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 doTransRegCode(struct trackDb *tdb, char *item, char *motifTable)
/* Display detailed info on a transcriptional regulatory code item. */
{
struct dnaMotif *motif = loadDnaMotif(item, motifTable);
int start = cartInt(cart, "o");
struct dnaSeq *seq = NULL;
char *table = tdb->table;
int rowOffset = hOffsetPastBin(database, seqName, table);
char query[256];
struct sqlResult *sr;
char **row;
struct sqlConnection *conn = hAllocConn(database);
struct transRegCode *trc = NULL;
cartWebStart(cart, database, "Regulatory Code Info");
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)
trc = transRegCodeLoad(row+rowOffset);
sqlFreeResult(&sr);
if (trc != NULL)
{
char strand[2];
seq = hDnaFromSeq(database, trc->chrom, trc->chromStart, trc->chromEnd, dnaLower);
if (seq->size != motif->columnCount)
{
printf("WARNING: seq->size = %d, motif->colCount=%d<BR>\n",
seq->size, motif->columnCount);
strand[0] = '?';
seq = NULL;
}
else
{
strand[0] = dnaMotifBestStrand(motif, seq->dna);
if (strand[0] == '-')
reverseComplement(seq->dna, seq->size);
}
strand[1] = 0;
printf("<B>Name:</B> ");
sacCerHgGeneLinkName(conn, trc->name);
printf("<BR>\n");
printf("<B>ChIP-chip Evidence:</B> %s<BR>\n", trc->chipEvidence);
printf("<B>Species conserved in:</B> %d of 2<BR>\n", trc->consSpecies);
if (seq != NULL)
printf("<B>Bit Score of Motif Hit:</B> %4.2f<BR>\n",
dnaMotifBitScore(motif, seq->dna));
printf("<B>Item score:</B> %d<BR>\n", trc->score);
printPosOnChrom(trc->chrom, trc->chromStart, trc->chromEnd, strand, TRUE, trc->name);
}
motifHitSection(seq, motif);
printTrackHtml(tdb);
}
struct dnaSeq *dnaFromChrom(char *db, char *chrom, int chromStart, int chromEnd, enum dnaCase seqCase)
/** Return the dna for the chromosome region specified. */
{
struct dnaSeq *seq = NULL;
if(chromNib != NULL)
{
seq = (struct dnaSeq *)nibLdPart(chromNib, chromNibFile, chromNibSize,
chromStart, chromEnd - chromStart);
}
else
seq = hDnaFromSeq(db, chrom, chromStart, chromEnd, seqCase);
return seq;
}
static double motifScoreHere(char *chrom, int start, int end,
char *motifName, char *motifTable)
/* Return score of motif at given position. */
{
double score;
struct dnaSeq *seq = hDnaFromSeq(database, chrom, start, end, dnaLower);
struct dnaMotif *motif = loadDnaMotif(motifName, motifTable);
char strand = dnaMotifBestStrand(motif, seq->dna);
if (strand == '-')
reverseComplement(seq->dna, seq->size);
score = dnaMotifBitScore(motif, seq->dna);
dnaMotifFree(&motif);
dnaSeqFree(&seq);
return score;
}
void chromFeatureSeq(struct sqlConnection *conn,
char *database, char *chrom, char *trackSpec,
FILE *bedFile, FILE *faFile,
int *retItemCount, int *retBaseCount)
/* Write out sequence file for features from one chromosome.
* This separate routine handles the non-merged case. It's
* reason for being is so that the feature names get preserved. */
{
boolean hasBin;
char t[512], *s = NULL;
char table[HDB_MAX_TABLE_STRING];
struct featureBits *fbList = NULL, *fb;
if (trackSpec[0] == '!')
errAbort("Sorry, '!' not available with fa output unless you use faMerge");
isolateTrackPartOfSpec(trackSpec, t);
s = strchr(t, '.');
if (s != NULL)
errAbort("Sorry, only database (not file) tracks allowed with "
"fa output unless you use faMerge");
// ignore isSplit return from hFindSplitTable()
(void) hFindSplitTable(database, chrom, t, table, &hasBin);
fbList = fbGetRangeQuery(database, trackSpec, chrom, 0, hChromSize(database, chrom),
where, TRUE, TRUE);
for (fb = fbList; fb != NULL; fb = fb->next)
{
int s = fb->start, e = fb->end;
if (bedFile != NULL)
{
fprintf(bedFile, "%s\t%d\t%d\t%s",
fb->chrom, fb->start, fb->end, fb->name);
if (fb->strand != '?')
