static void getMetaSequencesForEventsP(stSortedSet *metaSequences,
Flower *flower, stList *eventStrings) {
//Iterate over the sequences in the flower.
Flower_SequenceIterator *seqIt = flower_getSequenceIterator(flower);
Sequence *sequence;
while ((sequence = flower_getNextSequence(seqIt)) != NULL) {
MetaSequence *metaSequence = sequence_getMetaSequence(sequence);
if (stringIsInList(event_getHeader(sequence_getEvent(sequence)),
eventStrings) == 0) {
if (stSortedSet_search(metaSequences, metaSequence) == NULL) {
stSortedSet_insert(metaSequences, metaSequence);
}
}
}
flower_destructSequenceIterator(seqIt);
//Recurse over the flowers
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (group_getNestedFlower(group) != NULL) {
getMetaSequencesForEventsP(metaSequences,
group_getNestedFlower(group), eventStrings);
}
}
flower_destructGroupIterator(groupIt);
}
static void getMaximalHaplotypePathsCheck(Flower *flower,
stSortedSet *segmentSet, const char *eventString, stList *eventStrings) {
/*
* Do debug checks that the haplotypes paths are well formed.
*/
Flower_SegmentIterator *segmentIt = flower_getSegmentIterator(flower);
Segment *segment;
while ((segment = flower_getNextSegment(segmentIt)) != NULL) {
if (strcmp(event_getHeader(segment_getEvent(segment)), eventString) == 0) {
if (hasCapInEvents(cap_getEnd(segment_get5Cap(segment)), eventStrings)) { //isHaplotypeEnd(cap_getEnd(segment_get5Cap(segment)))) {
assert(stSortedSet_search(segmentSet, segment) != NULL
|| stSortedSet_search(segmentSet, segment_getReverse(
segment)) != NULL);
}
}
}
flower_destructSegmentIterator(segmentIt);
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (group_getNestedFlower(group) != NULL) {
getMaximalHaplotypePathsCheck(group_getNestedFlower(group),
segmentSet, eventString, eventStrings);
}
}
flower_destructGroupIterator(groupIt);
}
static void recoverBrokenAdjacencies(Flower *flower, stList *recoveredCaps, Name referenceEventName) {
/*
* Find reference intervals that are book-ended by stubs created in a child flower.
*/
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while((group = flower_getNextGroup(groupIt)) != NULL) {
Flower *nestedFlower;
if((nestedFlower = group_getNestedFlower(group)) != NULL) {
Flower_EndIterator *endIt = flower_getEndIterator(nestedFlower);
End *childEnd;
while((childEnd = flower_getNextEnd(endIt)) != NULL) {
if(end_isStubEnd(childEnd) && flower_getEnd(flower, end_getName(childEnd)) == NULL) { //We have a thread we need to promote
Cap *childCap = getCapForReferenceEvent(childEnd, referenceEventName); //The cap in the reference
assert(childCap != NULL);
assert(!end_isAttached(childEnd));
childCap = cap_getStrand(childCap) ? childCap : cap_getReverse(childCap);
if (!cap_getSide(childCap)) {
Cap *adjacentChildCap = NULL;
int64_t adjacencyLength = traceThreadLength(childCap, &adjacentChildCap);
Cap *cap = copyCapToParent(childCap, recoveredCaps);
assert(adjacentChildCap != NULL);
assert(!end_isAttached(cap_getEnd(adjacentChildCap)));
assert(!cap_getSide(cap));
Cap *adjacentCap = copyCapToParent(adjacentChildCap, recoveredCaps);
cap_makeAdjacent(cap, adjacentCap);
setAdjacencyLength(cap, adjacentCap, adjacencyLength);
}
}
}
flower_destructEndIterator(endIt);
}
}
flower_destructGroupIterator(groupIt);
}
void testGroup_serialisation(CuTest* testCase) {
cactusGroupTestSetup();
int64_t i;
Name name = group_getName(group);
void
*vA =
binaryRepresentation_makeBinaryRepresentation(group,
(void(*)(void *, void(*)(const void *, size_t,
size_t))) group_writeBinaryRepresentation,
&i);
CuAssertTrue(testCase, i > 0);
group_destruct(group);
void *vA2 = vA;
group = group_loadFromBinaryRepresentation(&vA2, flower);
free(vA);
CuAssertTrue(testCase, group_getName(group) == name);
CuAssertTrue(testCase, group_getFlower(group) == flower);
CuAssertTrue(testCase, group_getNestedFlower(group) == nestedFlower);
Group_EndIterator *iterator = group_getEndIterator(group);
CuAssertTrue(testCase, group_getNextEnd(iterator) == end1);
CuAssertTrue(testCase, group_getNextEnd(iterator) == end2);
CuAssertTrue(testCase, group_getNextEnd(iterator) == NULL);
group_destructEndIterator(iterator);
cactusGroupTestTeardown();
}
void makeCactusTree_chain(Chain *chain, FILE *fileHandle,
const char *parentNodeName, const char *parentEdgeColour) {
//Write the flower nodes.
