sr);
if (hit != NULL)
{
seq = hDnaFromSeq(database, hit->chrom, hit->chromStart, hit->chromEnd, dnaLower);
if (hit->strand[0] == '-')
reverseComplement(seq->dna, seq->size);
}
motifHitSection(seq, motif);
printTrackHtml(tdb);
}
void bigBedAddLinkedFeaturesFrom(struct track *track,
char *chrom, int start, int end, int scoreMin, int scoreMax, boolean useItemRgb,
int fieldCount, struct linkedFeatures **pLfList)
/* Read in items in chrom:start-end from bigBed file named in track->bbiFileName, convert
* them to linkedFeatures, and add to head of list. */
{
struct lm *lm = lmInit(0);
struct trackDb *tdb = track->tdb;
struct bigBedInterval *bb, *bbList = bigBedSelectRange(track, chrom, start, end, lm);
char *bedRow[32];
char startBuf[16], endBuf[16];
char *scoreFilter = cartOrTdbString(cart, track->tdb, "scoreFilter", NULL);
char *mouseOverField = cartOrTdbString(cart, track->tdb, "mouseOverField", NULL);
int minScore = 0;
if (scoreFilter)
minScore = atoi(scoreFilter);
int mouseOverIdx = bbExtraFieldIndex(tdb, mouseOverField);
for (bb = bbList; bb != NULL; bb = bb->next)
{
char* mouseOver = restField(bb, mouseOverIdx);
bigBedIntervalToRow(bb, chromName, startBuf, endBuf, bedRow, ArraySize(bedRow));
struct bed *bed = bedLoadN(bedRow, fieldCount);
struct linkedFeatures *lf = bedMungToLinkedFeatures(&bed, tdb, fieldCount,
scoreMin, scoreMax, useItemRgb);
if (scoreFilter == NULL || lf->score >= minScore)
slAddHead(pLfList, lf);
lf->mouseOver = mouseOver; // leaks some memory, cloneString handles NULL ifself
if (sameString(track->tdb->type, "bigGenePred"))
lf->original = genePredFromBigGenePred(chromName, bb);
}
lmCleanup(&lm);
}
static struct chromAnn* chromAnnBedReaderRead(struct chromAnnReader *car)
/* read next BED and convert to a chromAnn */
{
struct rowReader *rr = car->data;
if (!rowReaderNext(rr))
return NULL;
rowReaderExpectAtLeast(rr, 3);
char **rawCols = (car->opts & chromAnnSaveLines) ? rowReaderCloneColumns(rr) : NULL;
struct bed *bed = bedLoadN(rr->row, rr->numCols);
struct chromAnn *ca = chromAnnNew(bed->chrom, bed->strand[0], bed->name, rawCols,
strVectorWrite, strVectorFree);
if ((bed->blockCount == 0) || (car->opts & chromAnnRange))
{
if (car->opts & chromAnnCds)
{
if (bed->thickStart < bed->thickEnd)
chromAnnBlkNew(ca, bed->thickStart, bed->thickEnd);
}
else
chromAnnBlkNew(ca, bed->chromStart, bed->chromEnd);
}
else
addBedBlocks(ca, car->opts, bed);
chromAnnFinish(ca);
bedFree(&bed);
return ca;
}
void doPeakClusters(struct trackDb *tdb, char *item)
/* Display detailed info about a cluster of DNase peaks from other tracks. */
{
int start = cartInt(cart, "o");
char *table = tdb->table;
int rowOffset = hOffsetPastBin(database, seqName, table);
char query[256];
struct sqlResult *sr;
char **row;
struct bed *cluster = NULL;
struct sqlConnection *conn = hAllocConn(database);
cartWebStart(cart, database, "%s item details", tdb->shortLabel);
sqlSafef(query, sizeof(query),
"select * from %s where name = '%s' and chrom = '%s' and chromStart = %d",
table, item, seqName, start);
sr = sqlGetResult(conn, query);
row = sqlNextRow(sr);
if (row != NULL)
cluster = bedLoadN(row+rowOffset, 5);
sqlFreeResult(&sr);
if (cluster != NULL)
{
/* Get list of subgroups to display */
char *inputTableFieldDisplay = trackDbSetting(tdb, "inputTableFieldDisplay");
if (inputTableFieldDisplay != NULL)
{
struct slName *fieldList = stringToSlNames(inputTableFieldDisplay);