fprintf(bedFile, "\t0\t%c", fb->strand);
fprintf(bedFile, "\n");
}
if (faFile != NULL)
{
struct dnaSeq *seq = hDnaFromSeq(database, chrom, s, e, dnaLower);
if (fb->strand == '-')
reverseComplement(seq->dna, seq->size);
faWriteNext(faFile, fb->name, seq->dna, seq->size);
freeDnaSeq(&seq);
}
}
featureBitsFreeList(&fbList);
}
void runSamples(char *goodFile, char *badFile, char *newDb, char *oldDb, int numToRun)
/* run a bunch of tests */
{
int i,j,k;
FILE *good = mustOpen(goodFile, "w");
FILE *bad = mustOpen(badFile, "w");
char *tmp = NULL;
int numGood=0, numBad=0, tooManyNs=0;
boolean success = FALSE;
struct dnaSeq *seq = NULL;
printf("Running Tests\t");
for(i=0;i<numToRun;i++)
{
struct coordConvRep *ccr = NULL;
struct coordConv *cc = NULL;
if(!(i%10)) putTic();
cc = getRandomCoord(oldDb);
seq = hDnaFromSeq(cc->chrom, cc->chromStart, cc->chromEnd, dnaLower);
if(!(strstr(seq->dna, "nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn")))
{
chrom = cc->chrom;
chromStart = cc->chromStart;
chromEnd = cc->chromEnd;
success = convertCoordinates(good, bad, printReport, printReport);
if(success)
numGood++;
else
numBad++;
}
else
{
tooManyNs++;
}
freeDnaSeq(&seq);
coordConvFree(&cc);
}
carefulClose(&good);
carefulClose(&bad);
printf("\tDone.\n");
printf("Out of %d attempts got %d 'succesfully converted' and %d 'had problems', %d had too many N's\n",
(numGood + numBad), numGood, numBad, tooManyNs);
printf("After checking got %d of %d correctly called and %d incorrectly called.\n",
hgTestCorrect, hgTestCorrect+hgTestWrong, hgTestWrong);
}
void motifFinder(char *database, char *name, int fileCount, char *files[])
/* motifFinder - find largest scoring motif in bed items. */
{
struct sqlConnection *conn = sqlConnect(database);
int fileNum;
char where[256];
struct chromInfo *ci = createChromInfoList(NULL, database);
sqlSafefFrag(where, sizeof(where), "name = '%s'", name);
struct dnaMotif *motif = dnaMotifLoadWhere(conn, motifTable, where);
if(markovTable != NULL)
dnaMotifMakeLog2(motif);
if(motif == NULL)
errAbort("couldn't find motif '%s'", name);
for (fileNum = 0; fileNum < fileCount; fileNum++)
{
char *words[64], *line;
char **row;
struct lineFile *lf = lineFileOpen(files[fileNum], TRUE);
while (lineFileNextReal(lf, &line))
{
int dnaLength, i, j, rowOffset, length, wordCount = chopTabs(line, words);
unsigned chromSize;
boolean markovFound = FALSE;
double mark0[5];
double mark2[5][5][5];
struct dnaSeq *seq = NULL;
char *dupe = NULL;
if (0 == wordCount)
continue;
lineFileExpectAtLeast(lf, 3, wordCount);
dupe = cloneString(line);
char *chrom = words[0];
int chromStart = lineFileNeedNum(lf, words, 1);
if(markovTable != NULL)
chromStart = max(2, chromStart);
unsigned chromEnd = lineFileNeedNum(lf, words, 2);
if (chromEnd < 1)
errAbort("ERROR: line %d:'%s'\nchromEnd is less than 1\n",
lf->lineIx, dupe);
if (chromStart > chromEnd)
errAbort("ERROR: line %d:'%s'\nchromStart after chromEnd (%d > %d)\n",
lf->lineIx, dupe, chromStart, chromEnd);
length = chromEnd - chromStart;