char *chainNameString = cactusMisc_nameToString(chain_getName(chain));
const char *edgeColour = graphViz_getColour();
addNodeToGraph(chainNameString, fileHandle,
chain_getAverageInstanceBaseLength(chain) / totalProblemSize,
"box", chainNameString);
//Write in the parent edge.
if (parentNodeName != NULL) {
graphViz_addEdgeToGraph(parentNodeName, chainNameString, fileHandle,
"", parentEdgeColour, 10, 1, "forward");
}
//Create the linkers to the nested flowers.
Link *link = chain_getFirst(chain);
while(link != NULL) {
Group *group = link_getGroup(link);
assert(group != NULL);
assert(!group_isLeaf(group));
if (!group_isLeaf(group)) {
makeCactusTree_flower(group_getNestedFlower(group), fileHandle,
chainNameString, edgeColour);
}
link = link_getNextLink(link);
}
free(chainNameString);
}
void flower_destruct(Flower *flower, int64_t recursive) {
Flower_GroupIterator *iterator;
Sequence *sequence;
End *end;
Block *block;
Group *group;
Chain *chain;
Flower *nestedFlower;
if (recursive) {
iterator = flower_getGroupIterator(flower);
while ((group = flower_getNextGroup(iterator)) != NULL) {
nestedFlower = group_getNestedFlower(group);
if (nestedFlower != NULL) {
flower_destruct(nestedFlower, recursive);
}
}
flower_destructGroupIterator(iterator);
}
cactusDisk_removeFlower(flower->cactusDisk, flower);
flower_destructFaces(flower);
stSortedSet_destruct(flower->faces);
assert(flower_getEventTree(flower) != NULL);
eventTree_destruct(flower_getEventTree(flower));
while ((sequence = flower_getFirstSequence(flower)) != NULL) {
sequence_destruct(sequence);
}
stSortedSet_destruct(flower->sequences);
while ((chain = flower_getFirstChain(flower)) != NULL) {
chain_destruct(chain);
}
stSortedSet_destruct(flower->chains);
while ((end = flower_getFirstEnd(flower)) != NULL) {
end_destruct(end);
}
stSortedSet_destruct(flower->caps);
stSortedSet_destruct(flower->ends);
while ((block = flower_getFirstBlock(flower)) != NULL) {
block_destruct(block);
}
stSortedSet_destruct(flower->segments);
stSortedSet_destruct(flower->blocks);
while ((group = flower_getFirstGroup(flower)) != NULL) {
group_destruct(group);
}
stSortedSet_destruct(flower->groups);
free(flower);
}
Cap *getTerminalCap(Cap *cap) {
Flower *nestedFlower = group_getNestedFlower(end_getGroup(cap_getEnd(cap)));
if (nestedFlower != NULL) {
Cap *nestedCap = flower_getCap(nestedFlower, cap_getName(cap));
assert(nestedCap != NULL);
return getTerminalCap(cap_getOrientation(cap) ? nestedCap : cap_getReverse(nestedCap));
}
return cap;
}
void flower_checkRecursive(Flower *flower) {
flower_check(flower);
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (!group_isLeaf(group)) {
flower_checkRecursive(group_getNestedFlower(group));
}
}
flower_destructGroupIterator(groupIt);
}
static void getMaximalHaplotypePathsP(Flower *flower,
stList *maximalHaplotypePaths, stSortedSet *segmentSet,
const char *eventString,
stList *eventStrings) {
/*
* Iterate through the segments in this flower.