char *inputTrackTable = trackDbRequiredSetting(tdb, "inputTrackTable");
/* Print out some information about the cluster overall. */
printf("<B>Items in Cluster:</B> %s of %d<BR>\n", cluster->name,
sqlRowCount(conn, sqlCheckIdentifier(inputTrackTable)));
printf("<B>Cluster Score (out of 1000):</B> %d<BR>\n", cluster->score);
printPos(cluster->chrom, cluster->chromStart, cluster->chromEnd, NULL, TRUE, NULL);
/* In a new section put up list of hits. */
webNewSection("List of Items in Cluster");
webPrintLinkTableStart();
printClusterTableHeader(fieldList, FALSE, FALSE, TRUE);
printPeakClusterInfo(tdb, cart, conn, inputTrackTable, fieldList, cluster);
}
else
errAbort("Missing required trackDb setting %s for track %s",
"inputTableFieldDisplay", tdb->track);
webPrintLinkTableEnd();
}
printf("<A HREF=\"%s&g=htcListItemsAssayed&table=%s\" TARGET_blank>", hgcPathAndSettings(),
tdb->track);
printf("List all items assayed");
printf("</A><BR>\n");
webNewSection("Track Description");
printTrackHtml(tdb);
hFreeConn(&conn);
}
/* convert one line read from a bed file to a PSL */
void cnvBedRec(char *line, struct hash *chromSizes, FILE *pslFh)
{
char *row[12];
int numCols = chopByWhite(line, row, ArraySize(row));
if (numCols < 4)
errAbort("bed must have at least 4 columns");
struct bed *bed = bedLoadN(row, numCols);
struct psl* psl = bedToPsl(bed, chromSizes);
pslTabOut(psl, pslFh);
pslFree(&psl);
bedFree(&bed);
}
/* convert one line read from a bed file to a genePred */
void cnvBedRec(char *line, FILE *gpFh)
{
char *row[12];
int numCols = chopByWhite(line, row, ArraySize(row));
if (numCols < 4)
errAbort("bed must have at least 4 columns");
struct bed *bed = bedLoadN(row, numCols);
struct genePred* gp = bedToGenePred(bed);
genePredTabOut(gp, gpFh);
genePredFree(&gp);
bedFree(&bed);
}
struct bedPlus *bedPlusNLoad(char *row[], int wordCount, int bedWordCount)
/* Load a single row array like bedLoadN would... but handle extra info. */
{
struct bedPlus *newBp;
AllocVar(newBp);
if (bedWordCount > 0)
{
struct dyString *ds = newDyString(0);
int i;
newBp->bed = bedLoadN(row, bedWordCount);
for (i = bedWordCount; i < wordCount; i++)
{
dyStringAppend(ds, row[i]);
if (i < (wordCount-1))
dyStringAppendC(ds, '\t');
}
newBp->rest = dyStringCannibalize(&ds);
}
else
{
newBp->bed = bedLoadN(row, wordCount);
}
return newBp;
}
static struct bed *loadParTable(struct trackDb *tdb)
/* load all records in the par table */
{
struct bed *pars = NULL;
struct sqlConnection *conn = hAllocConn(database);
char query[512];
sqlSafef(query, sizeof(query), "select * from %s", tdb->table);
struct sqlResult *sr = sqlGetResult(conn, query);
char **row;
while ((row = sqlNextRow(sr)) != NULL)
slAddHead(&pars, bedLoadN(row, 4));
sqlFreeResult(&sr);
hFreeConn(&conn);
return pars;
}
void loadBed8(struct track *tg)
/* Convert bed 8 info in window to linked feature. */
{
struct bed *bed;
struct linkedFeatures *lfList = NULL, *lf;
struct trackDb *tdb = tg->tdb;
int scoreMin = atoi(trackDbSettingClosestToHomeOrDefault(tdb, "scoreMin", "0"));
int scoreMax = atoi(trackDbSettingClosestToHomeOrDefault(tdb, "scoreMax", "1000"));
boolean useItemRgb = FALSE;
useItemRgb = bedItemRgb(tdb);
if (tg->isBigBed)
{ // avoid opening an unneeded db connection for bigBed; required not to use mysql for parallel fetch tracks
bigBedAddLinkedFeaturesFrom(tg, chromName, winStart, winEnd,
