unsigned getDistance(Kmer source, Kmer target, unsigned kmer_length)
{
unsigned minDistance = kmer_length;
unsigned dpArray[2][kmer_length];
bzero(dpArray, kmer_length * 2 * sizeof(unsigned));
/*
#ifdef VERBOSE
printf("(dist) source k-mer: "); print_kmer(source, kmer_length); printf("\n");
printf("(dist) target k-mer: "); print_kmer(target, kmer_length); printf("\n");
#endif // VERBOSE
*/
for(unsigned is=0; is < kmer_length; ++is)
{
unsigned *previousRow = &dpArray[((is+1) % 2)][0];
unsigned *currentRow = &dpArray[is % 2][0];
bzero(currentRow, kmer_length * sizeof(int));
for(unsigned it=0; it < kmer_length; ++it)
{
if( getNucleotide(target, it) == getNucleotide(source, is) )
{
/*
#ifdef VERBOSE
printf("(dist) is=%d isN=%c it=%d itN=%c\n", is, CHAR_BASE_MAP[getNucleotide(source,is)], it, CHAR_BASE_MAP[getNucleotide(target,it)]);
#endif // VERBOSE
*/
unsigned previousCount = (it == 0 ? 0 : previousRow[it-1]);
unsigned currentCount = previousCount + 1;
currentRow[it] = currentCount;
// calculate distance
unsigned ms = min(is, kmer_length - is - currentCount);
unsigned mt = (ms < is ? it : kmer_length - it - currentCount);
unsigned distance = ms + (kmer_length - currentCount) + mt;
if( distance < minDistance )
{
/*
#ifdef VERBOSE
printf("previous: "); print_row(previousRow, 21, is-1, source, target);
printf(" current: "); print_row(currentRow, 21, is, source, target);
printf("(dist) is=%d it=%d ms=%d mt=%d len=%d distance=%d\n", is, it, ms, mt, currentCount, distance);
#endif // VERBOSE
*/
minDistance = distance;
}
}
}
}
/*
#ifdef VERBOSE
printf("(dist) minDistance = %d\n", minDistance);
#endif // VERBOSE
*/
return minDistance;
}
static boolean
compareSequences(TightString * sequence1, TightString * sequence2)
{
Coordinate i, j;
Coordinate length1 = getLength(sequence1);
Coordinate length2 = getLength(sequence2);
Coordinate maxLength;
Time Choice1, Choice2, Choice3;
Time maxScore;
if (length1 == 0 || length2 == 0)
return false;
maxLength = (length1 > length2 ? length1 : length2);
if (length1 < WORDLENGTH || length2 < WORDLENGTH)
if (maxLength - length1 > MAXGAPS
|| maxLength - length2 > MAXGAPS)
return false;
for (i = 0; i <= length1; i++)
Fmatrix[i][0] = 0;
for (j = 0; j <= length2; j++)
Fmatrix[0][j] = 0;
for (i = 1; i <= length1; i++) {
for (j = 1; j <= length2; j++) {
Choice1 =
Fmatrix[i - 1][j - 1] +
SIM[(int) getNucleotide(i - 1, sequence1)]
[(int) getNucleotide(j - 1, sequence2)];
Choice2 = Fmatrix[i - 1][j] + INDEL;
Choice3 = Fmatrix[i][j - 1] + INDEL;
Fmatrix[i][j] = max(Choice1, Choice2, Choice3);
}
}
maxScore = Fmatrix[length1][length2];
if (maxScore < maxLength - MAXGAPS)
return false;
if ((1 - maxScore / maxLength) > MAXDIVERGENCE)
return false;
return true;
}
int getDataSingle (char *label, char *data, int position){
int status, i;
struct sequence *seq;
status = 0;
/*
** Find the sequence label.
*/
seq = &(seqList);
while (strcmp (seq->label, label) != 0)
{
if (seq->next == NULL)
{
status = -1;
fprintf (stderr, "Unmatched sequence label in tree: '%s'.\n", label);
goto End_of_Routine;
}
seq = seq->next;