*/
Flower_SegmentIterator *segmentIt = flower_getSegmentIterator(flower);
Segment *segment;
while ((segment = flower_getNextSegment(segmentIt)) != NULL) {
if (stSortedSet_search(segmentSet, segment) == NULL
&& stSortedSet_search(segmentSet, segment_getReverse(segment))
== NULL) { //Check we haven't yet seen this segment
if (strcmp(event_getHeader(segment_getEvent(segment)), eventString)
== 0) { //Check if the segment is in the assembly
if (hasCapInEvents(cap_getEnd(segment_get5Cap(segment)), eventStrings)) { //Is a block in a haplotype segment
assert(hasCapInEvents(cap_getEnd(segment_get3Cap(segment)), eventStrings)); //isHaplotypeEnd(cap_getEnd(segment_get3Cap(segment))));
stList *maximalHaplotypePath = stList_construct();
stList_append(maximalHaplotypePaths, maximalHaplotypePath);
getMaximalHaplotypePathsP2(segment, maximalHaplotypePath,
segmentSet, eventStrings);
} else {
assert(!hasCapInEvents(cap_getEnd(segment_get3Cap(segment)), eventStrings));//assert(!isHaplotypeEnd(cap_getEnd(segment_get3Cap(segment))));
}
}
}
}
flower_destructSegmentIterator(segmentIt);
/*
* Now recurse on the contained flowers.
*/
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (group_getNestedFlower(group) != NULL) {
getMaximalHaplotypePathsP(group_getNestedFlower(group),
maximalHaplotypePaths, segmentSet, eventString, eventStrings);
}
}
flower_destructGroupIterator(groupIt);
}
static void setAdjacencyLengthsAndRecoverNewCapsAndBrokenAdjacencies(Cap *cap, stList *recoveredCaps) {
/*
* Sets the coordinates of the caps to be equal to the length of the adjacency sequence between them.
* Used to build the reference sequence bottom up.
*
* One complexity is that a reference thread between the two caps
* in each flower f may be broken into two in the children of f.
* Therefore, for each flower f first identify attached stub ends present in the children of f that are
* not present in f and copy them into f, reattaching the reference caps as needed.
*/
while (1) {
Cap *adjacentCap = cap_getAdjacency(cap);
assert(adjacentCap != NULL);
assert(cap_getCoordinate(cap) == INT64_MAX);
assert(cap_getCoordinate(adjacentCap) == INT64_MAX);
assert(cap_getStrand(cap) == cap_getStrand(adjacentCap));
assert(cap_getSide(cap) != cap_getSide(adjacentCap));
Group *group = end_getGroup(cap_getEnd(cap));
assert(group != NULL);
if (!group_isLeaf(group)) { //Adjacency is not terminal, so establish its sequence.
Flower *nestedFlower = group_getNestedFlower(group);
Cap *nestedCap = flower_getCap(nestedFlower, cap_getName(cap));
assert(nestedCap != NULL);
Cap *nestedAdjacentCap = flower_getCap(nestedFlower, cap_getName(adjacentCap));
assert(nestedAdjacentCap != NULL);
Cap *breakerCap;
int64_t adjacencyLength = traceThreadLength(nestedCap, &breakerCap);
assert(cap_getOrientation(nestedAdjacentCap));
if (cap_getPositiveOrientation(breakerCap) != nestedAdjacentCap) { //The thread is broken at the lower level.
//Copy cap into higher level graph.
