/* create object on the heap, copy values from psl and return, don't forget to free */
struct psl *copyPsl(struct psl *psl)
{
struct psl *ret = NULL;
AllocVar(ret);
ret->next = NULL;
ret->match = psl->match;
ret->misMatch = psl->misMatch;
ret->repMatch = psl->repMatch;
ret->nCount = psl->nCount;
ret->qNumInsert = psl->qNumInsert;
ret->qBaseInsert = psl->qBaseInsert;
ret->tNumInsert = psl->tNumInsert;
ret->tBaseInsert = psl ->tBaseInsert;
strcpy(ret->strand, psl->strand);
ret->qName = cloneString(psl->qName);
ret->qSize = psl->qSize;
ret->qStart = psl->qStart;
ret->qEnd = psl->qEnd;
ret->tName = cloneString(psl->tName);
ret->tSize = psl->tSize;
ret->tStart = psl->tStart;
ret->tEnd = psl->tEnd;
ret->blockCount = psl->blockCount;
ret->blockSizes = CloneArray(psl->blockSizes, psl->blockCount);
ret->qStarts = CloneArray(psl->qStarts, psl->blockCount);
ret->tStarts = CloneArray(psl->tStarts, psl->blockCount);
return ret;
}
CHqlExprMultiGuard(const CHqlExprMultiGuard * other)
{
if (other)
{
CloneArray(guarded, other->guarded);
}
}
struct wigSection *wigSectionRead(struct lineFile *lf)
/* Parse out next section of wig. */
{
static double *vals = NULL;
static int valAlloc = 0;
/* Get "fixedStep" line and parse it. */
char *line;
if (!lineFileNextReal(lf, &line))
return NULL;
char *pattern = "fixedStep ";
int patSize = 10;
if (!startsWith(pattern, line))
errAbort("Expecting fixedStep line %d of %s", lf->lineIx, lf->fileName);
line += patSize;
struct hash *varHash = hashVarLine(line, lf->lineIx);
int step = sqlUnsigned(requiredVal(lf, varHash, "step"));
int start = sqlUnsigned(requiredVal(lf, varHash, "start"));
char *chrom = cloneString(requiredVal(lf, varHash, "chrom"));
hashFree(&varHash);
/* Parse out numbers until next fixedStep. */
int valCount = 0;
int i;
for (;;)
{
if (!lineFileNextReal(lf, &line))
break;
if (startsWith(pattern, line))
{
lineFileReuse(lf);
break;
}
for (i=0; i<step; ++i)
{
if (valCount >= valAlloc)
{
int newAlloc = valAlloc + 1024;
ExpandArray(vals, valAlloc, newAlloc);
valAlloc = newAlloc;
}
vals[valCount] = lineFileNeedDouble(lf, &line, 0);
++valCount;
}
}
/* Create wigSection. */
struct wigSection *section;
AllocVar(section);
section->chrom = chrom;
section->chromStart = start;
section->chromEnd = start + valCount;
section->vals = CloneArray(vals, valCount);
return section;
}
void maxTranscriptomeExps(char *files[], int numFiles, char *outputFile, char *trackName)
{
struct sample **pSampList = NULL;
struct sample *s1;
struct sample *maxListSamp = NULL;
struct sample *maxSamp = NULL;
int i;
int count =0;
FILE *out = NULL;
AllocArray(pSampList, numFiles);
for(i=0;i<numFiles; i++)
{
warn("Reading %s.", files[i]);
pSampList[i] = sampleLoadAll(files[i]);
}
warn("Calculating Maxes.");
count = slCount(pSampList[0]);
for(i=0;i<count;i++)
{
AllocVar(maxSamp);
s1 = slElementFromIx(pSampList[0], i);
maxSamp->chrom = cloneString(s1->chrom);
maxSamp->chromStart = s1->chromStart;
maxSamp->chromEnd = s1->chromEnd;
maxSamp->name = cloneString(trackName);
snprintf(maxSamp->strand, sizeof(maxSamp->strand), "%s", s1->strand);
maxSamp->sampleCount = s1->sampleCount;
maxSamp->samplePosition = CloneArray(s1->samplePosition, maxSamp->sampleCount);
AllocArray(maxSamp->sampleHeight, maxSamp->sampleCount);
fillInMaxVals(maxSamp, pSampList, numFiles, i);
slAddHead(&maxListSamp, maxSamp);