}
/*
** Get the data from the requested columns.
*/
data[0] = seq->sequence[position-1];
data[1] = '\0';
/*
** Remember the observed combination.
*/
i = getNucleotide (data);
if (i == -1)
{
status = -1;
goto End_of_Routine;
}
obsCombs[i] = 1;
End_of_Routine:
/*
** Return the status.
*/
return (status);
}
static TightString *readPositivePassageMarker(PassageMarkerI marker,
TightString ** seqs,
int WORDLENGTH)
{
Coordinate index;
Nucleotide nucleotide;
TightString *tString =
seqs[getPassageMarkerSequenceID(marker) - 1];
TightString *res = newTightString(getPassageMarkerLength(marker));
for (index = 0; index < getLength(tString); index++) {
nucleotide =
getNucleotide(getPassageMarkerStart(marker) + index +
WORDLENGTH - 1, tString);
writeNucleotideAtPosition(nucleotide, index, res);
}
return res;
}
static TightString *readNegativePassageMarker(PassageMarkerI marker,
TightString ** seqs)
{
Coordinate index;
Nucleotide nucleotide;
TightString *tString =
seqs[getAbsolutePassMarkerSeqID(marker) - 1];
TightString *res = newTightString(getPassageMarkerLength(marker));
for (index = 0; index < getPassageMarkerLength(marker); index++) {
nucleotide =
getNucleotide(getPassageMarkerStart(marker) - index,
tString);
#ifndef COLOR
writeNucleotideAtPosition(3 - nucleotide, index, res);
#else
writeNucleotideAtPosition(nucleotide, index, res);
#endif
}
return res;
}
static void threadSequenceThroughGraph(TightString * tString,
KmerOccurenceTable * kmerTable,
Graph * graph,
IDnum seqID, Category category,
boolean readTracking,
boolean double_strand,
ReferenceMapping * referenceMappings,
Coordinate referenceMappingCount,
IDnum refCount,
Annotation * annotations,
IDnum annotationCount,
boolean second_in_pair)
{
Kmer word;
Kmer antiWord;
Coordinate readNucleotideIndex;
Coordinate kmerIndex;
KmerOccurence *kmerOccurence;
int wordLength = getWordLength(graph);
PassageMarkerI marker = NULL_IDX;
PassageMarkerI previousMarker = NULL_IDX;
Node *node = NULL;
Node *previousNode = NULL;
Coordinate coord = 0;
Coordinate previousCoord = 0;
Nucleotide nucleotide;
boolean reversed;
IDnum refID;
Coordinate refCoord = 0;
ReferenceMapping * refMap;
Annotation * annotation = annotations;
Coordinate index = 0;
Coordinate uniqueIndex = 0;
Coordinate annotIndex = 0;
IDnum annotCount = 0;
SmallNodeList * nodePile = NULL;
// Neglect any string shorter than WORDLENGTH :
if (getLength(tString) < wordLength)
return;
clearKmer(&word);
clearKmer(&antiWord);
// Fill in the initial word :
for (readNucleotideIndex = 0;
readNucleotideIndex < wordLength - 1; readNucleotideIndex++) {
nucleotide = getNucleotide(readNucleotideIndex, tString);
pushNucleotide(&word, nucleotide);
if (double_strand || second_in_pair) {
#ifdef COLOR
reversePushNucleotide(&antiWord, nucleotide);
#else
reversePushNucleotide(&antiWord, 3 - nucleotide);
#endif
}
}
// Go through sequence
while (readNucleotideIndex < getLength(tString)) {
nucleotide = getNucleotide(readNucleotideIndex++, tString);
pushNucleotide(&word, nucleotide);
if (double_strand || second_in_pair) {
#ifdef COLOR
reversePushNucleotide(&antiWord, nucleotide);
#else
reversePushNucleotide(&antiWord, 3 - nucleotide);
#endif
}
// Update annotation if necessary
if (annotCount < annotationCount && annotIndex == getAnnotationLength(annotation)) {
annotation = getNextAnnotation(annotation);
annotCount++;