breakerCap = copyCapToParent(breakerCap, recoveredCaps);
assert(cap_getSide(breakerCap));
cap_makeAdjacent(cap, breakerCap);
setAdjacencyLength(cap, breakerCap, adjacencyLength);
adjacencyLength = traceThreadLength(nestedAdjacentCap, &breakerCap);
assert(cap_getPositiveOrientation(breakerCap) != cap);
breakerCap = copyCapToParent(breakerCap, recoveredCaps);
assert(!cap_getSide(breakerCap));
cap_makeAdjacent(breakerCap, adjacentCap);
setAdjacencyLength(adjacentCap, breakerCap, adjacencyLength);
} else { //The thread is not broken at the lower level
setAdjacencyLength(cap, adjacentCap, adjacencyLength);
}
} else {
//Set the coordinates of the caps to the adjacency size
setAdjacencyLength(cap, adjacentCap, 0);
}
if ((cap = cap_getOtherSegmentCap(adjacentCap)) == NULL) {
break;
}
}
}
static void getOrderedSegmentsP(Flower *flower,
stSortedSet *segments) {
Flower_SegmentIterator *segmentIt = flower_getSegmentIterator(flower);
Segment *segment;
while ((segment = flower_getNextSegment(segmentIt)) != NULL) {
if (!segment_getStrand(segment)) {
segment = segment_getReverse(segment);
}
assert(stSortedSet_search(segments, segment) == NULL);
stSortedSet_insert(segments, segment);
}
flower_destructSegmentIterator(segmentIt);
//Recurse over the flowers
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (group_getNestedFlower(group) != NULL) {
getOrderedSegmentsP(group_getNestedFlower(group),
segments);
}
}
flower_destructGroupIterator(groupIt);
}
bool flower_removeIfRedundant(Flower *flower) {
if (!flower_isLeaf(flower) && flower_getParentGroup(flower) != NULL && flower_getBlockNumber(flower) == 0) { //We will remove this flower..
Group *parentGroup = flower_getParentGroup(flower); //This group will be destructed
//Deal with any parent chain..
if (group_isLink(parentGroup)) {
link_split(group_getLink(parentGroup));
}
Flower *parentFlower = group_getFlower(parentGroup); //We will add the groups in the flower to the parent
/*
* For each group in the flower we take its nested flower and attach it to the parent.
*/
Group *group;
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (!group_isLeaf(group)) {
//Copy the group into the parent..
Flower *nestedFlower = group_getNestedFlower(group);
assert(nestedFlower != NULL);
Group *newParentGroup = group_construct(parentFlower, nestedFlower);
flower_setParentGroup(nestedFlower, newParentGroup);
group_constructChainForLink(newParentGroup);
} else {
Group *newParentGroup = group_construct2(parentFlower);
End *end;
Group_EndIterator *endIt = group_getEndIterator(group);
while ((end = group_getNextEnd(endIt)) != NULL) {
End *parentEnd = flower_getEnd(parentFlower, end_getName(end));
assert(parentEnd != NULL);
end_setGroup(parentEnd, newParentGroup);
}
group_destructEndIterator(endIt);
group_constructChainForLink(newParentGroup);
}
}
flower_destructGroupIterator(groupIt);
//The group attached to the flower should now be empty
assert(group_getEndNumber(parentGroup) == 0);
group_destruct(parentGroup);
//Now wipe the flower out..
cactusDisk_deleteFlowerFromDisk(flower_getCactusDisk(flower), flower);
flower_destruct(flower, 0);
return 1;
}
return 0;
}
void flower_checkNotEmpty(Flower *flower, bool recursive) {
//First check the flower is not empty, unless it is the parent group.
if (flower_hasParentGroup(flower)) {
assert(flower_getGroupNumber(flower) > 0);
assert(flower_getEndNumber(flower) > 0);
assert(flower_getAttachedStubEndNumber(flower) > 0); //We must have some ends to tie us into the parent problem + flower_getBlockEndNumber(flower) > 0);
}
//Now Checks that each group contains at least one end and call recursive.
Group *group;
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
while ((group = flower_getNextGroup(groupIt)) != NULL) {
assert(group_getEndNumber(group) > 0);
assert(group_getAttachedStubEndNumber(group) + group_getBlockEndNumber(group) > 0);
if (recursive && !group_isLeaf(group)) {
flower_checkNotEmpty(group_getNestedFlower(group), 1);
}
}
flower_destructGroupIterator(groupIt);
}
void flower_delete2(Flower *flower, bool isOnDisk) {
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (!group_isLeaf(group)) {
flower_delete2(group_getNestedFlower(group), isOnDisk);
}
}
flower_destructGroupIterator(groupIt);
Group *parentGroup = flower_getParentGroup(flower);
if(parentGroup != NULL) {
parentGroup->leafGroup = 1;
}
//This needs modification so that we don't do this directly..