scoreMin, scoreMax, useItemRgb, 8, &lfList);
}
else
{
struct sqlConnection *conn = hAllocConn(database);
struct sqlResult *sr;
char **row;
int rowOffset;
char *scoreFilterClause = getScoreFilterClause(cart, tg->tdb,NULL);
if (scoreFilterClause != NULL)
{
sr = hRangeQuery(conn, tg->table, chromName, winStart, winEnd,scoreFilterClause, &rowOffset);
freeMem(scoreFilterClause);
}
else
{
sr = hRangeQuery(conn, tg->table, chromName, winStart, winEnd,
NULL, &rowOffset);
}
while ((row = sqlNextRow(sr)) != NULL)
{
bed = bedLoadN(row+rowOffset, 8);
lf = bedMungToLinkedFeatures(&bed, tdb, 8, scoreMin, scoreMax, useItemRgb);
slAddHead(&lfList, lf);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
}
slReverse(&lfList);
slSort(&lfList, linkedFeaturesCmp);
tg->items = lfList;
}
void loadWiki(struct track *tg)
/* Load the items in one custom track - just move beds in
* window... */
{
struct bed *bed, *list = NULL;
struct sqlConnection *conn = hAllocConn(database);
struct sqlResult *sr;
char **row;
int rowOffset;
sr = hRangeQuery(conn, tg->table, chromName, winStart, winEnd, NULL, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
bed = bedLoadN(row+rowOffset, 6);
freeMem(bed->name);
bed->name = malloc(MAX_WIKI_NAME);
snprintf(bed->name, MAX_WIKI_NAME, "%d-%d:%s", bed->chromStart, bed->chromEnd, bed->strand);
slAddHead(&list, bed);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
//slReverse(&list);
// add dummy links
AllocVar(bed);
bed->chrom = chromName;
bed->chromStart = winStart;
bed->chromEnd = winEnd;
bed->name = "Make new entry:+";
bed->strand[0] = '+';
bed->score = 100;
slAddHead(&list, bed);
AllocVar(bed);
bed->chrom = chromName;
bed->chromStart = winStart;
bed->chromEnd = winEnd;
bed->name = "Make new entry:-";
bed->strand[0] = '-';
bed->score = 100;
slAddHead(&list, bed);
tg->items = list;
}
static void wikiTrackFilteredBedOnRegion(
struct region *region, /* Region to get data from. */
struct hash *idHash, /* Hash of identifiers or NULL */
struct bedFilter *bf, /* Filter or NULL */
struct lm *lm, /* Local memory pool. */
struct bed **pBedList /* Output get's appended to this list */
)
/* Get the wikiTrack items passing filter on a single region. */
{
struct bed *bed;
int fieldCount = 6;
char query[512];
int rowOffset;
char **row;
struct sqlConnection *wikiConn = wikiConnect();
struct sqlResult *sr = NULL;
char where[512];
char *filter = filterClause(wikiDbName(), WIKI_TRACK_TABLE, region->chrom, NULL);
if (filter)
sqlSafefFrag(where, sizeof(where), "db='%s' AND %-s", database, filter);
else
sqlSafefFrag(where, sizeof(where), "db='%s'", database);
sqlSafef(query, sizeof(query), "select * from %s", WIKI_TRACK_TABLE);
sr = hRangeQuery(wikiConn, WIKI_TRACK_TABLE, region->chrom,
region->start, region->end, where, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
bed = bedLoadN(row+rowOffset, fieldCount);
if ((idHash == NULL || hashLookup(idHash, bed->name)) &&
(bf == NULL || bedFilterOne(bf, bed)))
{
struct bed *copy = lmCloneBed(bed, lm);
slAddHead(pBedList, copy);
}
}
sqlFreeResult(&sr);
wikiDisconnect(&wikiConn);
}
static struct bed *loadMsBed(struct trackDb *tdb, char *table, char *chrom, uint start, uint end)
/* load every thing from a bed 15 table in the given range */
{
struct sqlConnection *conn = hAllocConnTrack(database, tdb);
struct sqlResult *sr;
char **row;
int rowOffset;
struct bed *bedList = NULL, *bed;
