void mapAltGraphXFile(struct sqlConnection *conn, char *db, char *orthoDb, char *chrom,
char *netTable, char *altGraphXFileName, char *altGraphXTableName,
FILE *agxOut, FILE *selectedOut, int *foundCount, int *notFoundCount)
/* Map over altGraphX Structures from one organism to
another. Basically create a mapping for the vertices and then reverse
them if on '-' strand.*/
{
int count =0;
struct bed *bed = NULL;
struct altGraphX *agList = NULL, *ag = NULL, *agNew = NULL;
if(altGraphXFileName != NULL)
{
warn("Loading altGraphX Records from file %s.", altGraphXFileName);
agList = altGraphXLoadAll(altGraphXFileName);
}
else if(altGraphXTableName != NULL)
{
char query[256];
warn("Reading altGraphX Records from table %s.", altGraphXTableName);
sqlSafef(query, sizeof(query), "select * from %s where tName like '%s'", altGraphXTableName, chrom);
agList = altGraphXLoadByQuery(conn, query);
}
else
errAbort("orthoMap::mapAlGraphXFile() - Need a table name or file name to load altGraphX records");
warn("Mapping altGraphX records.");
for(ag = agList; ag != NULL; ag = ag->next)
{
if(differentString(ag->tName, chrom))
continue;
occassionalDot();
agNew = mapAltGraphX(ag, conn, db, netTable);
if(agNew == NULL)
(*notFoundCount)++;
else
{
(*foundCount)++;
altGraphXTabOut(agNew, agxOut);
altGraphXFree(&agNew);
if (selectedOut != NULL)
altGraphXTabOut(ag, selectedOut);
}
count++;
}
}
struct altGraphX *agFromAlignments(char *db, struct ggMrnaAli *maList, struct dnaSeq *seq, struct sqlConnection *conn,
int chromStart, int chromEnd, FILE *out )
/** Custer overlaps from maList into altGraphX structure. */
{
struct altGraphX *ag = NULL, *agList = NULL;
struct ggMrnaCluster *mcList=NULL, *mc=NULL;
struct ggMrnaInput *ci = NULL;
struct geneGraph *gg = NULL;
static int count = 0;
ci = ggMrnaInputFromAlignments(maList, seq);
mcList = ggClusterMrna(ci);
if(mcList == NULL)
{
freeGgMrnaInput(&ci);
return NULL;
}
clusterCount++;
for(mc = mcList; mc != NULL; mc = mc->next)
{
if(optionExists("consensus"))
{
gg = ggGraphConsensusCluster(db, mc, ci, tissLibHash, !optionExists("skipTissues"));
}
else
gg = ggGraphCluster(db, mc,ci);
assert(checkEvidenceMatrix(gg));
ag = ggToAltGraphX(gg);
if(ag != NULL)
{
char name[256];
freez(&ag->name);
safef(name, sizeof(name), "%s.%d", ag->tName, count++);
ag->name = cloneString(name);
/* Convert back to genomic coordinates. */
altGraphXoffset(ag, chromStart);
/* Sort vertices so that they are chromosomal order */
altGraphXVertPosSort(ag);
/* write to file */
binKeeperAdd(agxSeenBin, ag->tStart, ag->tEnd, ag);
slAddHead(&agList, ag);
}
}
/* Sometimes get nested, partial transcripts. Want to filter
those out. */
for(ag = agList; ag != NULL; ag = ag->next)
{
if(!agxIsRedundant(ag))
altGraphXTabOut(ag, out);
}
/* genoSeq and maList are freed with ci and gg */
ggFreeMrnaClusterList(&mcList);
freeGgMrnaInput(&ci);
freeGeneGraph(&gg);
return agList;
}
void txgToAgx(char *inTxg, char *outAgx)
/* txgToAgx - Convert from txg (txGraph) format to agx (altGraphX). */
{
struct lineFile *lf = lineFileOpen(inTxg, TRUE);
char *row[TXGRAPH_NUM_COLS];
FILE *f = mustOpen(outAgx, "w");
while (lineFileRow(lf, row))
{
struct txGraph *txg = txGraphLoad(row);
verbose(2, "loaded txGraph %s\n", txg->name);
struct altGraphX *agx = txGraphToAltGraphX(txg);
altGraphXTabOut(agx, f);
altGraphXFree(&agx);
txGraphFree(&txg);
}
carefulClose(&f);
}
void writeCassetteExon(struct bed *bedList, struct altGraphX *ag, int eIx, boolean *outputted,
FILE *bedOutFile, FILE *outfile, FILE *html, float conf )
/* Write out the information for a cassette exon. */
{
int i = eIx;
struct bed *bed=NULL;
if(bedOutFile != NULL)
bedTabOutN(bedList,12, bedOutFile);
writeBrowserLink(html, ag, conf, i);
if(!outputted)
{
altGraphXTabOut(ag, stdout);
*outputted = TRUE;
}
if(outfile != NULL)
{
struct dnaSeq *seq = hChromSeq(ag->tName, ag->vPositions[ag->edgeStarts[i]], ag->vPositions[ag->edgeEnds[i]]);
if(sameString(ag->strand , "+"))
reverseComplement(seq->dna, seq->size);
if(seq->size < 200)
faWriteNext(outfile, seq->name, seq->dna, seq->size);
freeDnaSeq(&seq);
}
}
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);
}