if(isOnDisk) {
cactusDisk_deleteFlowerFromDisk(flower_getCactusDisk(flower), flower);
}
flower_destruct(flower, 0);
}
void testGroup_makeNonLeaf(CuTest *testCase) {
cactusGroupTestSetup();
CuAssertTrue(testCase, group_isLeaf(group2));
end_setGroup(end4, group2);
group_makeNestedFlower(group2);
CuAssertTrue(testCase, !group_isLeaf(group2));
Flower *nestedFlower = group_getNestedFlower(group2);
CuAssertTrue(testCase, nestedFlower != NULL);
CuAssertTrue(testCase, !flower_builtBlocks(flower));
CuAssertTrue(testCase, !flower_builtTrees(flower));
CuAssertTrue(testCase, !flower_builtFaces(flower));
CuAssertTrue(testCase, flower_getName(nestedFlower) == group_getName(group2));
CuAssertTrue(testCase, flower_getParentGroup(nestedFlower) == group2);
CuAssertTrue(testCase, flower_getEndNumber(nestedFlower) == 1);
End *nestedEnd = flower_getFirstEnd(nestedFlower);
CuAssertTrue(testCase, end_getName(end4) == end_getName(nestedEnd));
CuAssertTrue(testCase, end_getGroup(nestedEnd) != NULL);
CuAssertTrue(testCase, flower_getGroupNumber(nestedFlower) == 1);
CuAssertTrue(testCase, flower_isTerminal(nestedFlower));
cactusGroupTestTeardown();
}
bool flower_deleteIfEmpty(Flower *flower) {
if (flower_getEndNumber(flower) == 0 && flower_getParentGroup(flower) != NULL) { //contains nothing useful..
assert(flower_getChainNumber(flower) == 0);
assert(flower_getBlockNumber(flower) == 0);
while (flower_getGroupNumber(flower) > 0) {
Group *group = flower_getFirstGroup(flower);
if (!group_isLeaf(group)) {
bool i = flower_deleteIfEmpty(group_getNestedFlower(group));
(void) i;
assert(i);
}
}
assert(flower_getGroupNumber(flower) == 0);
//This needs modification so that we don't do this directly..
cactusDisk_deleteFlowerFromDisk(flower_getCactusDisk(flower), flower);
Group *parentGroup = flower_getParentGroup(flower);
group_destruct(parentGroup);
flower_destruct(flower, 0);
return 1;
}
return 0;
}
void topDown(Flower *flower, Name referenceEventName) {
/*
* Run on each flower, top down. Sets the coordinates of each reference cap to the correct
* sequence, and sets the bases of the reference sequence to be consensus bases.
*/
Flower_EndIterator *endIt = flower_getEndIterator(flower);
End *end;
while ((end = flower_getNextEnd(endIt)) != NULL) {
Cap *cap = getCapForReferenceEvent(end, referenceEventName); //The cap in the reference
if (cap != NULL) {
cap = cap_getStrand(cap) ? cap : cap_getReverse(cap);
if (!cap_getSide(cap)) {
assert(cap_getCoordinate(cap) != INT64_MAX);
Sequence *sequence = cap_getSequence(cap);
assert(sequence != NULL);
Group *group = end_getGroup(end);
if (!group_isLeaf(group)) {
Flower *nestedFlower = group_getNestedFlower(group);
Cap *nestedCap = flower_getCap(nestedFlower, cap_getName(cap));
assert(nestedCap != NULL);
nestedCap = cap_getStrand(nestedCap) ? nestedCap : cap_getReverse(nestedCap);
assert(cap_getStrand(nestedCap));
assert(!cap_getSide(nestedCap));
int64_t endCoordinate = setCoordinates(nestedFlower, sequence_getMetaSequence(sequence),
nestedCap, cap_getCoordinate(cap));
(void) endCoordinate;
assert(endCoordinate == cap_getCoordinate(cap_getAdjacency(cap)));
assert(endCoordinate
== cap_getCoordinate(
flower_getCap(nestedFlower, cap_getName(cap_getAdjacency(cap)))));
}
}
}
}
flower_destructEndIterator(endIt);
}
int main(int argc, char *argv[]) {
/*
* Script for adding a reference genome to a flower.
*/
/*
* Arguments/options
*/
char * logLevelString = NULL;
char * cactusDiskDatabaseString = NULL;
char * secondaryDatabaseString = NULL;
char *referenceEventString = (char *) cactusMisc_getDefaultReferenceEventHeader();
bool bottomUpPhase = 0;
///////////////////////////////////////////////////////////////////////////
// (0) Parse the inputs handed by genomeCactus.py / setup stuff.