chromSize = getChromSize(ci, chrom);
if(markovTable == NULL)
{
dnaLength = length;
seq = hDnaFromSeq(database, chrom, chromStart, chromEnd, dnaUpper);
if(uniformBackground)
{
int i;
mark0[0] = 1;
for(i = 1; i <= 4; i++)
mark0[i] = 0.25;
}
else
{
dnaMark0(seq, mark0, NULL);
}
}
else
{
dnaLength = length + 4;
if(chromStart - 2 + dnaLength > chromSize)
// can't do analysis for potential peak hanging off the end of the chrom
continue;
seq = hDnaFromSeq(database, chrom, chromStart - 2, chromEnd + 2, dnaUpper);
struct sqlResult *sr = hRangeQuery(conn, markovTable, chrom, chromStart,
chromStart + 1, NULL, &rowOffset);
if((row = sqlNextRow(sr)) != NULL)
{
dnaMark2Deserialize(row[rowOffset + 3], mark2);
dnaMarkMakeLog2(mark2);
markovFound = TRUE;
}
else
errAbort("markov table '%s' is missing; non-markov analysis is current not supported", markovTable);
sqlFreeResult(&sr);
}
struct bed6FloatScore *hits = NULL;
for (i = 0; i < 2; i++)
{
double mark0Copy[5];
char strand = i == 0 ? '+' : '-';
for (j = 0; j <= 4; j++)
mark0Copy[j] = mark0[j];
if(strand == '-')
{
// reverse markov table too!
double tmp;
reverseComplement(seq->dna, dnaLength);
tmp = mark0Copy[1];
mark0Copy[1] = mark0Copy[3];
mark0Copy[3] = tmp;
tmp = mark0Copy[2];
mark0Copy[2] = mark0Copy[4];
mark0Copy[4] = tmp;
}
for (j = 0; j < length - motif->columnCount + 1; j++)
// tricky b/c if(markovFound) then seq includes the two bytes on either side of actual sequence.
{
double score;
if(markovFound)
score = dnaMotifBitScoreWithMarkovBg(motif, seq->dna + j, mark2);
else
score = dnaMotifBitScoreWithMark0Bg(motif, seq->dna + j, mark0Copy);
if(score >= minScoreCutoff)
{
int start;
if(strand == '-')
start = (chromEnd - j) - motif->columnCount;
else
start = chromStart + j;
struct bed6FloatScore *hit = NULL;
// Watch out for overlapping hits (on either strand; yes, I've seen that happen);
// we report only the highest scoring hit in this case.
// O(n^2) where n == number of motifs in a peak, but I expect n to be almost always very small.
if(!originalCoordinates)
{
for (hit = hits; hit != NULL; hit = hit->next)
{
if(hit->chromEnd > start && hit->chromStart <= (start + motif->columnCount))
{
verbose(3, "found overlapping hits: %d-%d overlaps with %d-%d\n", start, start + motif->columnCount, hit->chromStart, hit->chromEnd);
break;
}
}
}
if(hit == NULL || hit->score < score)
{
if(hit == NULL)
{
AllocVar(hit);
slAddHead(&hits, hit);
hit->chrom = cloneString(chrom);
}
hit->chromStart = originalCoordinates ? chromStart : start;
hit->chromEnd = originalCoordinates ? chromEnd : start + motif->columnCount;
hit->score = score;
hit->strand[0] = strand;
}
}
verbose(3, "j: %d; score: %.2f\n", j, score);
}
}
slSort(&hits, bed6FloatCmpDesc);
int count;
float currentPrior = prior;
for(count = 1; hits != NULL; count++, hits = hits->next)
{
if(topOnly && count > topOnly)
break;
// Use a progressively weaker prior for hits with lower scores