sr = hRangeQuery(conn, table, chrom, start, end, NULL, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
bed = bedLoadN(row+rowOffset, 15);
slAddHead(&bedList, bed);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
slReverse(&bedList);
return bedList;
}
struct genomeRangeTree *grtFromOpenBed(struct lineFile *lf, int size, boolean doPromoter)
/* Read an open bed file into a genomeRangeTree and return it. */
{
struct genomeRangeTree *grt = genomeRangeTreeNew();
char *row[size];
while (lineFileRow(lf, row))
{
struct bed *bed = bedLoadN(row, size);
if (doPromoter)
{
if (bed->strand[0] == '+')
genomeRangeTreeAdd(grt, bed->chrom, bed->chromStart - 500, bed->chromEnd + 500);
else
genomeRangeTreeAdd(grt, bed->chrom, bed->chromEnd - 500, bed->chromEnd + 500);
}
else
genomeRangeTreeAdd(grt, bed->chrom, bed->chromStart, bed->chromEnd);
}
return grt;
}
static struct bed * loadMsBedAll(char *table)
/* load every thing from a bed 15 table */
{
struct sqlConnection *conn = hAllocConn(database);
struct sqlResult *sr;
char **row;
struct bed *bedList = NULL, *bed;
char query[512];
sprintf(query, "select * from %s", table);
sr = sqlGetResult(conn, query);
while ((row = sqlNextRow(sr)) != NULL)
{
bed = bedLoadN(row, 15);
slAddHead(&bedList, bed);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
slReverse(&bedList);
return bedList;
}
struct bed *loadBed(char *database, char *chrom, char *track, int bedN, struct binKeeper *bk)
/* Load in info from a bed track to bk. */
{
struct sqlConnection *conn = hAllocConn(database);
struct sqlResult *sr;
char **row;
int rowOffset;
struct bed *list = NULL, *el;
sr = hChromQuery(conn, track, chrom, NULL, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
el = bedLoadN(row + rowOffset, bedN);
binKeeperAdd(bk, el->chromStart, el->chromEnd, el);
slAddHead(&list, el);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
slReverse(&list);
return list;
}
void loadBed6(struct track *tg)
/* Load the items in one custom track - just move beds in
* window... */
{
struct bed *bed, *list = NULL;
struct sqlConnection *conn = hAllocConn(database);
struct sqlResult *sr;
char **row;
int rowOffset;
sr = hRangeQuery(conn, tg->table, chromName, winStart, winEnd, NULL, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
bed = bedLoadN(row+rowOffset, 6);
slAddHead(&list, bed);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
slReverse(&list);
tg->items = list;
}
void bedMergeOverlappingBlocks(char *inBed, char *outBed)
/* bedMergeOverlappingBlocks - Fix faulty BED 12 files with illegal overlapping blocks. Also reports a summary of the changes.. */
{
int badBeds = 0;
FILE *log = NULL;
FILE *newBedFile = mustOpen(outBed, "w");
char *logName = optionVal("report", NULL);
struct lineFile *lf = lineFileOpen(inBed, TRUE);
char *line, *row[12];
boolean isItemRgb = FALSE;
if (logName)
log = mustOpen(logName, "w");
while (lineFileNext(lf, &line, NULL))
{
struct bed *bed;
int numFields = chopByWhite(line, row, ArraySize(row));
/* strange it's reading empty lines... whatever */
if (numFields == 0)
continue;
if (numFields < 12)
errAbort("file %s doesn't appear to be in blocked-bed format. At least 12 fields required, got %d", inBed, numFields);
if (bedParseRgb(row[8]))
isItemRgb = TRUE;
bed = bedLoadN(row, numFields);
badBeds += fixBed(bed, lf->lineIx, log);
if (isItemRgb)
bedTabOutNitemRgb(bed, numFields, newBedFile);
else
bedTabOutN(bed, numFields, newBedFile);
