void makeMotifs(char *inFile, struct hash *tfHash, char *outFile)
/* Parse input motifs and save them to outFile in dnaMotif format. */
{
struct lineFile *lf = lineFileOpen(inFile, TRUE);
FILE *f = mustOpen(outFile, "w");
struct hashEl *hel;
for (;;)
{
char *line;
char *words[256], *word;
int wordCount;
struct dnaMotif *motif;
if (!lineFileSkipTo(lf, "Probability matrix for"))
break;
lineFileNeedNext(lf, &line, NULL);
wordCount = chopLine(line, words);
if (wordCount >= ArraySize(words))
errAbort("Line %d of %s is too long\n", lf->lineIx, lf->fileName);
if (!sameString(words[0], "#"))
badFormat(lf);
AllocVar(motif);
motif->columnCount = wordCount-1;
readBaseProbs(lf, words, "#A", &motif->aProb, motif->columnCount);
readBaseProbs(lf, words, "#C", &motif->cProb, motif->columnCount);
readBaseProbs(lf, words, "#T", &motif->tProb, motif->columnCount);
readBaseProbs(lf, words, "#G", &motif->gProb, motif->columnCount);
if (!lineFileSkipTo(lf, "Source:"))
lineFileUnexpectedEnd(lf);
lineFileReuse(lf);
lineFileNeedNext(lf, &line, NULL);
word = nextWord(&line);
word = nextWord(&line);
if (word == NULL)
errAbort("Short Source: line %d of %s", lf->lineIx, lf->fileName);
motif->name = cloneString(word);
hel = hashLookup(tfHash, motif->name);
if (hel == NULL)
errAbort("%s in %s but not GFFs", motif->name, lf->fileName);
hel->val = motif;
dnaMotifTabOut(motif, f);
}
carefulClose(&f);
lineFileClose(&lf);
}
void iriToDnaMotif(char *inName, char *outName)
/* iriToDnaMotif - Convert improbRunInfo to dnaMotif. */
{
FILE *f = mustOpen(outName, "w");
static struct dnaMotif motif;
struct improbRunInfo *iriList = improbRunInfoLoadAll(inName);
struct improbRunInfo *iri;
for (iri = iriList; iri != NULL; iri = iri->next)
{
motif.name = iri->name;
motif.columnCount = iri->columnCount;
motif.aProb = iri->aProb;
motif.cProb = iri->cProb;
motif.gProb = iri->gProb;
motif.tProb = iri->tProb;
dnaMotifTabOut(&motif, f);
}
}
void emblMatrixToMotif(char *inName, char *outName)
/* emblMatrixToMotif - Convert transfac matrix in EMBL format to dnaMotif. */
{
struct hash *hash = NULL;
struct lineFile *lf = emblOpen(inName, NULL);
FILE *f = mustOpen(outName, "w");
struct dnaMotif *motif;
while ((hash = emblRecord(lf)) != NULL)
{
char *ac = hashFindVal(hash, "AC");
char *po = hashFindVal(hash, "P0");
if (ac != NULL && po != NULL && orgFits(hash))
{
motif = emblToMotif(ac, hash);
dnaMotifTabOut(motif, f);
dnaMotifFree(&motif);
}
}
}
struct hash *loadMotifWeights(struct sqlConnection *conn, char *fileName,
char *table)
/* Load in XML weight motif file and save it in tab-separated format
* and in hash keyed by motif name. */
{
struct esmMotifs *motifs = esmMotifsLoad(fileName);
struct esmMotif *motif;
FILE *f = hgCreateTabFile(tmpDir, table);
struct dyString *dy = dyStringNew(512);
struct hash *hash = newHash(16);
for (motif = motifs->esmMotif; motif != NULL; motif = motif->next)
{
struct esmWeights *weights = motif->esmWeights;
int posCount = slCount(weights->esmPosition);
struct esmPosition *pos;
struct dnaMotif *dm;
char name[64];
fixMotifName(motif->Name, name, sizeof(name));
AllocVar(dm);
dm->name = cloneString(name);
dm->columnCount = posCount;
AllocArray(dm->aProb, posCount);
AllocArray(dm->cProb, posCount);
AllocArray(dm->gProb, posCount);
AllocArray(dm->tProb, posCount);
for (pos = weights->esmPosition; pos != NULL; pos = pos->next)
{
char *row[5];
double odds[4], sumOdds = 0;
int i;
int ix = pos->Num;
int rowSize = chopString(pos->Weights, ";", row, ArraySize(row));
if (rowSize != 4)
errAbort("Expecting 4 values for weights in position %d of Motif %s",
pos->Num, motif->Name);
if (ix >= posCount)
errAbort("Num %d out of range in Motif %s", ix, motif->Name);
for (i=0; i<4; ++i)
{
odds[i] = exp(atof(row[0]));
sumOdds += odds[i];
}
dm->aProb[ix] = odds[0]/sumOdds;
dm->cProb[ix] = odds[1]/sumOdds;
dm->gProb[ix] = odds[2]/sumOdds;
dm->tProb[ix] = odds[3]/sumOdds;
}
dnaMotifTabOut(dm, f);
hashAdd(hash, dm->name, dm);
}
dyStringPrintf(dy,
"CREATE TABLE %s (\n"
" name varchar(16) not null, # Motif name.\n"
" columnCount int not null, # Count of columns in motif.\n"
" aProb longblob not null, # Probability of A's in each column.\n"
" cProb longblob not null, # Probability of C's in each column.\n"
" gProb longblob not null, # Probability of G's in each column.\n"
" tProb longblob not null, # Probability of T's in each column.\n"
" #Indices\n"
" PRIMARY KEY(name)\n"
")\n", table);
sqlRemakeTable(conn, table, dy->string);
hgLoadTabFile(conn, tmpDir, table, &f);
hgRemoveTabFile(tmpDir, table);
verbose(1, "Processed %d motifs into %s\n", slCount(motifs->esmMotif), table);
return hash;
}