verbose(3, "count: %d; score: %.2f; prior: %.2f; log2(prior / (1 - prior)): %.2f\n", count, hits->score, currentPrior, log2(currentPrior / (1 - currentPrior)));
if(hits->score >= minScoreCutoff - log2(currentPrior / (1 - currentPrior)))
{
printf("%s\t%d\t%d\t%s\t%.2f\t%c\n", chrom, originalCoordinates ? chromStart : hits->chromStart,
originalCoordinates ? chromEnd : hits->chromStart + motif->columnCount, name, hits->score, hits->strand[0]);
currentPrior = count == 1 ? priorBackoff : currentPrior * priorBackoff;
if(count > 2)
verbose(3, "hit for count: %d at %s:%d-%d\n", count, chrom, hits->chromStart, hits->chromStart + motif->columnCount);
}
else
break;
}
freeDnaSeq(&seq);
freeMem(dupe);
}
lineFileClose(&lf);
}
sqlDisconnect(&conn);
}
void intronEnds(char *database, char *table)
/* intronEnds - Gather stats on intron ends.. */
{
struct dyString *query = newDyString(1024);
struct sqlConnection *conn;
struct sqlResult *sr;
char **row;
struct genePred *gp;
int total = 0;
int gtag = 0;
int gcag = 0;
int atac = 0;
int ctac = 0;
DNA ends[4];
int exonIx, txStart;
struct dnaSeq *seq;
int rowOffset;
char strand;
rowOffset = hOffsetPastBin(database, NULL, table);
conn = hAllocConn(database);
sqlDyStringPrintf(query, "select * from %s", table);
if (chromName != NULL)
dyStringPrintf(query, " where chrom = '%s'", chromName);
if (cgiBoolean("withUtr"))
{
dyStringPrintf(query, " %s txStart != cdsStart",
(chromName == NULL ? "where" : "and"));
}
sr = sqlGetResult(conn, query->string);
while ((row = sqlNextRow(sr)) != NULL)
{
gp = genePredLoad(row+rowOffset);
strand = gp->strand[0];
txStart = gp->txStart;
seq = hDnaFromSeq(database, gp->chrom, txStart, gp->txEnd, dnaLower);
for (exonIx=1; exonIx < gp->exonCount; ++exonIx)
{
++total;
memcpy(ends, seq->dna + gp->exonEnds[exonIx-1] - txStart, 2);
memcpy(ends+2, seq->dna + gp->exonStarts[exonIx] - txStart - 2, 2);
if (strand == '-')
reverseComplement(ends, 4);
if (ends[0] == 'g' && ends[1] == 't' && ends[2] == 'a' && ends[3] == 'g')
++gtag;
if (ends[0] == 'g' && ends[1] == 'c' && ends[2] == 'a' && ends[3] == 'g')
++gcag;
if (ends[0] == 'a' && ends[1] == 't' && ends[2] == 'a' && ends[3] == 'c')
++atac;
if (ends[0] == 'c' && ends[1] == 't' && ends[2] == 'a' && ends[3] == 'c')
++ctac;
}
freeDnaSeq(&seq);
genePredFree(&gp);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
printf("gt/ag %d (%4.2f)\n", gtag, 100.0*gtag/total);
printf("gc/ag %d (%4.2f)\n", gcag, 100.0*gcag/total);
printf("at/ac %d (%4.2f)\n", atac, 100.0*atac/total);
printf("ct/ac %d (%4.2f)\n", ctac, 100.0*ctac/total);
printf("Total %d\n", total);
}
static void hgSeqConcatRegionsDb(char *db, char *chrom, int chromSize, char strand, char *name,
int rCount, unsigned *rStarts, unsigned *rSizes,
boolean *exonFlags, boolean *cdsFlags)
/* Concatenate and print out dna for a series of regions. */
{
// Note: this code use to generate different sequence ids if the global
// database in hdb was different than the db parameter. This functionality
// has been removed since the global database was removed and it didn't
// appear to be used.