}
lineFileClose(&lf);
if (log)
{
fprintf(log, "Fixed %d bad beds in all.\n", badBeds);
carefulClose(&log);
}
carefulClose(&newBedFile);
}
void loadOperon(struct track *tg)
/* Load the items in one custom track - just move beds in
* window... */
{
struct linkedFeatures *lfList = NULL, *lf;
struct bed *bed, *list = NULL;
struct sqlConnection *conn = hAllocConn(database);
struct sqlResult *sr;
char **row;
int rowOffset;
sr = hRangeQuery(conn, tg->table, chromName, winStart, winEnd, NULL, &rowOffset);
while ((row = sqlNextRow(sr)) != NULL)
{
bed = bedLoadN(row+rowOffset, 15);
slAddHead(&list, bed);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
slReverse(&list);
for (bed = list; bed != NULL; bed = bed->next)
{
struct simpleFeature *sf;
int i;
lf = lfFromBed(bed);
sf = lf->components;
for (i = 0; i < bed->expCount; i++)
{
if (sf == NULL)
break;
sf->grayIx = grayInRange((int)(bed->expScores[i]),0,1000);
sf = sf->next;
}
slAddHead(&lfList,lf);
}
tg->items = lfList;
}
static void addFilteredBedsOnRegion(struct bbiFile *bbi, struct region *region,
char *table, struct asFilter *filter, struct lm *bedLm, struct bed **pBedList)
/* Add relevant beds in reverse order to pBedList */
{
struct lm *bbLm = lmInit(0);
struct bigBedInterval *ivList = NULL, *iv;
ivList = bigBedIntervalQuery(bbi, region->chrom, region->start, region->end, 0, bbLm);
char *row[bbi->fieldCount];
char startBuf[16], endBuf[16];
for (iv = ivList; iv != NULL; iv = iv->next)
{
bigBedIntervalToRow(iv, region->chrom, startBuf, endBuf, row, bbi->fieldCount);
if (asFilterOnRow(filter, row))
{
struct bed *bed = bedLoadN(row, bbi->definedFieldCount);
struct bed *lmBed = lmCloneBed(bed, bedLm);
slAddHead(pBedList, lmBed);
bedFree(&bed);
}
}
lmCleanup(&bbLm);
}
static void bigBedClick(char *fileName, struct trackDb *tdb,
char *item, int start, int end, int bedSize)
/* Handle click in generic bigBed track. */
{
boolean showUrl = FALSE;
char *chrom = cartString(cart, "c");
/* Open BigWig file and get interval list. */
struct bbiFile *bbi = bigBedFileOpen(fileName);
struct lm *lm = lmInit(0);
int ivStart = start, ivEnd = end;
if (start == end)
{
// item is an insertion; expand the search range from 0 bases to 2 so we catch it:
ivStart = max(0, start-1);
ivEnd++;
}
struct bigBedInterval *bbList = bigBedIntervalQuery(bbi, chrom, ivStart, ivEnd, 0, lm);
/* Get bedSize if it's not already defined. */
if (bedSize == 0)
{
bedSize = bbi->definedFieldCount;
showUrl = TRUE;
}
char *scoreFilter = cartOrTdbString(cart, tdb, "scoreFilter", NULL);
int minScore = 0;
if (scoreFilter)
minScore = atoi(scoreFilter);
/* Find particular item in list - matching start, and item if possible. */
boolean found = FALSE;
boolean firstTime = TRUE;
struct bigBedInterval *bb;
for (bb = bbList; bb != NULL; bb = bb->next)
{
if (!(bb->start == start && bb->end == end))
continue;
if (bedSize > 3)
{
char *name = cloneFirstWordByTab(bb->rest);
boolean match = sameString(name, item);
freez(&name);
if (!match)
continue;
}
found = TRUE;
if (firstTime)
printf("<BR>\n");
int seq1Seq2Fields = 0;
// check for seq1 and seq2 in columns 7+8 (eg, pairedTagAlign)
boolean seq1Seq2 = sameOk(trackDbSetting(tdb, BASE_COLOR_USE_SEQUENCE), "seq1Seq2");
if (seq1Seq2 && bedSize == 6)
seq1Seq2Fields = 2;
char *fields[bedSize+seq1Seq2Fields];
char startBuf[16], endBuf[16];
char *rest = cloneString(bb->rest);
