void testEnd_getName(CuTest* testCase) {
cactusEndTestSetup();
CuAssertTrue(testCase, end_getName(end) != NULL_NAME);
CuAssertTrue(testCase, flower_getEnd(flower, end_getName(end)) == end);
CuAssertTrue(testCase, flower_getEnd(flower, end_getName(end_getReverse(end))) == end);
cactusEndTestTeardown();
}
void testEnd_setGroup(CuTest* testCase) {
cactusEndTestSetup();
Flower *flower2 = flower_construct(cactusDisk);
Group *group2 = group_construct2(flower2);
End *end2 = end_construct(1, flower2);
End *end3 = end_construct(1, flower2);
CuAssertTrue(testCase, group_getEndNumber(group2) == 0);
CuAssertTrue(testCase, end_getGroup(end2) == NULL);
CuAssertTrue(testCase, end_getGroup(end3) == NULL);
end_setGroup(end2, group2);
CuAssertTrue(testCase, group_getEndNumber(group2) == 1);
CuAssertTrue(testCase, end_getGroup(end2) == group2);
CuAssertTrue(testCase, group_getEnd(group2, end_getName(end2)) == end2);
CuAssertTrue(testCase, end_getGroup(end3) == NULL);
end_setGroup(end3, group2);
CuAssertTrue(testCase, group_getEndNumber(group2) == 2);
CuAssertTrue(testCase, end_getGroup(end2) == group2);
CuAssertTrue(testCase, group_getEnd(group2, end_getName(end2)) == end2);
CuAssertTrue(testCase, end_getGroup(end3) == group2);
CuAssertTrue(testCase, group_getEnd(group2, end_getName(end3)) == end3);
end_setGroup(end3, NULL);
end_setGroup(end2, group2);
CuAssertTrue(testCase, group_getEndNumber(group2) == 1);
CuAssertTrue(testCase, end_getGroup(end2) == group2);
CuAssertTrue(testCase, group_getEnd(group2, end_getName(end2)) == end2);
CuAssertTrue(testCase, end_getGroup(end3) == NULL);
cactusEndTestTeardown();
}
void testGroup_updateContainedEnds(CuTest* testCase) {
cactusGroupTestSetup();
end_copyConstruct(end3, nestedFlower);
CuAssertTrue(testCase, group_getEndNumber(group) == 2);
group_updateContainedEnds(group);
CuAssertTrue(testCase, group_getEndNumber(group) == 3);
CuAssertTrue(testCase, group_getEnd(group, end_getName(end1)) == end1);
CuAssertTrue(testCase, group_getEnd(group, end_getName(end2)) == end2);
CuAssertTrue(testCase, group_getEnd(group, end_getName(end3)) == end3);
cactusGroupTestTeardown();
}
void testEnd_copyConstruct(CuTest* testCase) {
cactusEndTestSetup();
Flower *flower2 = flower_construct(cactusDisk);
eventTree_copyConstruct(eventTree, flower2, testEnd_copyConstructP);
sequence_construct(metaSequence, flower2);
End *end2 = end_copyConstruct(end, flower2);
CuAssertTrue(testCase, end_getName(end2) != NULL_NAME);
CuAssertTrue(testCase, end_getName(end2) == end_getName(end));
CuAssertTrue(testCase, flower_getEnd(flower2, end_getName(end2)) == end2);
CuAssertTrue(testCase, cap_getName(end_getInstance(end2, cap_getName(rootCap))) == cap_getName(rootCap));
CuAssertTrue(testCase, cap_getName(end_getInstance(end2, cap_getName(leaf1Cap))) == cap_getName(leaf1Cap));
CuAssertTrue(testCase, cap_getName(end_getInstance(end2, cap_getName(leaf2Cap))) == cap_getName(leaf2Cap));
cactusEndTestTeardown();
}
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_addEnd(CuTest *testCase) {
cactusGroupTestSetup();
CuAssertTrue(testCase, group_getEndNumber(group2) == 0);
end_setGroup(end4, group2);
CuAssertTrue(testCase, group_getEndNumber(group2) == 1);
CuAssertTrue(testCase, end_getGroup(end4) == group2);
CuAssertTrue(testCase, group_getEnd(group2, end_getName(end4)) == end4);
cactusGroupTestTeardown();
}
void writeEndAlignmentToDisk(End *end, stSortedSet *endAlignment, FILE *fileHandle) {
fprintf(fileHandle, "%s %" PRIi64 "\n", cactusMisc_nameToStringStatic(end_getName(end)), stSortedSet_size(endAlignment));
stSortedSetIterator *it = stSortedSet_getIterator(endAlignment);
AlignedPair *aP;
while((aP = stSortedSet_getNext(it)) != NULL) {
fprintf(fileHandle, "%" PRIi64 " %" PRIi64 " %i %" PRIi64 " ", aP->subsequenceIdentifier, aP->position, aP->strand, aP->score);
aP = aP->reverse;
fprintf(fileHandle, "%" PRIi64 " %" PRIi64 " %i %" PRIi64 "\n", aP->subsequenceIdentifier, aP->position, aP->strand, aP->score);
}
stSortedSet_destructIterator(it);
}
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 endsAreConnected(End *end1, End *end2, stList *eventStrings) {
if (end_getName(end1) == end_getName(end2)) { //Then the ends are the same and are part of the same chromosome by definition.