struct dnaSeq *rSeq = NULL;
struct dnaSeq *cSeq = NULL;
char recName[256];
int seqStart, seqEnd;
int offset, cSize;
int i;
boolean isRc = (strand == '-') || cgiBoolean("hgSeq.revComp");
boolean maskRep = cgiBoolean("hgSeq.maskRepeats");
int padding5 = cgiOptionalInt("hgSeq.padding5", 0);
int padding3 = cgiOptionalInt("hgSeq.padding3", 0);
char *casing = cgiString("hgSeq.casing");
char *repMasking = cgiString("hgSeq.repMasking");
char *granularity = cgiOptionalString("hgSeq.granularity");
boolean concatRegions = granularity && sameString("gene", granularity);
if (rCount < 1)
return;
/* Don't support padding if granularity is gene (i.e. concat'ing all). */
if (concatRegions)
{
padding5 = padding3 = 0;
}
i = rCount - 1;
seqStart = rStarts[0] - (isRc ? padding3 : padding5);
seqEnd = rStarts[i] + rSizes[i] + (isRc ? padding5 : padding3);
/* Padding might push us off the edge of the chrom; if so, truncate: */
if (seqStart < 0)
{
if (isRc)
padding3 += seqStart;
else
padding5 += seqStart;
seqStart = 0;
}
/* if we know the chromSize, don't pad out beyond it */
if ((chromSize > 0) && (seqEnd > chromSize))
{
if (isRc)
padding5 += (chromSize - seqEnd);
else
padding3 += (chromSize - seqEnd);
seqEnd = chromSize;
}
if (seqEnd <= seqStart)
{
printf("# Null range for %s_%s (range=%s:%d-%d 5'pad=%d 3'pad=%d) (may indicate a query-side insert)\n",
db,
name,
chrom, seqStart+1, seqEnd,
padding5, padding3);
return;
}
if (maskRep)
{
rSeq = hDnaFromSeq(db, chrom, seqStart, seqEnd, dnaMixed);
if (sameString(repMasking, "N"))
lowerToN(rSeq->dna, strlen(rSeq->dna));
if (!sameString(casing, "upper"))
tolowers(rSeq->dna);
}
else if (sameString(casing, "upper"))
rSeq = hDnaFromSeq(db, chrom, seqStart, seqEnd, dnaUpper);
else
rSeq = hDnaFromSeq(db, chrom, seqStart, seqEnd, dnaLower);
/* Handle casing and compute size of concatenated sequence */
cSize = 0;
for (i=0; i < rCount; i++)
{
if ((sameString(casing, "exon") && exonFlags[i]) ||
(sameString(casing, "cds") && cdsFlags[i]))
{
int rStart = rStarts[i] - seqStart;
toUpperN(rSeq->dna+rStart, rSizes[i]);
}
cSize += rSizes[i];
}
cSize += (padding5 + padding3);
AllocVar(cSeq);
cSeq->dna = needLargeMem(cSize+1);
cSeq->size = cSize;
offset = 0;
for (i=0; i < rCount; i++)
{
int start = rStarts[i] - seqStart;
int size = rSizes[i];
if (i == 0)
{
start -= (isRc ? padding3 : padding5);
assert(start == 0);
size += (isRc ? padding3 : padding5);
}
if (i == rCount-1)
{
size += (isRc ? padding5 : padding3);
}
memcpy(cSeq->dna+offset, rSeq->dna+start, size);
offset += size;
}
assert(offset == cSeq->size);
cSeq->dna[offset] = 0;
freeDnaSeq(&rSeq);
if (isRc)
reverseComplement(cSeq->dna, cSeq->size);
safef(recName, sizeof(recName),
"%s_%s range=%s:%d-%d 5'pad=%d 3'pad=%d "
"strand=%c repeatMasking=%s",
db,
name,
chrom, seqStart+1, seqEnd,
padding5, padding3,
(isRc ? '-' : '+'),
(maskRep ? repMasking : "none"));
faWriteNext(stdout, recName, cSeq->dna, cSeq->size);
freeDnaSeq(&cSeq);
}
void doClusterMotifDetails(struct sqlConnection *conn, struct trackDb *tdb,
struct factorSource *cluster)
/* Display details about TF binding motif(s) in cluster */
{
char *motifTable = trackDbSetting(tdb, "motifTable"); // localizations
char *motifPwmTable = trackDbSetting(tdb, "motifPwmTable"); // PWM used to draw sequence logo
