SEXP r_cgiParameters() {
int i;
char *name;
Stringa value=stringCreate(16);
Texta keys=textCreate(8);
Texta values=textCreate(8);
SEXP r_keys, r_values;
cgiGetInit();
while(name = cgiGetNext(value)) {
textAdd(keys, name);
textAdd(values, string(value));
}
int n=arrayMax(keys);
PROTECT(r_keys=allocVector(STRSXP, n));
PROTECT(r_values=allocVector(STRSXP, n));
for(i=0; i<n; ++i) {
SET_STRING_ELT(r_keys, i, mkChar(textItem(keys,i)));
SET_STRING_ELT(r_values, i, mkChar(textItem(values,i)));
}
setNames(r_values, r_keys);
stringDestroy(value);
textDestroy(keys);
textDestroy(values);
UNPROTECT(2);
return(r_values);
}
int main (int argc, char *argv[])
{
GfrEntry *currGE;
int count;
int countRemoved;
int i;
if (argc != 3) {
usage ("%s <offsetCutoff> <minNumUniqueReads>",argv[0]);
}
count = 0;
countRemoved = 0;
int offsetCutOff = atoi (argv[1]);
int minNumUniqueReads = atoi (argv[2]);
gfr_init ("-");
puts (gfr_writeHeader ());
while (currGE = gfr_nextEntry ()) {
Array starts = arrayCreate( 100, int);
for (i = 0; i < arrayMax( currGE->interReads ); i++) {
int currStart = arrp(currGE->interReads, i, GfrInterRead)->readStart1 + arrp(currGE->interReads, i, GfrInterRead)->readStart2;
array(starts, arrayMax(starts), int) = currStart;
}
arraySort( starts, (ARRAYORDERF) arrayIntcmp );
arrayUniq( starts, NULL, (ARRAYORDERF) arrayIntcmp ) ;
int numUniqeOffsets = arrayMax( starts );
arrayDestroy( starts );
if (arrayMax( currGE->readsTranscript1 ) != arrayMax( currGE->readsTranscript2 ) )
die( "The two ends have a different number of reads");
Texta reads = textCreate(arrayMax(currGE->readsTranscript1));
for (i = 0; i < arrayMax(currGE->readsTranscript1); i++) {
Stringa strA = stringCreate( strlen(textItem( currGE->readsTranscript1, i) ) * 2 + 1);
stringAppendf( strA, textItem( currGE->readsTranscript1,i));
stringAppendf( strA, textItem( currGE->readsTranscript2,i));
textAdd( reads, string(strA));
stringDestroy( strA );
}
textUniqKeepOrder( reads );
int numRemaining = arrayMax( reads );
textDestroy ( reads );
if (numRemaining <= minNumUniqueReads || numUniqeOffsets <= offsetCutOff) {
countRemoved++;
continue;
}
puts (gfr_writeGfrEntry (currGE));
count++;
}
gfr_deInit ();
warn("%s_PCRFilter: offset=%d minNumUniqueReads=%d",
argv[0],offsetCutOff, minNumUniqueReads);
warn("%s_numRemoved: %d",argv[0],countRemoved);
warn("%s_numGfrEntries: %d",argv[0],count);
return 0;
}
SEXP c_read_biokit_exprs (SEXP filename) {
LineStream ls;
char* line;
const int MAND_NCOL=7; // the first column is the row name, and column 2-7 are mandatory
int add_ncol=0;
Texta it;
Texta rnames=textCreate(128);
Array mrpkms=arrayCreate(128, double);
Array mreads=arrayCreate(128, int);
Array srpkms=arrayCreate(128, double);
Array sreads=arrayCreate(128, int);
Array mprop=arrayCreate(128, double);
Array allmap = arrayCreate(128, int);
Array annos=arrayCreate(128, Texta);