annotIndex = 0;
}
// Search for reference mapping
if (category == REFERENCE) {
if (referenceMappings)
refMap = findReferenceMapping(seqID, index, referenceMappings, referenceMappingCount);
else
refMap = NULL;
if (refMap) {
node = getNodeInGraph(graph, refMap->nodeID);
if (refMap->nodeID > 0) {
coord = refMap->nodeStart + (index - refMap->referenceStart);
} else {
coord = getNodeLength(node) - refMap->nodeStart - refMap->length + (index - refMap->referenceStart);
}
} else {
node = NULL;
if (previousNode)
break;
}
}
// Search for reference-based mapping
else if (annotCount < annotationCount && uniqueIndex >= getPosition(annotation) && getAnnotSequenceID(annotation) <= refCount && getAnnotSequenceID(annotation) >= -refCount) {
refID = getAnnotSequenceID(annotation);
if (refID > 0)
refCoord = getStart(annotation) + annotIndex;
else
refCoord = getStart(annotation) - annotIndex;
refMap = findReferenceMapping(refID, refCoord, referenceMappings, referenceMappingCount);
// If success
if (refMap) {
if (refID > 0) {
node = getNodeInGraph(graph, refMap->nodeID);
if (refMap->nodeID > 0) {
coord = refMap->nodeStart + (refCoord - refMap->referenceStart);
} else {
coord = getNodeLength(node) - refMap->nodeStart - refMap->length + (refCoord - refMap->referenceStart);
}
} else {
node = getNodeInGraph(graph, -refMap->nodeID);
if (refMap->nodeID > 0) {
coord = getNodeLength(node) - refMap->nodeStart - (refCoord - refMap->referenceStart) - 1;
} else {
coord = refMap->nodeStart + refMap->length - (refCoord - refMap->referenceStart) - 1;
}
}
} else {
node = NULL;
if (previousNode)
break;
}
}
// Search in table
else {
reversed = false;
if (double_strand) {
if (compareKmers(&word, &antiWord) <= 0) {
kmerOccurence =
findKmerInKmerOccurenceTable(&word,
kmerTable);
} else {
kmerOccurence =
findKmerInKmerOccurenceTable(&antiWord,
kmerTable);
reversed = true;
}
} else {
if (!second_in_pair) {
kmerOccurence =
findKmerInKmerOccurenceTable(&word,
kmerTable);
} else {
kmerOccurence =
findKmerInKmerOccurenceTable(&antiWord,
kmerTable);
reversed = true;
}
}
if (kmerOccurence) {
if (!reversed) {
node = getNodeInGraph(graph, getKmerOccurenceNodeID(kmerOccurence));
coord = getKmerOccurencePosition(kmerOccurence);
} else {
node = getNodeInGraph(graph, -getKmerOccurenceNodeID(kmerOccurence));
coord = getNodeLength(node) - getKmerOccurencePosition(kmerOccurence) - 1;
}
} else {
node = NULL;
if (previousNode)
break;
}
}
// Increment positions
if (annotCount < annotationCount && uniqueIndex >= getPosition(annotation))
annotIndex++;
else
uniqueIndex++;
// Fill in graph
if (node)
{
#ifdef OPENMP
lockNode(node);
#endif
kmerIndex = readNucleotideIndex - wordLength;
if (previousNode == node
&& previousCoord == coord - 1) {
if (category / 2 >= CATEGORIES) {
setPassageMarkerFinish(marker,
kmerIndex +
1);
setFinishOffset(marker,
getNodeLength(node)
- coord - 1);
} else {
#ifndef SINGLE_COV_CAT
incrementVirtualCoverage(node, category / 2, 1);
incrementOriginalVirtualCoverage(node, category / 2, 1);
#else
incrementVirtualCoverage(node, 1);
#endif
}
#ifdef OPENMP
unLockNode(node);
#endif
} else {
if (category / 2 >= CATEGORIES) {
marker =
newPassageMarker(seqID,
kmerIndex,
kmerIndex + 1,
coord,
getNodeLength
(node) -
coord - 1);
transposePassageMarker(marker,
node);
connectPassageMarkers
(previousMarker, marker,
graph);
previousMarker = marker;