char *motifMapTable = trackDbSetting(tdb, "motifMapTable"); // map target to motif
struct slName *motifNames = NULL, *mn; // list of canonical motifs for the factor
struct dnaMotif *motif = NULL;
struct bed6FloatScore *hit = NULL, *maxHit = NULL;
char **row;
char query[256];
if (motifTable != NULL && sqlTableExists(conn, motifTable))
{
struct sqlResult *sr;
int rowOffset;
char where[256];
if (motifMapTable == NULL || !sqlTableExists(conn, motifMapTable))
{
// Assume cluster name is motif name if there is no map table
motifNames = slNameNew(cluster->name);
}
else
{
sqlSafef(query, sizeof(query),
"select motif from %s where target = '%s'", motifMapTable, cluster->name);
char *ret = sqlQuickString(conn, query);
if (ret == NULL)
{
// missing target from table -- no canonical motif
webNewEmptySection();
return;
}
motifNames = slNameListFromString(ret, ',');
}
for (mn = motifNames; mn != NULL; mn = mn->next)
{
sqlSafefFrag(where, sizeof(where), "name='%s' order by score desc limit 1", mn->name);
sr = hRangeQuery(conn, motifTable, cluster->chrom, cluster->chromStart,
cluster->chromEnd, where, &rowOffset);
if ((row = sqlNextRow(sr)) != NULL)
{
hit = bed6FloatScoreLoad(row + rowOffset);
if (maxHit == NULL || maxHit->score < hit->score)
maxHit = hit;
}
sqlFreeResult(&sr);
}
}
if (maxHit == NULL)
{
// Maintain table layout
webNewEmptySection();
return;
}
hit = maxHit;
webNewSection("Canonical Motif in Cluster");
char posLink[1024];
safef(posLink, sizeof(posLink),"<a href=\"%s&db=%s&position=%s%%3A%d-%d\">%s:%d-%d</a>",
hgTracksPathAndSettings(), database,
cluster->chrom, hit->chromStart+1, hit->chromEnd,
cluster->chrom, hit->chromStart+1, hit->chromEnd);
printf("<b>Motif Name:</b> %s<br>\n", hit->name);
printf("<b>Motif Score");
printf(":</b> %.2f<br>\n", hit->score);
printf("<b>Motif Position:</b> %s<br>\n", posLink);
printf("<b>Motif Strand:</b> %c<br>\n", (int)hit->strand[0]);
struct dnaSeq *seq = hDnaFromSeq(database, seqName, hit->chromStart, hit->chromEnd, dnaLower);
if (seq == NULL)
return;
if (hit->strand[0] == '-')
reverseComplement(seq->dna, seq->size);
if (motifPwmTable != NULL && sqlTableExists(conn, motifPwmTable))
{
motif = loadDnaMotif(hit->name, motifPwmTable);
if (motif == NULL)
return;
motifLogoAndMatrix(&seq, 1, motif);
}
}
void processVariant(char *database, char *proteinID, int aaSeqLen, int varStart, int varLen,
char *origSeq, char *varSeq, char *varId, FILE *outf)
{
char query[256];
struct sqlResult *sr;
char **row;
struct sqlConnection *conn;
char *qNameStr;
char *qSizeStr;
char *qStartStr;
char *qEndStr;
char *tNameStr=NULL;
char *tSizeStr;
char *tStartStr;
char *tEndStr;
char *blockCountStr;
char *blockSizesStr;
char *qStartsStr;
char *tStartsStr;
char *strand = NULL;
int blockCount=0;
char *exonStartStr = NULL;
int exonGenomeStartPos, exonGenomeEndPos;
char *exonGenomeStartStr = NULL;
char *chp, *chp2, *chp3;
char *exonSizeStr = NULL;
int j;
int exonStart, exonEnd;
int lastStart;
int lastLen;
int accumLen;
int exonLen;
int dvStart;
int varEnd;
struct dnaSeq *dnaseq;
conn= hAllocConn();
/* NOTE: the query below may not always return single answer, */
/* and kgProtMap and knownGene alignments may not be identical, so pick the closest one. */
