static IDnum chooseNextInternalPreNode(IDnum currentPreNodeID,
IDnum sequenceIndex,
PreGraph * preGraph, IDnum * chains)
{
if (currentPreNodeID >= preNodeCount_pg(preGraph))
return 0;
if (sequenceIndex >= sequenceCount_pg(preGraph))
return currentPreNodeID + 1;
if (currentPreNodeID + 1 < chains[sequenceIndex + 1])
return currentPreNodeID + 1;
return 0;
}
// Threads each sequences and creates preArcs according to road map indications
static void connectPreNodes(RoadMapArray * rdmaps, PreGraph * preGraph,
IDnum * chains)
{
IDnum sequenceIndex;
IDnum referenceCount = rdmaps->referenceCount;
#ifdef _OPENMP
annotationOffset = mallocOrExit(rdmaps->length + 1, Coordinate);
annotationOffset[0] = 0;
for (sequenceIndex = 1; sequenceIndex <= rdmaps->length; sequenceIndex++)
annotationOffset[sequenceIndex] = annotationOffset[sequenceIndex - 1] +
getAnnotationCount(getRoadMapInArray(rdmaps, sequenceIndex - 1));
#else
Annotation *annot = rdmaps->annotations;
#endif
if (rdmaps->referenceCount > 0)
allocatePreMarkerCountSpace_pg(preGraph);
#ifdef _OPENMP
int threads = omp_get_max_threads();
if (threads > 8)
threads = 8;
#pragma omp parallel for num_threads(threads)
#endif
for (sequenceIndex = 1;
sequenceIndex <= sequenceCount_pg(preGraph);
sequenceIndex++) {
#ifdef _OPENMP
Annotation *annot = getAnnotationInArray(rdmaps->annotations, annotationOffset[sequenceIndex - 1]);
#endif
RoadMap *rdmap;
Coordinate currentPosition, currentInternalPosition;
IDnum currentPreNodeID, nextInternalPreNodeID;
IDnum annotIndex, lastAnnotIndex;
boolean isReference;
if (sequenceIndex % 1000000 == 0)
velvetLog("Connecting %li / %li\n", (long) sequenceIndex,
(long) sequenceCount_pg(preGraph));
rdmap = getRoadMapInArray(rdmaps, sequenceIndex - 1);
annotIndex = 0;
lastAnnotIndex = getAnnotationCount(rdmap);
nextInternalPreNodeID = chooseNextInternalPreNode
(chains[sequenceIndex] - 1, sequenceIndex,
preGraph, chains);
isReference = (sequenceIndex <= referenceCount);
currentPosition = 0;
currentInternalPosition = 0;
currentPreNodeID = 0;
// Recursion up to last annotation
while (annotIndex < lastAnnotIndex
|| nextInternalPreNodeID != 0) {
if (annotIndex == lastAnnotIndex
|| (nextInternalPreNodeID != 0
&& currentInternalPosition <
getPosition(annot))) {
connectPreNodeToTheNext(¤tPreNodeID,
nextInternalPreNodeID,
¤tPosition,
sequenceIndex,
isReference,
preGraph);
nextInternalPreNodeID =
chooseNextInternalPreNode
(currentPreNodeID, sequenceIndex,
preGraph, chains);
currentInternalPosition +=
getPreNodeLength_pg(currentPreNodeID,
preGraph);
} else {
connectAnnotation(¤tPreNodeID, annot,
¤tPosition,
sequenceIndex, isReference,
preGraph);
annot = getNextAnnotation(annot);
annotIndex++;
}
}
}
if (rdmaps->referenceCount > 0) {
allocatePreMarkerSpace_pg(preGraph);
createPreMarkers(rdmaps, preGraph, chains);
}
#ifdef _OPENMP
free(annotationOffset);
annotationOffset = NULL;
#endif
}
static void createPreMarkers(RoadMapArray * rdmaps, PreGraph * preGraph,
IDnum * chains)
{
IDnum sequenceIndex;
IDnum referenceCount = rdmaps->referenceCount;
#ifndef _OPENMP
Annotation *annot = rdmaps->annotations;
#endif
#ifdef _OPENMP
int threads = omp_get_max_threads();
if (threads > 8)
threads = 8;
#pragma omp parallel for num_threads(threads)
#endif
for (sequenceIndex = 1;
sequenceIndex <= referenceCount;