} else {
if (readTracking) {
if (!isNodeMemorized(node, nodePile)) {
addReadStart(node,
seqID,
coord,
graph,
kmerIndex);
memorizeNode(node, &nodePile);
} else {
blurLastShortReadMarker
(node, graph);
}
}
#ifndef SINGLE_COV_CAT
incrementVirtualCoverage(node, category / 2, 1);
incrementOriginalVirtualCoverage(node, category / 2, 1);
#else
incrementVirtualCoverage(node, 1);
#endif
}
#ifdef OPENMP
lockTwoNodes(node, previousNode);
#endif
createArc(previousNode, node, graph);
#ifdef OPENMP
unLockTwoNodes(node, previousNode);
#endif
}
previousNode = node;
previousCoord = coord;
}
index++;
}
if (readTracking && category / 2 < CATEGORIES)
unMemorizeNodes(&nodePile);
}
static void ghostThreadSequenceThroughGraph(TightString * tString,
KmerOccurenceTable *
kmerTable, Graph * graph,
IDnum seqID, Category category,
boolean readTracking,
boolean double_strand,
ReferenceMapping * referenceMappings,
Coordinate referenceMappingCount,
IDnum refCount,
Annotation * annotations,
IDnum annotationCount,
boolean second_in_pair)
{
Kmer word;
Kmer antiWord;
Coordinate readNucleotideIndex;
KmerOccurence *kmerOccurence;
int wordLength = getWordLength(graph);
Nucleotide nucleotide;
IDnum refID;
Coordinate refCoord;
ReferenceMapping * refMap = NULL;
Coordinate uniqueIndex = 0;
Coordinate annotIndex = 0;
IDnum annotCount = 0;
boolean reversed;
SmallNodeList * nodePile = NULL;
Annotation * annotation = annotations;
Node *node;
Node *previousNode = NULL;
// Neglect any read which will not be short paired
if ((!readTracking && category % 2 == 0)
|| category / 2 >= CATEGORIES)
return;
// Neglect any string shorter than WORDLENGTH :
if (getLength(tString) < wordLength)
return;
// Verify that all short reads are reasonnably short
if (getLength(tString) > USHRT_MAX) {
velvetLog("Short read of length %lli, longer than limit %i\n",
(long long) getLength(tString), SHRT_MAX);
velvetLog("You should better declare this sequence as long, because it genuinely is!\n");
exit(1);
}
clearKmer(&word);
clearKmer(&antiWord);
// Fill in the initial word :
for (readNucleotideIndex = 0;
readNucleotideIndex < wordLength - 1; readNucleotideIndex++) {
nucleotide = getNucleotide(readNucleotideIndex, tString);
pushNucleotide(&word, nucleotide);
if (double_strand || second_in_pair) {
#ifdef COLOR
reversePushNucleotide(&antiWord, nucleotide);
#else
reversePushNucleotide(&antiWord, 3 - nucleotide);
#endif
}
}
// Go through sequence
while (readNucleotideIndex < getLength(tString)) {
// Shift word:
nucleotide = getNucleotide(readNucleotideIndex++, tString);
pushNucleotide(&word, nucleotide);
if (double_strand || second_in_pair) {
#ifdef COLOR
reversePushNucleotide(&antiWord, nucleotide);
#else
reversePushNucleotide(&antiWord, 3 - nucleotide);
#endif
}
// Update annotation if necessary
if (annotCount < annotationCount && annotIndex == getAnnotationLength(annotation)) {
annotation = getNextAnnotation(annotation);
annotCount++;
annotIndex = 0;
}
// Search for reference mapping
if (annotCount < annotationCount && uniqueIndex >= getPosition(annotation) && getAnnotSequenceID(annotation) <= refCount && getAnnotSequenceID(annotation) >= -refCount) {
refID = getAnnotSequenceID(annotation);
if (refID > 0)
refCoord = getStart(annotation) + annotIndex;
else
refCoord = getStart(annotation) - annotIndex;