Texta anno=NULL; // must have a NULL assigned; otherwise textCreateClear leads to memory error
Stringa str=stringCreate(8);
SEXP R_rnames, R_mrpkms, R_mreads, R_srpkms, R_sreads, R_mprop, R_allmap, R_res;
SEXP R_colnames, R_class;
int nprot=0;
int i=0;
int j=0;
int nrow=0;
const char* fn=CHAR(STRING_ELT(filename, 0));
ls = ls_createFromFile(strdup(fn));
ls_nextLine(ls); // skip the first header line
while(line = ls_nextLine(ls)) {
it = textFieldtokP(line, "\t");
if(arrayMax(it)<MAND_NCOL)
error("Input file must contain no less than %d columns", MAND_NCOL);
textAdd(rnames, textItem(it, 0));
array(mrpkms, arrayMax(mrpkms), double)=atof(textItem(it, 1));
array(mreads, arrayMax(mreads), int)=atoi(textItem(it, 2));
array(srpkms, arrayMax(srpkms), double)=atof(textItem(it, 3));
array(sreads, arrayMax(sreads), int)=atoi(textItem(it, 4));
array(mprop, arrayMax(mprop), double)=atof(textItem(it, 5));
array(allmap, arrayMax(allmap), int)=atoi(textItem(it, 6));
add_ncol = max(arrayMax(it)-MAND_NCOL, add_ncol);
textCreateClear(anno, arrayMax(it)-MAND_NCOL);
for(i=MAND_NCOL; i<arrayMax(it); ++i) {
textAdd(anno, textItem(it, i));
}
array(annos, arrayMax(annos), Texta)=textClone(anno);
nrow++;
}
R_rnames=PROTECT(allocVector(STRSXP, nrow)); nprot++;
R_mrpkms=PROTECT(allocVector(REALSXP, nrow)); nprot++;
R_mreads=PROTECT(allocVector(INTSXP, nrow)); nprot++;
R_srpkms=PROTECT(allocVector(REALSXP, nrow)); nprot++;
R_sreads=PROTECT(allocVector(INTSXP, nrow)); nprot++;
R_mprop=PROTECT(allocVector(REALSXP, nrow)); nprot++;
R_allmap=PROTECT(allocVector(INTSXP, nrow)); nprot++;
for(i=0; i<nrow; ++i) {
SET_STRING_ELT(R_rnames, i, mkChar(textItem(rnames, i)));
REAL(R_mrpkms)[i]=arru(mrpkms, i, double);
INTEGER(R_mreads)[i]=arru(mreads, i, int);
REAL(R_srpkms)[i]=arru(srpkms, i, double);
INTEGER(R_sreads)[i]=arru(sreads, i, int);
REAL(R_mprop)[i]=arru(mprop, i, double);
INTEGER(R_allmap)[i]=arru(allmap, i, int);
}
R_res=PROTECT(allocVector(VECSXP, MAND_NCOL+add_ncol-1)); nprot++;
SET_VECTOR_ELT(R_res, 0, R_mrpkms);
SET_VECTOR_ELT(R_res, 1, R_mreads);
SET_VECTOR_ELT(R_res, 2, R_srpkms);
SET_VECTOR_ELT(R_res, 3, R_sreads);
SET_VECTOR_ELT(R_res, 4, R_mprop);
SET_VECTOR_ELT(R_res, 5, R_allmap);
for(i=0; i<add_ncol; ++i) {
SEXP R_anno=NULL;
R_anno=PROTECT(allocVector(STRSXP, nrow));
for(j=0; j<nrow; ++j) {
anno=array(annos, j, Texta);
if(arrayMax(anno)>i) {
SET_STRING_ELT(R_anno, j, mkChar(textItem(anno, i)));
} else {
SET_STRING_ELT(R_anno, j, R_NaString);
}
}
SET_VECTOR_ELT(R_res, i+MAND_NCOL-1, R_anno); // -1 because the first column is row name
UNPROTECT(1);
}
PROTECT(R_colnames=allocVector(STRSXP, MAND_NCOL+add_ncol-1)); nprot++;
PROTECT(R_class=allocVector(STRSXP, 1)); nprot++;
SET_STRING_ELT(R_colnames, 0, mkChar("RPKM_MultiMap"));
SET_STRING_ELT(R_colnames, 1, mkChar("ReadCount_MultiMap"));
SET_STRING_ELT(R_colnames, 2, mkChar("RPKM_UniqMap"));
SET_STRING_ELT(R_colnames, 3, mkChar("ReadCount_UniqMap"));