///////////////////////////////////////////////////////////////////////////
while (1) {
static struct option long_options[] = { { "logLevel", required_argument, 0, 'a' }, { "cactusDisk", required_argument, 0, 'b' }, { "secondaryDisk", required_argument, 0, 'd' }, { "referenceEventString", required_argument, 0, 'g' }, { "help", no_argument,
0, 'h' }, { "bottomUpPhase", no_argument, 0, 'j' }, { 0, 0, 0, 0 } };
int option_index = 0;
int key = getopt_long(argc, argv, "a:b:c:d:e:g:hi:j", long_options, &option_index);
if (key == -1) {
break;
}
switch (key) {
case 'a':
logLevelString = stString_copy(optarg);
break;
case 'b':
cactusDiskDatabaseString = stString_copy(optarg);
break;
case 'd':
secondaryDatabaseString = stString_copy(optarg);
break;
case 'g':
referenceEventString = stString_copy(optarg);
break;
case 'h':
usage();
return 0;
case 'j':
bottomUpPhase = 1;
break;
default:
usage();
return 1;
}
}
///////////////////////////////////////////////////////////////////////////
// (0) Check the inputs.
///////////////////////////////////////////////////////////////////////////
assert(cactusDiskDatabaseString != NULL);
//////////////////////////////////////////////
//Set up logging
//////////////////////////////////////////////
st_setLogLevelFromString(logLevelString);
//////////////////////////////////////////////
//Load the database
//////////////////////////////////////////////
st_logInfo("referenceEventString = %s\n", referenceEventString);
st_logInfo("bottomUpPhase = %i\n", bottomUpPhase);
stKVDatabaseConf *kvDatabaseConf = stKVDatabaseConf_constructFromString(cactusDiskDatabaseString);
CactusDisk *cactusDisk = cactusDisk_construct(kvDatabaseConf, false, true);
stKVDatabaseConf_destruct(kvDatabaseConf);
st_logInfo("Set up the flower disk\n");
stKVDatabase *sequenceDatabase = NULL;
if (secondaryDatabaseString != NULL) {
kvDatabaseConf = stKVDatabaseConf_constructFromString(secondaryDatabaseString);
sequenceDatabase = stKVDatabase_construct(kvDatabaseConf, 0);
stKVDatabaseConf_destruct(kvDatabaseConf);
}
FlowerStream *flowerStream = flowerWriter_getFlowerStream(cactusDisk, stdin);
Flower *flower;
while ((flower = flowerStream_getNext(flowerStream)) != NULL) {
st_logDebug("Processing flower %" PRIi64 "\n", flower_getName(flower));
///////////////////////////////////////////////////////////////////////////
// Get the appropriate event names
///////////////////////////////////////////////////////////////////////////
st_logInfo("%s\n", eventTree_makeNewickString(flower_getEventTree(flower)));
Event *referenceEvent = eventTree_getEventByHeader(flower_getEventTree(flower), referenceEventString);
if (referenceEvent == NULL) {
st_errAbort("Reference event %s not found in tree. Check your "
"--referenceEventString option", referenceEventString);
}
Name referenceEventName = event_getName(referenceEvent);
///////////////////////////////////////////////////////////////////////////
// Now do bottom up or top down, depending
///////////////////////////////////////////////////////////////////////////
stList *flowers = stList_construct();
stList_append(flowers, flower);
preCacheNestedFlowers(cactusDisk, flowers);
if (bottomUpPhase) {
assert(sequenceDatabase != NULL);
cactusDisk_preCacheSegmentStrings(cactusDisk, flowers);
bottomUp(flowers, sequenceDatabase, referenceEventName, !flower_hasParentGroup(flower), generateJukesCantorMatrix);
// Unload the nested flowers to save memory. They haven't
// been changed, so we don't write them to the cactus
// disk.
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (!group_isLeaf(group)) {
flower_unload(group_getNestedFlower(group));
}
}
flower_destructGroupIterator(groupIt);
assert(!flower_isParentLoaded(flower));
// Write this flower to disk.
cactusDisk_addUpdateRequest(cactusDisk, flower);
} else {
topDown(flower, referenceEventName);
// We've changed the nested flowers, but not this
// flower. We write the nested flowers to disk, then
// unload them to save memory. This flower will be
// unloaded by the flower-stream code.