sequenceIndex++) {
#ifdef _OPENMP
Annotation *annot = getAnnotationInArray(rdmaps->annotations, annotationOffset[sequenceIndex - 1]);
#endif
RoadMap *rdmap;
Coordinate currentPosition, currentInternalPosition;
IDnum currentPreNodeID, nextInternalPreNodeID;
IDnum annotIndex, lastAnnotIndex;
PreMarker * previous;
if (sequenceIndex % 1000000 == 0)
velvetLog("Connecting %li / %li\n", (long) sequenceIndex,
(long) sequenceCount_pg(preGraph));
rdmap = getRoadMapInArray(rdmaps, sequenceIndex - 1);
annotIndex = 0;
lastAnnotIndex = getAnnotationCount(rdmap);
nextInternalPreNodeID = chooseNextInternalPreNode
(chains[sequenceIndex] - 1, sequenceIndex,
preGraph, chains);
previous = NULL;
currentPosition = 0;
currentInternalPosition = 0;
currentPreNodeID = 0;
// Recursion up to last annotation
while (annotIndex < lastAnnotIndex
|| nextInternalPreNodeID != 0) {
if (annotIndex == lastAnnotIndex
|| (nextInternalPreNodeID != 0
&& currentInternalPosition <
getPosition(annot))) {
#ifdef _OPENMP
lockNode(nextInternalPreNodeID);
#endif
previous = addPreMarker_pg(preGraph,
nextInternalPreNodeID,
sequenceIndex,
¤tPosition,
previous);
#ifdef _OPENMP
unLockNode(nextInternalPreNodeID);
#endif
currentPreNodeID = nextInternalPreNodeID;
nextInternalPreNodeID =
chooseNextInternalPreNode
(currentPreNodeID, sequenceIndex,
preGraph, chains);
currentInternalPosition +=
getPreNodeLength_pg(currentPreNodeID,
preGraph);
} else {
reConnectAnnotation(¤tPreNodeID, annot,
¤tPosition,
sequenceIndex,
preGraph,
&previous);
annot = getNextAnnotation(annot);
annotIndex++;
}
}
}
}
// Creates the preNode using insertion marker and annotation lists for each sequence
static void
// Creates the preNode using insertion marker and annotation lists for each sequence
createPreNodes(RoadMapArray * rdmaps, PreGraph * preGraph,
IDnum * markerCounters, InsertionMarker * insertionMarkers,
InsertionMarker * veryLastMarker, IDnum * chains,
SequencesReader *seqReadInfo, int WORDLENGTH)
{
char *sequenceFilename = seqReadInfo->m_seqFilename;
Annotation *annot = rdmaps->annotations;
IDnum latestPreNodeID;
InsertionMarker *currentMarker = insertionMarkers;
IDnum sequenceIndex;
Coordinate currentPosition, nextStop;
IDnum preNodeCounter = 1;
FILE *file = NULL;
char line[50000];
int lineLength = 50000;
Coordinate readIndex;
boolean tooShort;
Kmer initialKmer;
char c;
RoadMap *rdmap;
IDnum annotIndex, lastAnnotIndex;
IDnum markerIndex, lastMarkerIndex;
if (!seqReadInfo->m_bIsBinary) {
file = fopen(sequenceFilename, "r");
if (file == NULL)
exitErrorf(EXIT_FAILURE, true, "Could not read %s", sequenceFilename);
// Reading sequence descriptor in first line
if (sequenceCount_pg(preGraph) > 0 && !fgets(line, lineLength, file))
exitErrorf(EXIT_FAILURE, true, "%s incomplete.", sequenceFilename);
seqReadInfo->m_pFile = file;
}
// Now that we have read all of the annotations, we go on to create the preNodes and tie them up
for (sequenceIndex = 1;
sequenceIndex <= sequenceCount_pg(preGraph);
sequenceIndex++) {
if (sequenceIndex % 1000000 == 0)
velvetLog("Sequence %li / %li\n", (long) sequenceIndex,
(long) sequenceCount_pg(preGraph));
if (!seqReadInfo->m_bIsBinary) {
while (line[0] != '>')
if (!fgets(line, lineLength, file))
exitErrorf(EXIT_FAILURE, true, "%s incomplete.", sequenceFilename);
}