SET_STRING_ELT(R_colnames, 4, mkChar("MultiProp"));
SET_STRING_ELT(R_colnames, 5, mkChar("AllMappingReads"));
for(i=0; i<add_ncol; ++i) {
stringPrintf(str, "Annotation%d", i+1);
SET_STRING_ELT(R_colnames, i+MAND_NCOL-1,
mkChar(string(str)));
}
SET_STRING_ELT(R_class, 0, mkChar("data.frame"));
setAttrib(R_res, install("names"), R_colnames);
setAttrib(R_res, install("row.names"), R_rnames);
setAttrib(R_res, install("class"), R_class);
for(i=0; i<nrow; ++i) {
textDestroy(array(annos, i, Texta));
}
arrayDestroy(annos);
arrayDestroy(rnames);
arrayDestroy(mrpkms);
arrayDestroy(mreads);
arrayDestroy(srpkms);
arrayDestroy(sreads);
arrayDestroy(mprop);
arrayDestroy(allmap);
stringDestroy(str);
ls_destroy(ls);
UNPROTECT(nprot);
return(R_res);
}
int main (int argc, char *argv[])
{
GfrEntry *currGE;
int i,j,k,l, h,index;
Stringa buffer,cmd,fnSequencesToAlign;
FILE *fp;
FILE *fp1;
FILE *fp2;
FILE *freads1;
FILE *freads2;
Array gfrEntries;
BowtieQuery *currBQ,testBQ;
BowtieEntry *currBE;
Texta seqNames;
int readSize1, readSize2, minReadSize;
Array bowtieQueries;
char transcriptNumber;
int isHomologous,homologousCount;
int count;
int countRemoved;
unsigned short int tooMany;
BlatQuery *blQ;
config *conf;
if ((conf = confp_open(getenv("FUSIONSEQ_CONFPATH"))) == NULL) {
die("%s:\tCannot find .fusionseqrc", argv[0]);
return EXIT_FAILURE;
}
if ( (confp_get( conf, "BLAT_TWO_BIT_TO_FA")) == NULL) {
die("%s:\tCannot find BLAT_TWO_BIT_TO_FA in the configuration file: %s", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}
if ( (confp_get( conf,"BLAT_DATA_DIR")) == NULL) {
die("%s:\tCannot find BLAT_DATA_DIR in the configuration file: %sc", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}
if( confp_get( conf, "TMP_DIR")==NULL ) {
die("%s:\tCannot find TMP_DIR in the configuration file: %s)", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}
if( confp_get( conf, "BLAT_GFSERVER")==NULL ) {
die("%s:\tCannot find BLAT_GFSERVER in the configuration file: %s)", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}
if( confp_get( conf, "BLAT_GFCLIENT")==NULL ) {
die("%s:\tCannot find BLAT_GFCLIENT in the configuration file: %s)", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}
if( confp_get( conf, "BLAT_GFSERVER_HOST")==NULL ) {
die("%s:\tCannot find BLAT_GFSERVER_HOST in the configuration file: %s)", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}if( confp_get( conf, "BLAT_GFSERVER_PORT")==NULL ) {
die("%s:\tCannot find BLAT_GFSERVER_PORT in the configuration file: %s)", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}
if( confp_get( conf, "PSEUDOGENE_DIR")==NULL ) {
die("%s:\tCannot find PSEUDOGENE_DIR in the configuration file: %s)", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}
if( confp_get( conf, "PSEUDOGENE_FILENAME")==NULL ) {
die("%s:\tCannot find PSEUDOGENE_FILENAME in the configuration file: %s)", argv[0], getenv("FUSIONSEQ_CONFPATH") );