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (!group_isLeaf(group)) {
cactusDisk_addUpdateRequest(cactusDisk, group_getNestedFlower(group));
flower_unload(group_getNestedFlower(group));
}
}
flower_destructGroupIterator(groupIt);
}
stList_destruct(flowers);
}
///////////////////////////////////////////////////////////////////////////
// Write the flower(s) back to disk.
///////////////////////////////////////////////////////////////////////////
cactusDisk_write(cactusDisk);
st_logInfo("Updated the flower on disk\n");
///////////////////////////////////////////////////////////////////////////
//Clean up.
///////////////////////////////////////////////////////////////////////////
if (sequenceDatabase != NULL) {
stKVDatabase_destruct(sequenceDatabase);
}
cactusDisk_destruct(cactusDisk);
return 0; //Exit without clean up is quicker, enable cleanup when doing memory leak detection.
free(cactusDiskDatabaseString);
free(referenceEventString);
free(logLevelString);
st_logInfo("Cleaned stuff up and am finished\n");
return 0;
}
void testGroup_getNestedFlower(CuTest* testCase) {
cactusGroupTestSetup();
CuAssertTrue(testCase, group_getNestedFlower(group) == nestedFlower);
cactusGroupTestTeardown();
}
int mapGene(Cap *cap, int level, int exon, struct bed *gene, FILE *fileHandle){
/*
*Following cactus adjacencies, starting from 'cap', find regions that overlap with
*exons of input gene. Report chain relations of these regions with the exons.
*cap: current cap. Level = chain level. exon = exon number. gene = bed record of gene
*/
int64_t exonStart, exonEnd;
if(isStubCap(cap)){
Group *group = end_getGroup(cap_getEnd(cap));
Flower *nestedFlower = group_getNestedFlower(group);
if(nestedFlower != NULL){//recursive call
Cap *childCap = flower_getCap(nestedFlower, cap_getName(cap));
assert(childCap != NULL);
exon = mapGene(childCap, level + 1, exon, gene, fileHandle);
exonStart = gene->chromStarts->list[exon] + gene->chromStart;
exonEnd = exonStart + gene->blockSizes->list[exon];
}
}
cap = cap_getAdjacency(cap);
Cap *nextcap;
int64_t capCoor;
exonStart = gene->chromStarts->list[exon] + gene->chromStart;
exonEnd = exonStart + gene->blockSizes->list[exon];
Block *block = end_getBlock(cap_getEnd(cap));
if(block == NULL){
moveCapToNextBlock(&cap);
}
while(!isStubCap(cap) && exon < gene->blockCount){
End *cend = cap_getEnd(cap);
capCoor = cap_getCoordinate(cap);//Cap coordinate is always the coordinate on + strand
nextcap = cap_getAdjacency(cap_getOtherSegmentCap(cap));
st_logInfo("capCoor: %d, nextCap: %d, eStart: %d, eEnd: %d. Exon: %d\n",
capCoor, cap_getCoordinate(nextcap), exonStart, exonEnd, exon);
//keep moving if nextBlock Start is still upstream of current exon
if(cap_getCoordinate(nextcap) <= exonStart){
moveCapToNextBlock(&cap);
st_logInfo("Still upstream, nextcap <= exonStart. Move to next chainBlock\n");
}else if(capCoor >= exonEnd){//Done with current exon, move to next
st_logInfo("Done with current exon, move to next one\n\n");
fprintf(fileHandle, "\t\t</exon>\n");//end previous exon
exon++;
if(exon < gene->blockCount){
exonStart = gene->chromStarts->list[exon] + gene->chromStart;
exonEnd = exonStart + gene->blockSizes->list[exon];
fprintf(fileHandle, "\t\t<exon id=\"%d\" start=\"%" PRIi64 "\" end=\"%" PRIi64 "\">\n", exon, exonStart, exonEnd);
}
}else{//current exon overlaps with current block Or with lower level flower
Cap *oppcap = cap_getOtherSegmentCap(cap);
st_logInfo("Current exon overlaps with current block or with lower flower\n");