rdmap = getRoadMapInArray(rdmaps, sequenceIndex - 1);
annotIndex = 0;
lastAnnotIndex = getAnnotationCount(rdmap);
markerIndex = 0;
lastMarkerIndex = markerCounters[sequenceIndex];
currentPosition = 0;
// Reading first (k-1) nucleotides
tooShort = false;
clearKmer(&initialKmer);
//velvetLog("Initial kmer: ");
TightString *tString = NULL;
char *strString = NULL;
if (seqReadInfo->m_bIsBinary) {
tString = getTightStringInArray(seqReadInfo->m_sequences->tSequences, sequenceIndex - 1);
strString = readTightString(tString);
}
for (readIndex = 0; readIndex < WORDLENGTH - 1;
readIndex++) {
if (seqReadInfo->m_bIsBinary) {
if (readIndex >= tString->length) {
tooShort = true;
break;
}
c = strString[readIndex];
} else {
c = getc(file);
while (c == '\n' || c == '\r')
c = getc(file);
if (c == '>' || c == 'M' || c == EOF) {
ungetc(c, file);
tooShort = true;
break;
}
}
switch (c) {
case 'A':
case 'N':
pushNucleotide(&initialKmer, ADENINE);
break;
case 'C':
pushNucleotide(&initialKmer, CYTOSINE);
break;
case 'G':
pushNucleotide(&initialKmer, GUANINE);
break;
case 'T':
pushNucleotide(&initialKmer, THYMINE);
break;
default:
velvetLog
("Irregular sequence file: are you sure your Sequence and Roadmap file come from the same source?\n");
fflush(stdout);
abort();
}
}
if (tooShort) {
//velvetLog("Skipping short read.. %d\n", sequenceIndex);
chains[sequenceIndex] = preNodeCounter;
if (seqReadInfo->m_bIsBinary) {
free(strString);
} else {
if (!fgets(line, lineLength, file) && sequenceIndex < sequenceCount_pg(preGraph))
exitErrorf(EXIT_FAILURE, true, "%s incomplete.", sequenceFilename);
}
continue;
}
char *currString = NULL;
if (seqReadInfo->m_bIsBinary) {
currString = &strString[readIndex];
seqReadInfo->m_ppCurrString = &currString;
}
latestPreNodeID = 0;
while (annotIndex < lastAnnotIndex) {
if (markerIndex == lastMarkerIndex
|| getPosition(annot) <=
getInsertionMarkerPosition(currentMarker))
nextStop = getPosition(annot);
else {
nextStop =
getInsertionMarkerPosition
(currentMarker);
}
if (currentPosition != nextStop) {
if (seqReadInfo->m_bIsBinary) {
if (readIndex >= tString->length) {
velvetLog("readIndex %ld beyond string len %ld\n", (uint64_t) readIndex, (uint64_t) tString->length);
exit(1);
}
}
//if (sequenceIndex == 481)
// velvetLog("Adding pre nodes from %lli to %lli\n", (long long) currentPosition, (long long) nextStop);
addPreNodeToPreGraph_pg(preGraph,
currentPosition,
nextStop,
seqReadInfo,
&initialKmer,
preNodeCounter);
if (latestPreNodeID == 0) {
chains[sequenceIndex] =
preNodeCounter;
}
latestPreNodeID = preNodeCounter++;
currentPosition = nextStop;
}
while (markerIndex < lastMarkerIndex
&& getInsertionMarkerPosition(currentMarker)
== nextStop) {
convertMarker(currentMarker,
latestPreNodeID);
currentMarker++;
markerIndex++;
}
while (annotIndex < lastAnnotIndex
&& getPosition(annot) == nextStop) {
for (readIndex = 0;
readIndex <
getAnnotationLength(annot);
readIndex++) {
if (seqReadInfo->m_bIsBinary) {
c = *currString;
currString += 1; // increment the pointer
} else {
c = getc(file);
while (!isalpha(c))
c = getc(file);
}
//if (sequenceIndex == 481)
// velvetLog("(%c)", c);
switch (c) {
case 'A':
case 'N':
pushNucleotide(&initialKmer, ADENINE);
break;
case 'C':
pushNucleotide(&initialKmer, CYTOSINE);
break;
case 'G':
pushNucleotide(&initialKmer, GUANINE);