}
slReverse(&maxListSamp);
warn("Saving Maxes");
out = mustOpen(outputFile, "w");
for(maxSamp = maxListSamp; maxSamp != NULL; maxSamp = maxSamp->next)
{
sampleTabOut(maxSamp, out);
}
carefulClose(&out);
warn("Cleaning up");
sampleFreeList(&maxListSamp);
for(i=0;i<numFiles; i++)
sampleFreeList(&pSampList[i]);
freez(&pSampList);
warn("Done.");
}
struct improbRunInfo * analyseOneMotifRun(char *runName, char *seqDir,
char *motifDir, int controlCount, char *controls[])
/* Bundle up data on one improbizer run and associated control runs. */
{
char fileName[512];
char motifName[256];
int seqCount, baseCount;
struct improbRunInfo *iriList = NULL, *iri;
struct lineFile *lf = NULL;
struct motif motif;
int motifIx = 0;
int i;
float acc, best, mean, x;
printf("%s\n", runName);
/* Count bases in sequences - this will be used in each iri. */
sprintf(fileName, "%s/%s.fa", seqDir, runName);
countSeq(fileName, &seqCount, &baseCount);
/* Allocate iri and read the main run. */
sprintf(fileName, "%s/%s", motifDir, runName);
lf = lineFileOpen(fileName, TRUE);
while (readMotif(lf, &motif))
{
AllocVar(iri);
slAddTail(&iriList, iri);
++motifIx;
snprintf(motifName, sizeof(motifName), "%s.%d", runName, motifIx);
iri->name = cloneString(motifName);
iri->seqCount = seqCount;
iri->runScore = motif.score;
iri->runPos = motif.pos;
iri->runPosSd = motif.posSd;
iri->columnCount = motif.size;
iri->consensus = cloneString(motif.consensus);
iri->aProb = CloneArray(motif.profile[0], motif.size);
iri->cProb = CloneArray(motif.profile[1], motif.size);
iri->gProb = CloneArray(motif.profile[2], motif.size);
iri->tProb = CloneArray(motif.profile[3], motif.size);
iri->controlCount = controlCount;
AllocArray(iri->controlScores, controlCount);
}
lineFileClose(&lf);
/* Read the control runs. */
for (i=0; i<controlCount; ++i)
{
sprintf(fileName, "%s/%s", controls[i], runName);
lf = lineFileOpen(fileName, TRUE);
for (iri = iriList; iri != NULL; iri = iri->next)
{
if (!readMotif(lf, &motif))
errAbort("%s doesn't contain the expected number of motifs", lf->fileName);
iri->controlScores[i] = motif.score;
}
lineFileClose(&lf);
}
/* Calculate best and mean on control runs. */
for (iri = iriList; iri != NULL; iri = iri->next)
{
acc = best = 0;
for (i=0; i<controlCount; ++i)
{
x = iri->controlScores[i];
acc += x;
if (x > best)
best = x;
}
iri->bestControlScore = best;
iri->meanControlScore = acc/controlCount;
}
/* Calculate standard deviation of control runs. */
for (iri = iriList; iri != NULL; iri = iri->next)
{
acc = 0;
mean = iri->meanControlScore;
for (i=0; i<controlCount; ++i)
{
x = iri->controlScores[i] - mean;
acc += x*x;
}
if (controlCount > 1)
acc /= controlCount;
iri->sdControlScore = sqrt(acc);
}
return iriList;
}
void reportCassette(struct altGraphX *ag, bool **em, int vs, int ve1, int ve2,
int altBpStart, int altBpEnd, int startV, int endV, FILE *out)
/* Write out both an altGraphX and two bed files. For a cassette exon the
edges are -
Name Vertexes Class
------ ---------- -----
exon1: startV->vs constitutive (cons 0)
junction1: vs->ve1 alternative1 (alt1 1)
exon2: ve1->altBpEnd alternative1 (alt1 1)
junction2: altBpEnd->ve2 alternative1 (alt1 1)
exon3: ve2->endV constitutive (cons 0)
junction3: vs->ve2 alternative2 (alt2 2)
*/
{
struct altGraphX *agLoc = NULL; /* Local altGraphX. */
struct evidence *ev = NULL, *evLoc = NULL;