int bbFieldCount = bigBedIntervalToRow(bb, chrom, startBuf, endBuf, fields,
bedSize+seq1Seq2Fields);
if (bbFieldCount != bedSize+seq1Seq2Fields)
{
errAbort("Disagreement between trackDb field count (%d) and %s fieldCount (%d)",
bedSize, fileName, bbFieldCount);
}
struct bed *bed = bedLoadN(fields, bedSize);
if (bedSize >= 6 && scoreFilter && bed->score < minScore)
continue;
if (showUrl && (bedSize >= 4))
printCustomUrl(tdb, item, TRUE);
bedPrintPos(bed, bedSize, tdb);
// display seq1 and seq2
if (seq1Seq2 && bedSize+seq1Seq2Fields == 8)
printf("<table><tr><th>Sequence 1</th><th>Sequence 2</th></tr>"
"<tr><td> %s </td><td> %s </td></tr></table>", fields[6], fields[7]);
else if (isNotEmpty(rest))
{
char *restFields[256];
int restCount = chopTabs(rest, restFields);
int restBedFields = bedSize - 3;
if (restCount > restBedFields)
{
if (0 == extraFieldsPrint(tdb,NULL,restFields + restBedFields,restCount - restBedFields))
{
int i;
char label[20];
safef(label, sizeof(label), "nonBedFieldsLabel");
printf("<B>%s </B>",
trackDbSettingOrDefault(tdb, label, "Non-BED fields:"));
for (i = restBedFields; i < restCount; i++)
printf("%s%s", (i > 0 ? "\t" : ""), restFields[i]);
printf("<BR>\n");
}
}
}
if (isCustomTrack(tdb->track))
{
time_t timep = bbiUpdateTime(bbi);
printBbiUpdateTime(&timep);
}
}
if (!found)
{
printf("No item %s starting at %d\n", emptyForNull(item), start);
}
lmCleanup(&lm);
bbiFileClose(&bbi);
}
/* bedGraphLoadItems - an ordinary bed load, but we are interested
* in only the chrom, start, end, and the graphColumn
*/
static void bedGraphLoadItems(struct track *tg)
{
struct sqlConnection *conn;
struct sqlResult *sr = (struct sqlResult *) NULL;
char **row = (char **)NULL;
int rowOffset = 0;
struct bedGraphItem *bgList = NULL;
int itemsLoaded = 0;
int colCount = 0;
struct wigCartOptions *wigCart = (struct wigCartOptions *) tg->wigCartData;
int graphColumn = 5;
char *tableName;
if(sameString(tg->table, "affyTranscription"))
wigCart->colorTrack = "affyTransfrags";
graphColumn = wigCart->graphColumn;
#ifndef GBROWSE
if (isCustomTrack(tg->table) && tg->customPt)
{
struct customTrack *ct = (struct customTrack *) tg->customPt;
tableName = ct->dbTableName;
conn = hAllocConn(CUSTOM_TRASH);
}
else
#endif /* GBROWSE */
{
tableName = tg->table;
conn = hAllocConnTrack(database, tg->tdb);
}
sr = hRangeQuery(conn, tableName, chromName, winStart, winEnd, NULL,
&rowOffset);
colCount = sqlCountColumns(sr) - rowOffset;
/* Must have at least four good columns */
if (colCount < 4)
errAbort("bedGraphLoadItems: table %s only has %d data columns, must be at least 4", tableName, colCount);
if (colCount < graphColumn)
errAbort("bedGraphLoadItems: table %s only has %d data columns, specified graph column %d does not exist",
tableName, colCount, graphColumn);
/* before loop, determine actual row[graphColumn] index */
graphColumn += (rowOffset - 1);
while ((row = sqlNextRow(sr)) != NULL)
{
struct bedGraphItem *bg;
struct bed *bed;
++itemsLoaded;
/* load chrom, start, end */
bed = bedLoadN(row+rowOffset, 3);
AllocVar(bg);
bg->start = bed->chromStart;
bg->end = bed->chromEnd;
if ((colCount > 4) && ((graphColumn + rowOffset) != 4))
bg->name = cloneString(row[3+rowOffset]);
else
{
char name[128];
safef(name,ArraySize(name),"%s.%d", bed->chrom, itemsLoaded);
bg->name = cloneString(name);
}
bg->dataValue = sqlFloat(row[graphColumn]);
/* filled in by DrawItems */
bg->graphUpperLimit = wigEncodeStartingUpperLimit;