return EXIT_FAILURE;
}
cmd = stringCreate (100);
// initializing the gfServers
stringPrintf( cmd, "%s status %s %s &> /dev/null", confp_get( conf, "BLAT_GFSERVER"), confp_get( conf, "BLAT_GFSERVER_HOST"), confp_get( conf, "BLAT_GFSERVER_PORT") );
int ret = hlr_system( string(cmd), 1 );
if( ret != 0 ) { // not initialized
stringPrintf( cmd , "%s -repMatch=100000 -tileSize=12 -canStop -log=%s/gfServer_genome.log start %s %s %s/%s &", confp_get( conf, "BLAT_GFSERVER"), confp_get(conf, "TMP_DIR"),confp_get( conf, "BLAT_GFSERVER_HOST"), confp_get( conf, "BLAT_GFSERVER_PORT"), confp_get(conf, "BLAT_DATA_DIR"), confp_get(conf, "BLAT_TWO_BIT_DATA_FILENAME"));
hlr_system( string( cmd ), 0 );
long int startTime = time(0);
stringPrintf( cmd , "%s status %s %s &2> /dev/null", confp_get( conf, "BLAT_GFSERVER"), confp_get( conf, "BLAT_GFSERVER_HOST"), confp_get( conf, "BLAT_GFSERVER_PORT"));
while( hlr_system( string(cmd), 1) && (time(0)-startTime)<600 ) ;
if( hlr_system( string(cmd), 1 ) != 0 ) {
die("gfServer for %s/%s not initialized: %s %s %s", confp_get(conf, "BLAT_DATA_DIR"), confp_get(conf, "BLAT_TWO_BIT_DATA_FILENAME"), confp_get( conf, "BLAT_GFSERVER"), confp_get( conf, "BLAT_GFSERVER_HOST"), confp_get( conf, "BLAT_GFSERVER_PORT"));
return EXIT_FAILURE;
}
}
// end initialization
gfr_init ("-");
gfrEntries = gfr_parse ();
if (arrayMax (gfrEntries) == 0){
puts (gfr_writeHeader ());
gfr_deInit ();
return 0;
}
seqNames = textCreate (10000);
buffer = stringCreate (100);
fnSequencesToAlign = stringCreate (100);
count = 0;
countRemoved = 0;
stringPrintf( buffer, "%s/%s", confp_get( conf, "PSEUDOGENE_DIR"), confp_get( conf, "PSEUDOGENE_FILENAME") );
intervalFind_addIntervalsToSearchSpace (string(buffer),0);
puts (gfr_writeHeader ());
for (i = 0; i < arrayMax (gfrEntries); i++) {
currGE = arrp (gfrEntries,i,GfrEntry);
homologousCount = 0;
minReadSize=10000;
// creating two fasta files with the two genes
stringPrintf( cmd, "%s %s/%s -seq=%s -start=%d -end=%d %s/%s_transcript1.fa", confp_get(conf, "BLAT_TWO_BIT_TO_FA") , confp_get(conf, "BLAT_DATA_DIR"), confp_get(conf, "BLAT_TWO_BIT_DATA_FILENAME"), currGE->chromosomeTranscript1, currGE->startTranscript1, currGE->endTranscript1, confp_get(conf, "TMP_DIR"), currGE->id);
hlr_system( string(cmd) , 0);
stringPrintf( cmd, "%s %s/%s -seq=%s -start=%d -end=%d %s/%s_transcript2.fa", confp_get(conf, "BLAT_TWO_BIT_TO_FA"), confp_get(conf, "BLAT_DATA_DIR"), confp_get(conf, "BLAT_TWO_BIT_DATA_FILENAME"), currGE->chromosomeTranscript2, currGE->startTranscript2, currGE->endTranscript2, confp_get(conf, "TMP_DIR"), currGE->id);
hlr_system( string(cmd) , 0);
Stringa fa1 = stringCreate( 100 );
Stringa fa2 = stringCreate( 100 );
// creating the two fasta files with the reads
stringPrintf( fa1, "%s/%s_reads1.fa", confp_get(conf, "TMP_DIR"), currGE->id);
if (!(freads1 = fopen ( string(fa1) ,"w"))) {