if(cap_getCoordinate(oppcap) >= exonStart && exonEnd > capCoor){
mapBlockToExon(cap, level, fileHandle);
if(exonEnd <= cap_getCoordinate(oppcap) + 1){
st_logInfo("Done with current exon, move to next one\n\n");
fprintf(fileHandle, "\t\t</exon>\n");//end previous exon
exon++;
if(exon < gene->blockCount){
exonStart = gene->chromStarts->list[exon] + gene->chromStart;
exonEnd = exonStart + gene->blockSizes->list[exon];
fprintf(fileHandle, "\t\t<exon id=\"%d\" start=\"%" PRIi64 "\" end=\"%" PRIi64 "\">\n", exon, exonStart, exonEnd);
}
continue;
}
}
//Traverse lower level flowers if exists
Group *group = end_getGroup(end_getOtherBlockEnd(cend));
Flower *nestedFlower = group_getNestedFlower(group);
if(nestedFlower != NULL){//recursive call
Cap *childCap = flower_getCap(nestedFlower, cap_getName(cap_getOtherSegmentCap(cap)));
assert(childCap != NULL);
exon = mapGene(childCap, level + 1, exon, gene, fileHandle);
exonStart = gene->chromStarts->list[exon] + gene->chromStart;
exonEnd = exonStart + gene->blockSizes->list[exon];
}
moveCapToNextBlock(&cap);
}
}
return exon;
}
void makeCactusTree_flower(Flower *flower, FILE *fileHandle, const char *parentNodeName,
const char *parentEdgeColour) {
if(flower_isTerminal(flower)) {
makeCactusTree_terminalNode(flower, fileHandle, parentNodeName, parentEdgeColour);
}
else {
//Write the flower nodes.
char *flowerNameString = cactusMisc_nameToString(flower_getName(flower));
const char *edgeColour = graphViz_getColour();
addNodeToGraph(flowerNameString, fileHandle, flower_getTotalBaseLength(flower)
/ totalProblemSize, "ellipse", flowerNameString);
//Write in the parent edge.
if (parentNodeName != NULL) {
graphViz_addEdgeToGraph(parentNodeName, flowerNameString, fileHandle, "",
parentEdgeColour, 10, 1, "forward");
}
//Create the chains.
Flower_ChainIterator *chainIterator = flower_getChainIterator(flower);
Chain *chain;
while ((chain = flower_getNextChain(chainIterator)) != NULL) {
makeCactusTree_chain(chain, fileHandle, flowerNameString, edgeColour);
}
flower_destructChainIterator(chainIterator);
//Create the diamond node
char *diamondNodeNameString = st_malloc(sizeof(char) * (strlen(
flowerNameString) + 2));
sprintf(diamondNodeNameString, "z%s", flowerNameString);
const char *diamondEdgeColour = graphViz_getColour();
//Create all the groups linked to the diamond.
Flower_GroupIterator *groupIterator = flower_getGroupIterator(flower);
Group *group;
double size = 0.0; //get the size of the group organising node..
int64_t nonTrivialGroupCount = 0;
while ((group = flower_getNextGroup(groupIterator)) != NULL) {
assert(!group_isLeaf(group));
if (group_isTangle(group)) {
size += group_getTotalBaseLength(group);
nonTrivialGroupCount++;
}
}
flower_destructGroupIterator(groupIterator);
if(nonTrivialGroupCount == 0) {
assert(flower_getParentGroup(flower) == 0);
}
else {
//assert(nonTrivialGroupCount > 0);
addNodeToGraph(diamondNodeNameString, fileHandle, size
/ totalProblemSize, "diamond", "");
graphViz_addEdgeToGraph(flowerNameString, diamondNodeNameString,
fileHandle, "", edgeColour, 10, 1, "forward");
groupIterator = flower_getGroupIterator(flower);
while ((group = flower_getNextGroup(groupIterator)) != NULL) {
if (group_isTangle(group)) {
assert(!group_isLeaf(group));
makeCactusTree_flower(group_getNestedFlower(group), fileHandle,
diamondNodeNameString, diamondEdgeColour);
}
}
flower_destructGroupIterator(groupIterator);
}
free(flowerNameString);
free(diamondNodeNameString);
}
}