refMap = findReferenceMapping(refID, refCoord, referenceMappings, referenceMappingCount);
// If success
if (refMap) {
if (refID > 0)
node = getNodeInGraph(graph, refMap->nodeID);
else
node = getNodeInGraph(graph, -refMap->nodeID);
} else {
node = NULL;
if (previousNode)
break;
}
}
// if not.. look in table
else {
reversed = false;
if (double_strand) {
if (compareKmers(&word, &antiWord) <= 0) {
kmerOccurence =
findKmerInKmerOccurenceTable(&word,
kmerTable);
} else {
kmerOccurence =
findKmerInKmerOccurenceTable(&antiWord,
kmerTable);
reversed = true;
}
} else {
if (!second_in_pair) {
kmerOccurence =
findKmerInKmerOccurenceTable(&word,
kmerTable);
} else {
kmerOccurence =
findKmerInKmerOccurenceTable(&antiWord,
kmerTable);
reversed = true;
}
}
if (kmerOccurence) {
if (!reversed)
node = getNodeInGraph(graph, getKmerOccurenceNodeID(kmerOccurence));
else
node = getNodeInGraph(graph, -getKmerOccurenceNodeID(kmerOccurence));
} else {
node = NULL;
if (previousNode)
break;
}
}
if (annotCount < annotationCount && uniqueIndex >= getPosition(annotation))
annotIndex++;
else
uniqueIndex++;
previousNode = node;
// Fill in graph
if (node && !isNodeMemorized(node, nodePile))
{
#ifdef OPENMP
lockNode(node);
#endif
incrementReadStartCount(node, graph);
#ifdef OPENMP
unLockNode(node);
#endif
memorizeNode(node, &nodePile);
}
}
unMemorizeNodes(&nodePile);
}
static void threadSequenceThroughGraph(TightString * tString,
KmerOccurenceTable * kmerOccurences,
Graph * graph,
IDnum seqID, Category category,
boolean readTracking,
boolean double_strand)
{
Kmer word;
Kmer antiWord;
Coordinate readNucleotideIndex;
Coordinate kmerIndex;
KmerOccurence *kmerOccurence;
int wordLength = getWordLength(graph);
PassageMarker *marker = NULL;
PassageMarker *previousMarker = NULL;
Node *node;
Node *previousNode = NULL;
Coordinate coord;
Coordinate previousCoord = 0;
Nucleotide nucleotide;
clearKmer(&word);
clearKmer(&antiWord);
// Neglect any string shorter than WORDLENGTH :
if (getLength(tString) < wordLength)
return;
// Fill in the initial word :
for (readNucleotideIndex = 0;
readNucleotideIndex < wordLength - 1; readNucleotideIndex++) {
nucleotide = getNucleotide(readNucleotideIndex, tString);
pushNucleotide(&word, nucleotide);
if (double_strand) {
#ifdef COLOR
reversePushNucleotide(&antiWord, nucleotide);
#else
reversePushNucleotide(&antiWord, 3 - nucleotide);
#endif
}
}
// Go through sequence
while (readNucleotideIndex < getLength(tString)) {
nucleotide = getNucleotide(readNucleotideIndex++, tString);
pushNucleotide(&word, nucleotide);
if (double_strand) {
#ifdef COLOR
reversePushNucleotide(&antiWord, nucleotide);
#else
reversePushNucleotide(&antiWord, 3 - nucleotide);
#endif
}
// Search in table
if ((!double_strand || compareKmers(&word, &antiWord) <= 0)
&& (kmerOccurence =
findKmerOccurenceInSortedTable(&word,
kmerOccurences))) {
node =
getNodeInGraph(graph, kmerOccurence->nodeID);
coord = kmerOccurence->position;
} else if ((double_strand && compareKmers(&word, &antiWord) > 0)
&& (kmerOccurence =
findKmerOccurenceInSortedTable(&antiWord,
kmerOccurences)))
{
node =
getNodeInGraph(graph, -kmerOccurence->nodeID);
coord =
getNodeLength(node) - kmerOccurence->position -
1;
} else {
node = NULL;
if (previousNode) {
break;
}
}
// Fill in graph
if (node) {
kmerIndex = readNucleotideIndex - wordLength;
if (previousNode == node
&& previousCoord == coord - 1) {