int *vPos = ag->vPositions;
unsigned char *vT = ag->vTypes;
int *vPosLoc = NULL; /* Vertex Positions. */
int *eStartsLoc = NULL; /* Edge Starts. */
int *eEndsLoc = NULL; /* Edge ends. */
unsigned char *vTLoc = NULL; /* Vertex Types. */
int *eTLoc = NULL; /* Edge Types. */
int vCLoc = 0;
int eCLoc = 0;
int i =0;
struct dyString *dy = NULL;
if(out == NULL)
return;
AllocVar(agLoc);
agLoc->tName = cloneString(ag->tName);
agLoc->name = cloneString(ag->name);
agLoc->tStart = vPos[startV];
agLoc->tEnd = vPos[endV];
agLoc->strand[0] = ag->strand[0];
agLoc->vertexCount = vCLoc = 6;
agLoc->edgeCount = eCLoc = 6;
agLoc->id = altCassette;
/* Allocate some arrays. */
AllocArray(vPosLoc, vCLoc);
AllocArray(eStartsLoc, vCLoc);
AllocArray(eEndsLoc, vCLoc);
AllocArray(vTLoc, vCLoc);
AllocArray(eTLoc, vCLoc);
/* Fill in the vertex positions. */
vPosLoc[0] = vPos[startV];
vPosLoc[1] = vPos[vs];
vPosLoc[2] = vPos[ve1];
vPosLoc[3] = vPos[altBpEnd];
vPosLoc[4] = vPos[ve2];
vPosLoc[5] = vPos[endV];
/* Fill in the vertex types. */
vTLoc[0] = vT[startV];
vTLoc[1] = vT[vs];
vTLoc[2] = vT[ve1];
vTLoc[3] = vT[altBpEnd];
vTLoc[4] = vT[ve2];
vTLoc[5] = vT[endV];
/* Fill in the edges. */
/* Constitutive first exon. */
eStartsLoc[0] = 0;
eEndsLoc[0] = 1;
eTLoc[0] = 0;
ev = evidenceForEdge(ag, startV, vs);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
/* Exon inclusion junction. */
eStartsLoc[1] = 1;
eEndsLoc[1] = 2;
eTLoc[1] = 1;
ev = evidenceForEdge(ag, vs, ve1);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
/* Exon exclusion junction. */
eStartsLoc[2] = 1;
eEndsLoc[2] = 4;
eTLoc[2] = 2;
ev = evidenceForEdge(ag, vs, ve2);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
/* Cassette exon. */
eStartsLoc[3] = 2;
eEndsLoc[3] = 3;
eTLoc[3] = 1;
ev = evidenceForEdge(ag, ve1, altBpEnd);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
/* Exon inclusion junction. */
eStartsLoc[4] = 3;
eEndsLoc[4] = 4;
eTLoc[4] = 1;
ev = evidenceForEdge(ag, altBpEnd, ve2);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
/* Constitutive second exon. */
eStartsLoc[5] = 4;
eEndsLoc[5] = 5;
eTLoc[5] = 0;
ev = evidenceForEdge(ag, ve2, endV);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
slReverse(&agLoc->evidence);
dy = newDyString(ag->mrnaRefCount*36);
agLoc->mrnaRefCount = ag->mrnaRefCount;
for(i=0; i<ag->mrnaRefCount; i++)
dyStringPrintf(dy, "%s,", ag->mrnaRefs[i]);
sqlStringDynamicArray(dy->string, &agLoc->mrnaRefs, &i);
dyStringFree(&dy);
agLoc->mrnaTissues = CloneArray(ag->mrnaTissues, ag->mrnaRefCount);
agLoc->mrnaLibs = CloneArray(ag->mrnaLibs, ag->mrnaRefCount);
agLoc->vPositions = vPosLoc;
agLoc->edgeStarts = eStartsLoc;
agLoc->edgeEnds = eEndsLoc;
agLoc->vTypes = vTLoc;
agLoc->edgeTypes = eTLoc;
altGraphXTabOut(agLoc, out);
altGraphXFree(&agLoc);
}
void reportAlt3Prime(struct altGraphX *ag, bool **em, int vs, int ve1, int ve2,
int altBpStart, int altBpEnd, int startV, int endV, FILE *out)
/* Write out an altGraphX record for an alt3Prime splicing
event. Variable names are consistent with the rest of the program, but
can be misleading. Specifically vs = start of alt splicing, ve1 =
first end of alt splicing, etc. even though "vs" is really the end of
an exon. For an alt5Prime splice the edges are:
Name Vertexes Class
------ ---------- -----
exon1: startV->vs constituative (0)