die ("Unable to open file: %s",string (fa1));
}
// writing the reads of the first end into file
for (l = 0; l < arrayMax (currGE->readsTranscript1); l++) {
char* currRead1 = hlr_strdup( textItem (currGE->readsTranscript1,l)); // read1
readSize1 = strlen( currRead1 );
if( readSize1 == 0 ) die("Read size cannot be zero: read1[ %s ]", currRead1);
if( readSize1 < minReadSize ) minReadSize = readSize1;
fprintf( freads1, ">%d\n%s\n", l, currRead1 );
hlr_free( currRead1 );
}
fclose( freads1 );
stringPrintf( fa2, "%s/%s_reads2.fa", confp_get(conf, "TMP_DIR"), currGE->id);
if (!(freads2 = fopen ( string(fa2) ,"w"))) {
die ("Unable to open file: %s",string (fa2));
}
// writing the reads of the second end into file
for (l = 0; l < arrayMax (currGE->readsTranscript2); l++) {
char* currRead2 = hlr_strdup( textItem (currGE->readsTranscript2,l)); // read2
readSize2 = strlen( currRead2 );
if( readSize2 == 0 ) die("Read size cannot be zero: read2[ %s ]", currRead2);
if( readSize2 < minReadSize ) minReadSize = readSize2;
fprintf( freads2, ">%d\n%s\n", l, currRead2 );
hlr_free( currRead2 );
}
fclose( freads2 );
// collapse the reads 2 ## requires the FASTX package
stringPrintf( cmd, "%s -i %s/%s_reads2.fa -o %s/%s_reads2.collapsed.fa", confp_get(conf, "FASTX_COLLAPSER"), confp_get(conf, "TMP_DIR"), currGE->id, confp_get(conf, "TMP_DIR"), currGE->id );
hlr_system (string (cmd),0);
//blat of reads2 against the first transcript
stringPrintf( cmd, "%s -t=dna -out=psl -fine -tileSize=15 %s/%s_transcript1.fa %s/%s_reads2.collapsed.fa stdout",confp_get(conf, "BLAT_BLAT"), confp_get(conf, "TMP_DIR"), currGE->id, confp_get(conf, "TMP_DIR"), currGE->id );
// reading the results of blast from Pipe
blatParser_initFromPipe( string(cmd) );
while( blQ = blatParser_nextQuery() ) {
int nucleotideOverlap = getNucleotideOverlap ( blQ );
if ( nucleotideOverlap > ( ((double)readSize2)* atof(confp_get(conf,"MAX_OVERLAP_ALLOWED"))) ) {
char* value = strchr(blQ->qName,'-');
homologousCount+=atoi(value+1);
}
}
blatParser_deInit();
// collapse the reads 1 ## requires the FASTX package on the path
stringPrintf( cmd, "%s -i %s/%s_reads1.fa -o %s/%s_reads1.collapsed.fa", confp_get(conf, "FASTX_COLLAPSER"), confp_get(conf, "TMP_DIR"), currGE->id, confp_get(conf, "TMP_DIR"), currGE->id );
hlr_system (string (cmd),0);
//blat of reads1 against the second transcript
stringPrintf( cmd, "%s -t=dna -out=psl -fine -tileSize=15 %s/%s_transcript2.fa %s/%s_reads1.collapsed.fa stdout",confp_get(conf, "BLAT_BLAT"), confp_get(conf, "TMP_DIR"), currGE->id, confp_get(conf, "TMP_DIR"), currGE->id );
blatParser_initFromPipe( string(cmd) );
while( blQ = blatParser_nextQuery() ) {
int nucleotideOverlap = getNucleotideOverlap ( blQ );
if ( nucleotideOverlap > ( ((double)readSize1)* atof(confp_get(conf,"MAX_OVERLAP_ALLOWED"))) ) {
char* value = strchr(blQ->qName,'-');
homologousCount+=atoi(value+1);
}
}
blatParser_deInit();