break;
case 'T':
pushNucleotide(&initialKmer, THYMINE);
break;
default:
velvetLog
("Irregular sequence file: are you sure your Sequence and Roadmap file come from the same source?\n");
fflush(stdout);
#ifdef DEBUG
abort();
#endif
exit(1);
}
}
annot = getNextAnnotation(annot);
annotIndex++;
}
}
while (markerIndex < lastMarkerIndex) {
if (currentPosition ==
getInsertionMarkerPosition(currentMarker)) {
convertMarker(currentMarker,
latestPreNodeID);
currentMarker++;
markerIndex++;
} else {
nextStop =
getInsertionMarkerPosition
(currentMarker);
//if (sequenceIndex == 481)
// velvetLog("Adding pre nodes from %lli to %lli\n", (long long) currentPosition, (long long) nextStop);
addPreNodeToPreGraph_pg(preGraph,
currentPosition,
nextStop, seqReadInfo,
&initialKmer,
preNodeCounter);
if (latestPreNodeID == 0)
chains[sequenceIndex] =
preNodeCounter;
latestPreNodeID = preNodeCounter++;
currentPosition =
getInsertionMarkerPosition
(currentMarker);
}
}
if (seqReadInfo->m_bIsBinary) {
free(strString);
} else {
// End of sequence
if (!fgets(line, lineLength, file) && sequenceIndex < sequenceCount_pg(preGraph))
exitErrorf(EXIT_FAILURE, true, "%s incomplete.", sequenceFilename);
//velvetLog(" \n");
}
if (latestPreNodeID == 0)
chains[sequenceIndex] = preNodeCounter;
}
free(markerCounters);
if (!seqReadInfo->m_bIsBinary) {
fclose(file);
}
}
// Counts how many preNodes are to be created to allocate appropriate memory
static void
countPreNodes(RoadMapArray * rdmaps, PreGraph * preGraph,
IDnum * markerCounters, InsertionMarker * insertionMarkers,
InsertionMarker * veryLastMarker)
{
Annotation *annot = rdmaps->annotations;
InsertionMarker *currentMarker = insertionMarkers;
IDnum markerIndex, lastMarkerIndex;
IDnum sequenceIndex;
Coordinate currentPosition, nextStop;
IDnum preNodeCounter = 0;
RoadMap *rdmap;
IDnum annotIndex, lastAnnotIndex;
// Now that we have read all of the annotations, we go on to create the preNodes and tie them up
for (sequenceIndex = 1;
sequenceIndex <= sequenceCount_pg(preGraph);
sequenceIndex++) {
rdmap = getRoadMapInArray(rdmaps, sequenceIndex - 1);
annotIndex = 0;
lastAnnotIndex = getAnnotationCount(rdmap);
markerIndex = 0;
lastMarkerIndex = markerCounters[sequenceIndex];
currentPosition = 0;
while (annotIndex < lastAnnotIndex) {
if (markerIndex == lastMarkerIndex
|| getPosition(annot) <=
getInsertionMarkerPosition(currentMarker))
nextStop = getPosition(annot);
else
nextStop =
getInsertionMarkerPosition
(currentMarker);
if (currentPosition != nextStop) {
preNodeCounter++;
currentPosition = nextStop;
}
while (markerIndex < lastMarkerIndex
&& getInsertionMarkerPosition(currentMarker)
== currentPosition) {
currentMarker++;
markerIndex++;
}
while (annotIndex < lastAnnotIndex
&& getPosition(annot) == currentPosition) {
annot = getNextAnnotation(annot);
annotIndex++;
}
}
while (markerIndex < lastMarkerIndex) {
if (currentPosition ==
getInsertionMarkerPosition(currentMarker)) {
currentMarker++;
markerIndex++;
} else {
preNodeCounter++;
currentPosition =
getInsertionMarkerPosition
(currentMarker);
}
}
}
allocatePreNodeSpace_pg(preGraph, preNodeCounter);
}
// Threads each sequences and creates preArcs according to road map indications
static void connectPreNodes(RoadMapArray * rdmaps, PreGraph * preGraph,
IDnum * chains)