safef(query,sizeof(query), "select qName, qSize, qStart, qEnd, tName, tSize, tStart, tEnd, blockCount, blockSizes, qStarts, tStarts, strand from %s.%s where qName='%s';",
database, "kgProtMap", proteinID);
sr = sqlMustGetResult(conn, query);
row = sqlNextRow(sr);
if (row == NULL)
{
sqlFreeResult(&sr);
hFreeConn(&conn);
return;
}
while (row != NULL)
{
qNameStr = cloneString(row[0]);
qSizeStr = cloneString(row[1]);
qStartStr = cloneString(row[2]);
qEndStr = cloneString(row[3]);
tNameStr = cloneString(row[4]);
tSizeStr = cloneString(row[5]);
tStartStr = cloneString(row[6]);
tEndStr = cloneString(row[7]);
blockCountStr = cloneString(row[8]);
blockSizesStr = cloneString(row[9]);
qStartsStr = cloneString(row[10]);
tStartsStr = cloneString(row[11]);
strand = cloneString(row[12]);
blockCount = atoi(blockCountStr);
if (sameWord(strand, "+"))
{
exonStartStr = qStartsStr;
exonGenomeStartStr = tStartsStr;
exonSizeStr = blockSizesStr;
accumLen = 0;
lastStart = 0;
chp2 = exonGenomeStartStr;
chp3 = blockSizesStr;
for (j=0; j< blockCount; j++)
{
chp = strstr(chp2, ",");
*chp = '\0';
exonStart = atoi(chp2);
chp2 = chp+1;
chp = strstr(chp3, ",");
*chp = '\0';
exonLen = atoi(chp3);
chp3 = chp + 1;
if (((accumLen + exonLen)/3) >= varStart)
{
dvStart = exonStart + (varStart - accumLen/3)*3;
np++;
dnaseq = hDnaFromSeq(tNameStr, dvStart, dvStart+varLen*3, dnaUpper);
/* check if first AA of variant agrees with genomic codon */
if (lookupCodon(dnaseq->dna) == origSeq[0])
{
np0++;
fprintf(outf, "%s\t%d\t%d\t%s\n", tNameStr, dvStart, dvStart+varLen*3, varId);
}
else
{
/* try next position */
dvStart = exonStart + (varStart - accumLen/3)*3 - 1;
dnaseq = hDnaFromSeq(tNameStr, dvStart, dvStart+varLen*3, dnaUpper);
if (lookupCodon(dnaseq->dna) == origSeq[0])
{
np1++;
fprintf(outf, "%s\t%d\t%d\t%s\n",
tNameStr, dvStart, dvStart+varLen*3, varId);
}
else
{
/* could further scan wider range for valid AA mapping */
}
}
break;
}
lastStart = exonStart;
accumLen = accumLen + exonLen;
}
}
else
{
exonStartStr = qStartsStr;
exonGenomeStartStr = tStartsStr;
exonSizeStr = blockSizesStr;
varEnd = aaSeqLen - varStart - varLen;
accumLen = 0;
lastStart = 0;
chp2 = exonGenomeStartStr;
chp3 = blockSizesStr;
for (j=0; j< blockCount; j++)
{
chp = strstr(chp2, ",");
*chp = '\0';
exonStart = atoi(chp2);
chp2 = chp+1;
chp = strstr(chp3, ",");
*chp = '\0';
exonLen = atoi(chp3);
chp3 = chp + 1;
if (((accumLen + exonLen)/3) >= varEnd)
{
nn++;
dvStart = exonStart + (varEnd - accumLen/3)*3;
dnaseq = hDnaFromSeq(tNameStr, dvStart, dvStart+varLen*3, dnaUpper);
reverseComplement(dnaseq->dna, (long)3);
if (lookupCodon(dnaseq->dna) == origSeq[0])
{
nn0++;
fprintf(outf, "%s\t%d\t%d\t%s\n", tNameStr, dvStart, dvStart+varLen*3, varId);
}
else
{
dvStart = exonStart + (varEnd - accumLen/3)*3 - 1;
dnaseq = hDnaFromSeq(tNameStr, dvStart, dvStart+varLen*3, dnaUpper);
reverseComplement(dnaseq->dna, (long)3);
if (lookupCodon(dnaseq->dna) == origSeq[0])
{
nn1++;
fprintf(outf, "%s\t%d\t%d\t%s\n", tNameStr, dvStart, dvStart+varLen*3, varId);
}
else
{
/* deal with this later */
}
}
fflush(outf);
break;
}
lastStart = exonStart;
accumLen = accumLen + exonLen;
}
}
row = sqlNextRow(sr);
}
sqlFreeResult(&sr);
hFreeConn(&conn);
}