stringPrintf (cmd,"cd %s;rm -rf %s_reads?.fa %s_reads?.collapsed.fa %s_transcript?.fa", confp_get(conf, "TMP_DIR"), currGE->id,currGE->id,currGE->id);
hlr_system( string(cmd) , 0);
if (((double)homologousCount / (double)arrayMax(currGE->readsTranscript1)) <= atof(confp_get(conf, "MAX_FRACTION_HOMOLOGOUS")) ) {
homologousCount = 0;
// there is no homology between the two genes, but what about the rest of the genome
writeFasta( currGE, &minReadSize, confp_get(conf, "TMP_DIR") );
stringPrintf(cmd, "cd %s; %s %s %s / -t=dna -q=dna -minScore=%d -out=psl %s_reads.fa %s.smallhomology.psl &>/dev/null", confp_get(conf, "TMP_DIR"), confp_get( conf, "BLAT_GFCLIENT"), confp_get( conf, "BLAT_GFSERVER_HOST"), confp_get( conf, "BLAT_GFSERVER_PORT"), minReadSize - (int)(0.1 * minReadSize) > 20 ? minReadSize - (int) (0.1 * minReadSize) : 20 , currGE->id, currGE->id);
int attempts=0;
ret = hlr_system( string(cmd), 1 );
while( hlr_system( string(cmd), 1 ) && attempts<5000 ) attempts++;
if( attempts == 5000 ) {
die("Cannot map the reads %s", string( cmd ));
return EXIT_FAILURE;
}
// reading the results of blast from File
stringPrintf(cmd, "%s/%s.smallhomology.psl", confp_get( conf, "TMP_DIR"), currGE->id);
blatParser_initFromFile( string(cmd) );
tooMany = 1;
while( blQ = blatParser_nextQuery() ) {
tooMany = 0;
checkPseudogeneOverlap( blQ );
if( arrayMax( blQ->entries ) > 1 ) {
homologousCount+= arrayMax( blQ->entries ) - 1;
char* value = strchr( blQ->qName,'/' );
if( value ) *value = '\0'; else die("Not a valid index in the blat query name:\t%s", blQ->qName );
int indexOfInter = atoi( blQ->qName ); // the following three lines should removed the read if writing the GFR entry
GfrInterRead *currGIR = arrp( currGE->interReads, indexOfInter, GfrInterRead );
currGIR->flag = 1;
}
}
blatParser_deInit();
if ( tooMany == 1 || ( ( (double) homologousCount / (double) ( arrayMax(currGE->readsTranscript1) + arrayMax(currGE->readsTranscript2) ) ) > atof(confp_get(conf, "MAX_FRACTION_HOMOLOGOUS")) ) ) {
countRemoved++;
stringPrintf (cmd,"cd %s; rm -rf %s_reads*.fa %s_reads?.collapsed.fa %s_transcript?.fa %s.smallhomology.psl", confp_get(conf, "TMP_DIR"), currGE->id,currGE->id,currGE->id,currGE->id);
hlr_system( string(cmd), 1 );
continue;
}
// writing the gfrEntry, if everthing else didn't stop
if( homologousCount > 0 ) updateStats( currGE );
puts (gfr_writeGfrEntry (currGE));
count++;
// removing temporary files
stringPrintf (cmd,"cd %s;rm -rf %s_reads*.fa %s_reads?.collapsed.fa %s_transcript?.fa %s.smallhomology.psl", confp_get(conf, "TMP_DIR"), currGE->id,currGE->id,currGE->id,currGE->id);
hlr_system( string(cmd) , 1);
} else {
countRemoved++;
}
}
gfr_deInit ();
stringDestroy (fnSequencesToAlign);
stringDestroy (cmd);
stringDestroy (buffer);
warn ("%s_numRemoved: %d",argv[0],countRemoved);
warn ("%s_numGfrEntries: %d",argv[0],count);
confp_close(conf);
return EXIT_SUCCESS;
}