bg->graphLowerLimit = wigEncodeStartingLowerLimit;
slAddHead(&bgList, bg);
bedFree(&bed);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
slReverse(&bgList);
tg->items = bgList;
} /* bedGraphLoadItems() */
void bedItemOverlapCount(struct hash *chromHash, char *infile, char *outfile){
unsigned maxChromSize = 0;
unitSize *counts = (unitSize *)NULL;
FILE *f = mustOpen(outfile, "w");
struct hashCookie hc = hashFirst(chromHash);
struct hashEl *hel;
while( (hel = hashNext(&hc)) != NULL) {
unsigned num = (unsigned) ptToInt(hel->val);
maxChromSize = max(num, maxChromSize);
}
verbose(2,"#\tmaxChromSize: %u\n", maxChromSize);
if (maxChromSize < 1)
errAbort("maxChromSize is zero ?");
/* Allocate just once for the largest chrom and reuse this array */
counts = needHugeMem(sizeof(unitSize) * maxChromSize);
/* Reset the array to be zero to be reused */
memset((void *)counts, 0, sizeof(unitSize)*(size_t)maxChromSize);
unsigned chromSize = 0;
char *prevChrom = (char *)NULL;
boolean outputToDo = FALSE;
struct hash *seenHash = newHash(5);
struct lineFile *bf = lineFileOpen(infile , TRUE);
struct bed *bed = (struct bed *)NULL;
char *row[12];
int numFields = doBed12 ? 12 : 3;
while (lineFileNextRow(bf,row, numFields))
{
int i;
bed = bedLoadN(row, numFields);
verbose(3,"#\t%s\t%d\t%d\n",bed->chrom,bed->chromStart, bed->chromEnd);
if (prevChrom && differentWord(bed->chrom,prevChrom)) // End a chr
{
verbose(2,"#\tchrom %s done, size %d\n", prevChrom, chromSize);
if (outputToDo)
outputCounts(counts, prevChrom, chromSize, f);
outputToDo = FALSE;
memset((void *)counts, 0,
sizeof(unitSize)*(size_t)maxChromSize); /* zero counts */
freez(&prevChrom);
// prevChrom is now NULL so it will be caught by next if!
}
if ((char *)NULL == prevChrom) // begin a chr
{
if (hashLookup(seenHash, bed->chrom))
errAbort("ERROR:input file not sorted. %s seen before on line %d\n",
bed->chrom, bf->lineIx);
hashAdd(seenHash, bed->chrom, NULL);
prevChrom = cloneString(bed->chrom);
chromSize = hashIntVal(chromHash, prevChrom);
verbose(2,"#\tchrom %s starting, size %d\n", prevChrom,chromSize);
}
if (bed->chromEnd > chromSize)
{
// check for circular chrM
if (doBed12 || bed->chromStart>=chromSize
|| differentWord(bed->chrom,"chrM"))
{
warn("ERROR: %s\t%d\t%d", bed->chrom, bed->chromStart,
bed->chromEnd);
errAbort("chromEnd > chromSize ? %d > %d",
bed->chromEnd,chromSize);
}
for (i = bed->chromStart; i < chromSize; ++i)
INCWOVERFLOW(counts,i);
for (i = 0; i < (bed->chromEnd - chromSize); ++i)
INCWOVERFLOW(counts,i);
}
else if (doBed12)
{
int *starts = bed->chromStarts;
int *sizes = bed->blockSizes;
int *endStarts = &bed->chromStarts[bed->blockCount];
for(; starts < endStarts; starts++, sizes++)
{
unsigned int end = *starts + *sizes + bed->chromStart;
for (i = *starts + bed->chromStart; i < end; ++i)
INCWOVERFLOW(counts,i);
}
}
else
{
for (i = bed->chromStart; i < bed->chromEnd; ++i)
INCWOVERFLOW(counts, i);
}
outputToDo = TRUE;
bedFree(&bed); // plug the memory leak
}
lineFileClose(&bf);
// Note, next file could be on same chr!
if (outputToDo)
outputCounts(counts, prevChrom, chromSize, f);
if (doOutBounds)
fprintf(stderr, "min %lu max %lu\n", (unsigned long)overMin, (unsigned long)overMax);
verbose(2,"#\tchrom %s done, size %d\n", prevChrom, chromSize);
carefulClose(&f);
freeMem(counts);
freez(&prevChrom);
// hashFreeWithVals(&chromHash, freez);
freeHash(&seenHash);
}