if (category / 2 >= CATEGORIES) {
setPassageMarkerFinish(marker,
kmerIndex +
1);
setFinishOffset(marker,
getNodeLength(node)
- coord - 1);
} else {
incrementVirtualCoverage(node,
category /
2, 1);
incrementOriginalVirtualCoverage
(node, category / 2, 1);
}
} else {
if (category / 2 >= CATEGORIES) {
marker =
newPassageMarker(seqID,
kmerIndex,
kmerIndex + 1,
coord,
getNodeLength
(node) -
coord - 1);
transposePassageMarker(marker,
node);
connectPassageMarkers
(previousMarker, marker,
graph);
previousMarker = marker;
} else {
if (readTracking) {
if (!getNodeStatus(node)) {
addReadStart(node,
seqID,
coord,
graph,
kmerIndex);
setSingleNodeStatus
(node, true);
memorizeNode(node);
} else {
blurLastShortReadMarker
(node, graph);
}
}
incrementVirtualCoverage(node,
category /
2, 1);
incrementOriginalVirtualCoverage
(node, category / 2, 1);
}
createArc(previousNode, node, graph);
}
previousNode = node;
previousCoord = coord;
}
}
unlockMemorizedNodes();
}
static void ghostThreadSequenceThroughGraph(TightString * tString,
KmerOccurenceTable *
kmerOccurences, Graph * graph,
IDnum seqID, Category category,
boolean readTracking,
boolean double_strand)
{
Kmer word;
Kmer antiWord;
Coordinate readNucleotideIndex;
KmerOccurence *kmerOccurence;
int wordLength = getWordLength(graph);
Nucleotide nucleotide;
Node *node;
Node *previousNode = NULL;
clearKmer(&word);
clearKmer(&antiWord);
// Neglect any read which will not be short paired
if ((!readTracking && category % 2 == 0)
|| category / 2 >= CATEGORIES)
return;
// Neglect any string shorter than WORDLENGTH :
if (getLength(tString) < wordLength)
return;
// Verify that all short reads are reasonnably short
if (getLength(tString) > USHRT_MAX) {
printf("Short read of length %lli, longer than limit %i\n",
(long long) getLength(tString), SHRT_MAX);
puts("You should better declare this sequence as long, because it genuinely is!");
exit(1);
}
// Allocate memory for the read pairs
if (!readStartsAreActivated(graph))
activateReadStarts(graph);
// Fill in the initial word :
for (readNucleotideIndex = 0;
readNucleotideIndex < wordLength - 1; readNucleotideIndex++) {
nucleotide = getNucleotide(readNucleotideIndex, tString);
pushNucleotide(&word, nucleotide);
if (double_strand) {
#ifdef COLOR
reversePushNucleotide(&antiWord, nucleotide);
#else
reversePushNucleotide(&antiWord, 3 - nucleotide);
#endif
}
}
// Go through sequence
while (readNucleotideIndex < getLength(tString)) {
// Shift word:
nucleotide = getNucleotide(readNucleotideIndex++, tString);
pushNucleotide(&word, nucleotide);
if (double_strand) {
#ifdef COLOR
reversePushNucleotide(&antiWord, nucleotide);
#else
reversePushNucleotide(&antiWord, 3 - nucleotide);
#endif
}
// Search in table
if ((!double_strand || compareKmers(&word, &antiWord) <= 0)
&& (kmerOccurence =
findKmerOccurenceInSortedTable(&word,
kmerOccurences))) {
node =
getNodeInGraph(graph, kmerOccurence->nodeID);
} else if ((double_strand && compareKmers(&word, &antiWord) > 0)
&& (kmerOccurence =
findKmerOccurenceInSortedTable(&antiWord,
kmerOccurences)))
{
node =
getNodeInGraph(graph, -kmerOccurence->nodeID);
} else {
node = NULL;
if (previousNode)
break;
}
previousNode = node;
// Fill in graph
if (node && !getNodeStatus(node)) {
incrementReadStartCount(node, graph);
setSingleNodeStatus(node, true);
memorizeNode(node);
}
}
unlockMemorizedNodes();
}