{
Coordinate currentPosition, currentInternalPosition;
IDnum sequenceIndex;
Annotation *annot = rdmaps->annotations;
IDnum referenceCount = rdmaps->referenceCount;
IDnum currentPreNodeID, nextInternalPreNodeID;
RoadMap *rdmap;
IDnum annotIndex, lastAnnotIndex;
boolean isReference;
if (rdmaps->referenceCount > 0)
allocatePreMarkerCountSpace_pg(preGraph);
for (sequenceIndex = 1;
sequenceIndex <= sequenceCount_pg(preGraph);
sequenceIndex++) {
if (sequenceIndex % 100000 == 0)
velvetLog("Connecting %li / %li\n", (long) sequenceIndex,
(long) sequenceCount_pg(preGraph));
rdmap = getRoadMapInArray(rdmaps, sequenceIndex - 1);
annotIndex = 0;
lastAnnotIndex = getAnnotationCount(rdmap);
nextInternalPreNodeID = chooseNextInternalPreNode
(chains[sequenceIndex] - 1, sequenceIndex,
preGraph, chains);
isReference = (sequenceIndex <= referenceCount);
currentPosition = 0;
currentInternalPosition = 0;
currentPreNodeID = 0;
// Recursion up to last annotation
while (annotIndex < lastAnnotIndex
|| nextInternalPreNodeID != 0) {
if (annotIndex == lastAnnotIndex
|| (nextInternalPreNodeID != 0
&& currentInternalPosition <
getPosition(annot))) {
connectPreNodeToTheNext(¤tPreNodeID,
nextInternalPreNodeID,
¤tPosition,
sequenceIndex,
isReference,
preGraph);
nextInternalPreNodeID =
chooseNextInternalPreNode
(currentPreNodeID, sequenceIndex,
preGraph, chains);
currentInternalPosition +=
getPreNodeLength_pg(currentPreNodeID,
preGraph);
} else {
connectAnnotation(¤tPreNodeID, annot,
¤tPosition,
sequenceIndex, isReference,
preGraph);
annot = getNextAnnotation(annot);
annotIndex++;
}
}
}
if (rdmaps->referenceCount > 0) {
allocatePreMarkerSpace_pg(preGraph);
createPreMarkers(rdmaps, preGraph, chains);
}
}
static void createPreMarkers(RoadMapArray * rdmaps, PreGraph * preGraph,
IDnum * chains)
{
Coordinate currentPosition, currentInternalPosition;
IDnum sequenceIndex;
Annotation *annot = rdmaps->annotations;
IDnum referenceCount = rdmaps->referenceCount;
IDnum currentPreNodeID, nextInternalPreNodeID;
RoadMap *rdmap;
IDnum annotIndex, lastAnnotIndex;
PreMarker * previous;
for (sequenceIndex = 1;
sequenceIndex <= referenceCount;
sequenceIndex++) {
if (sequenceIndex % 100000 == 0)
velvetLog("Connecting %li / %li\n", (long) sequenceIndex,
(long) sequenceCount_pg(preGraph));
rdmap = getRoadMapInArray(rdmaps, sequenceIndex - 1);
annotIndex = 0;
lastAnnotIndex = getAnnotationCount(rdmap);
nextInternalPreNodeID = chooseNextInternalPreNode
(chains[sequenceIndex] - 1, sequenceIndex,
preGraph, chains);
previous = NULL;
currentPosition = 0;
currentInternalPosition = 0;
currentPreNodeID = 0;
// Recursion up to last annotation
while (annotIndex < lastAnnotIndex
|| nextInternalPreNodeID != 0) {
if (annotIndex == lastAnnotIndex
|| (nextInternalPreNodeID != 0
&& currentInternalPosition <
getPosition(annot))) {
previous = addPreMarker_pg(preGraph,
nextInternalPreNodeID,
sequenceIndex,
¤tPosition,
previous);
currentPreNodeID = nextInternalPreNodeID;
nextInternalPreNodeID =
chooseNextInternalPreNode
(currentPreNodeID, sequenceIndex,
preGraph, chains);
currentInternalPosition +=
getPreNodeLength_pg(currentPreNodeID,
preGraph);
} else {
reConnectAnnotation(¤tPreNodeID, annot,
¤tPosition,
sequenceIndex,
preGraph,
&previous);
annot = getNextAnnotation(annot);
annotIndex++;
}
}
}
}