void makeTmpSai(struct sqlConnection *conn, struct cdwValidFile *vf, char *genoFile,
char **retSampleFile, char **retSaiFile)
/* Given a fastq file, make a subsample of it 100k reads long and align it with
* bwa producing a sai file of given name. */
{
/* Get fastq record */
long long fileId = vf->fileId;
struct cdwFastqFile *fqf = cdwFastqFileFromFileId(conn, fileId);
if (fqf == NULL)
errAbort("No cdwFastqFile record for file id %lld", fileId);
/* Create downsampled fastq in temp directory - downsampled more than default even. */
char sampleFastqName[PATH_LEN];
cdwMakeTempFastqSample(fqf->sampleFileName, FASTQ_SAMPLE_SIZE, sampleFastqName);
verbose(1, "downsampled %s into %s\n", vf->licensePlate, sampleFastqName);
/* Do alignment */
char cmd[3*PATH_LEN];
char *saiName = cloneString(rTempName(cdwTempDir(), "cdwPairSample", ".sai"));
safef(cmd, sizeof(cmd), "bwa aln -t 3 %s %s > %s", genoFile, sampleFastqName, saiName);
mustSystem(cmd);
/* Save return variables, clean up, and go home. */
*retSampleFile = cloneString(sampleFastqName);
*retSaiFile = saiName;
cdwFastqFileFree(&fqf);
}
void fastqRepeatQa(struct sqlConnection *conn, struct cdwFile *ef, struct cdwValidFile *vf)
/* Do repeat QA if possible on fastq file. */
{
/* First see if total repeat content is already in our table, in which case we are done. */
long long fileId = ef->id;
char query[512];
sqlSafef(query, sizeof(query),
"select count(*) from cdwQaRepeat where fileId=%lld and repeatClass='total'" , fileId);
if (sqlQuickNum(conn, query) != 0)
return; /* We've done this already */
/* Get sample file name from fastq table. */
struct cdwFastqFile *fqf = cdwFastqFileForFileId(conn, fileId);
if (fqf == NULL)
errAbort("No edqFastqRecord for %s", vf->licensePlate);
char *fastqPath = fqf->sampleFileName;
char bwaIndex[PATH_LEN];
safef(bwaIndex, sizeof(bwaIndex), "%s%s/repeatMasker/repeatMasker.fa",
cdwValDataDir, vf->ucscDb);
char cmd[3*PATH_LEN];
char *saiName = cloneString(rTempName(cdwTempDir(), "cdwQaRepeat", ".sai"));
safef(cmd, sizeof(cmd), "bwa aln %s %s > %s", bwaIndex, fastqPath, saiName);
mustSystem(cmd);
char *samName = cloneString(rTempName(cdwTempDir(), "cdwQaRepeat", ".sam"));
safef(cmd, sizeof(cmd), "bwa samse %s %s %s > %s", bwaIndex, saiName, fastqPath, samName);
mustSystem(cmd);
remove(saiName);
char *raName = cloneString(rTempName(cdwTempDir(), "cdwQaRepeat", ".ra"));
safef(cmd, sizeof(cmd), "edwSamRepeatAnalysis %s %s", samName, raName);
mustSystem(cmd);
verbose(2, "mustSystem(%s)\n", cmd);
remove(samName);
raIntoCdwRepeatQa(raName, conn, fileId);
remove(raName);
#ifdef SOON
#endif /* SOON */
freez(&saiName);
freez(&samName);
freez(&raName);
cdwFastqFileFree(&fqf);
}
void pairedEndQa(struct sqlConnection *conn, struct cdwFile *ef, struct cdwValidFile *vf)
/* Look for other end, do a pairwise alignment, and save results in database. */
{
verbose(2, "pairedEndQa on %u %s %s\n", ef->id, ef->cdwFileName, ef->submitFileName);
/* Get other end, return if not found. */
struct cdwValidFile *otherVf = cdwOppositePairedEnd(conn, ef, vf);
if (otherVf == NULL)
return;
if (otherVf->fileId > vf->fileId)
return;
struct cdwValidFile *vf1, *vf2;
struct cdwQaPairedEndFastq *pair = cdwQaPairedEndFastqFromVfs(conn, vf, otherVf, &vf1, &vf2);
if (pair != NULL)
{
cdwValidFileFree(&otherVf);
return;
}
/* Get target assembly and figure out path for BWA index. */
struct cdwAssembly *assembly = cdwAssemblyForUcscDb(conn, vf->ucscDb);
assert(assembly != NULL);
char genoFile[PATH_LEN];
safef(genoFile, sizeof(genoFile), "%s%s/bwaData/%s.fa",
cdwValDataDir, assembly->ucscDb, assembly->ucscDb);
verbose(1, "aligning subsamples on %u vs. %u paired reads\n", vf1->fileId, vf2->fileId);
/* Make alignments of subsamples. */
char *sample1 = NULL, *sample2 = NULL, *sai1 = NULL, *sai2 = NULL;
makeTmpSai(conn, vf1, genoFile, &sample1, &sai1);
makeTmpSai(conn, vf2, genoFile, &sample2, &sai2);
/* Make paired end alignment */
char *tmpSam = cloneString(rTempName(cdwTempDir(), "cdwPairSample", ".sam"));
char command[6*PATH_LEN];
safef(command, sizeof(command),
"bwa sampe -n 1 -N 1 -f %s %s %s %s %s %s"
, tmpSam, genoFile, sai1, sai2, sample1, sample2);
mustSystem(command);
/* Make ra file with pairing statistics */
char *tmpRa = cloneString(rTempName(cdwTempDir(), "cdwPairSample", ".ra"));
safef(command, sizeof(command),
"edwSamPairedEndStats -maxInsert=%d %s %s", maxInsert, tmpSam, tmpRa);
mustSystem(command);
/* Read RA file into variables. */
struct cdwQaPairedEndFastq *pe = cdwQaPairedEndFastqOneFromRa(tmpRa);
/* Update database with record. */
struct sqlConnection *freshConn = cdwConnectReadWrite();
char query[256];
sqlSafef(query, sizeof(query),
"insert into cdwQaPairedEndFastq "
"(fileId1,fileId2,concordance,distanceMean,distanceStd,distanceMin,distanceMax,recordComplete) "
" values (%u,%u,%g,%g,%g,%g,%g,1)"
, vf1->fileId, vf2->fileId, pe->concordance, pe->distanceMean
, pe->distanceStd, pe->distanceMin, pe->distanceMax);
sqlUpdate(conn, query);
sqlDisconnect(&freshConn);
/* Clean up and go home. */
cdwValidFileFree(&otherVf);
remove(sample1);
remove(sample2);
remove(sai1);
remove(sai2);
remove(tmpSam);
remove(tmpRa);
#ifdef SOON
#endif /* SOON */
freez(&sample1);
freez(&sample2);
freez(&sai1);
freez(&sai2);
freez(&tmpSam);
freez(&tmpRa);
cdwQaPairedEndFastqFree(&pe);
cdwValidFileFree(&otherVf);
}