junction1: vs->ve1 alternative (1)
junction2: vs->ve2 alternative (2)
exon2: ve1->e2 alternative (1)
exon3: ve2->endV constituative (0)
*/
{
struct altGraphX *agLoc = NULL; /* Local altGraphX. */
struct evidence *ev = NULL, *evLoc = NULL;
int *vPos = ag->vPositions;
unsigned char *vT = ag->vTypes;
int *vPosLoc = NULL; /* Vertex Positions. */
int *eStartsLoc = NULL; /* Edge Starts. */
int *eEndsLoc = NULL; /* Edge ends. */
unsigned char *vTLoc = NULL; /* Vertex Types. */
int *eTLoc = NULL; /* Edge Types. */
int vCLoc = 0;
int eCLoc = 0;
int edgeIx = 0, vertexIx = 0;
int i =0;
struct dyString *dy = NULL;
if(out == NULL)
return;
AllocVar(agLoc);
agLoc->tName = cloneString(ag->tName);
agLoc->name = cloneString(ag->name);
agLoc->tStart = vPos[startV];
agLoc->tEnd = vPos[endV];
agLoc->strand[0] = ag->strand[0];
agLoc->vertexCount = vCLoc = 6;
agLoc->edgeCount = eCLoc = 5;
agLoc->id = alt3Prime;
/* Allocate some arrays. */
AllocArray(vPosLoc, vCLoc);
AllocArray(eStartsLoc, eCLoc);
AllocArray(eEndsLoc, eCLoc);
AllocArray(vTLoc, vCLoc);
AllocArray(eTLoc, eCLoc);
/* Fill in the vertex positions. */
vertexIx = 0;
vPosLoc[vertexIx++] = vPos[startV]; /* 0 */
vPosLoc[vertexIx++] = vPos[vs]; /* 1 */
vPosLoc[vertexIx++] = vPos[ve1]; /* 2 */
vPosLoc[vertexIx++] = vPos[ve2]; /* 3 */
vPosLoc[vertexIx++] = vPos[ve2]; /* 4 */
vPosLoc[vertexIx++] = vPos[endV]; /* 5 */
/* Fill in the vertex types. */
vertexIx = 0;
vTLoc[vertexIx++] = vT[startV];
vTLoc[vertexIx++] = vT[vs];
vTLoc[vertexIx++] = vT[ve1];
vTLoc[vertexIx++] = vT[vs]; /* Faking a separate exon for the alt spliced portion. */
vTLoc[vertexIx++] = vT[ve2];
vTLoc[vertexIx++] = vT[endV];
edgeIx = 0;
/* Constitutive first exon. */
eStartsLoc[edgeIx] = 0;
eEndsLoc[edgeIx] = 1;
eTLoc[edgeIx] = 0;
ev = evidenceForEdge(ag, startV, vs);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
edgeIx++;
/* Alternative1 junction (shorter). */
eStartsLoc[edgeIx] = 1;
eEndsLoc[edgeIx] = 2;
eTLoc[edgeIx] = 1;
ev = evidenceForEdge(ag, vs, ve1);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
edgeIx++;
/* Alt2 junction (longer). */
eStartsLoc[edgeIx] = 1;
eEndsLoc[edgeIx] = 4;
eTLoc[edgeIx] = 2;
ev = evidenceForEdge(ag, vs, ve2);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
edgeIx++;
/* Alt1 portion of second exon. */
eStartsLoc[edgeIx] = 2;
eEndsLoc[edgeIx] = 3;
eTLoc[edgeIx] = 1;
ev = evidenceForEdge(ag, ve1, endV);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
edgeIx++;
/* Exon 2 constitutive (shorter exon) */
eStartsLoc[edgeIx] = 4;
eEndsLoc[edgeIx] = 5;
eTLoc[edgeIx] = 0;
ev = evidenceForEdge(ag, ve2, endV);
evLoc = CloneVar(ev);
evLoc->mrnaIds = CloneArray(ev->mrnaIds, ev->evCount);
slAddHead(&agLoc->evidence, evLoc);
edgeIx++;
/* Package up the evidence, tissues, etc. */
slReverse(&agLoc->evidence);
dy = newDyString(ag->mrnaRefCount*36);
agLoc->mrnaRefCount = ag->mrnaRefCount;
for(i=0; i<ag->mrnaRefCount; i++)
dyStringPrintf(dy, "%s,", ag->mrnaRefs[i]);
sqlStringDynamicArray(dy->string, &agLoc->mrnaRefs, &i);
dyStringFree(&dy);
agLoc->mrnaTissues = CloneArray(ag->mrnaTissues, ag->mrnaRefCount);
agLoc->mrnaLibs = CloneArray(ag->mrnaLibs, ag->mrnaRefCount);
agLoc->vPositions = vPosLoc;
agLoc->edgeStarts = eStartsLoc;
agLoc->edgeEnds = eEndsLoc;
agLoc->vTypes = vTLoc;
agLoc->edgeTypes = eTLoc;
altGraphXTabOut(agLoc, out);
altGraphXFree(&agLoc);
}