End_InstanceIterator *instanceIterator = end_getInstanceIterator(end1);
Cap *cap1;
while ((cap1 = end_getNext(instanceIterator)) != NULL) {
if (capHasGivenEvents(cap1, eventStrings)) {
end_destructInstanceIterator(instanceIterator);
return 1;
}
}
return 0;
}
End_InstanceIterator *instanceIterator = end_getInstanceIterator(end1);
Cap *cap1;
while ((cap1 = end_getNext(instanceIterator)) != NULL) {
if (capHasGivenEvents(cap1, eventStrings)) {
End_InstanceIterator *instanceIterator2 = end_getInstanceIterator(end2);
Cap *cap2;
while ((cap2 = end_getNext(instanceIterator2)) != NULL) {
assert(cap_getName(cap2) != cap_getName(cap1)); //This could only happen if end1 == end2
if (sequence_getMetaSequence(cap_getSequence(cap1)) == sequence_getMetaSequence(cap_getSequence(cap2))) {
assert(strcmp(event_getHeader(cap_getEvent(cap1)),
event_getHeader(cap_getEvent(cap2))) == 0);
assert(cap_getPositiveOrientation(cap1)
!= cap_getPositiveOrientation(cap2));
assert(cap_getName(cap1) != cap_getName(cap2));
//they could have the same coordinate if they represent two ends of a block of length 1.
end_destructInstanceIterator(instanceIterator);
end_destructInstanceIterator(instanceIterator2);
return 1;
}
}
end_destructInstanceIterator(instanceIterator2);
}
}
end_destructInstanceIterator(instanceIterator);
return 0;
}
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;
}
stSortedSet *makeEndAlignment(StateMachine *sM, End *end, int64_t spanningTrees, int64_t maxSequenceLength,
bool useProgressiveMerging, float gapGamma,
PairwiseAlignmentParameters *pairwiseAlignmentBandingParameters) {
//Make an alignment of the sequences in the ends
//Get the adjacency sequences to be aligned.
Cap *cap;
End_InstanceIterator *it = end_getInstanceIterator(end);
stList *sequences = stList_construct3(0, (void (*)(void *))adjacencySequence_destruct);
stList *seqFrags = stList_construct3(0, (void (*)(void *))seqFrag_destruct);
stHash *endInstanceNumbers = stHash_construct2(NULL, free);
while((cap = end_getNext(it)) != NULL) {
if(cap_getSide(cap)) {
cap = cap_getReverse(cap);
}
AdjacencySequence *adjacencySequence = adjacencySequence_construct(cap, maxSequenceLength);
stList_append(sequences, adjacencySequence);
assert(cap_getAdjacency(cap) != NULL);
End *otherEnd = end_getPositiveOrientation(cap_getEnd(cap_getAdjacency(cap)));
stList_append(seqFrags, seqFrag_construct(adjacencySequence->string, 0, end_getName(otherEnd)));
//Increase count of seqfrags with a given end.
int64_t *c = stHash_search(endInstanceNumbers, otherEnd);
if(c == NULL) {
c = st_calloc(1, sizeof(int64_t));
assert(*c == 0);
stHash_insert(endInstanceNumbers, otherEnd, c);
}
(*c)++;
}
end_destructInstanceIterator(it);
//Get the alignment.
MultipleAlignment *mA = makeAlignment(sM, seqFrags, spanningTrees, 100000000, useProgressiveMerging, gapGamma, pairwiseAlignmentBandingParameters);
//Build an array of weights to reweight pairs in the alignment.
int64_t *pairwiseAlignmentsPerSequenceNonCommonEnds = st_calloc(stList_length(seqFrags), sizeof(int64_t));
int64_t *pairwiseAlignmentsPerSequenceCommonEnds = st_calloc(stList_length(seqFrags), sizeof(int64_t));
//First build array on number of pairwise alignments to each sequence, distinguishing alignments between sequences sharing
//common ends.
for(int64_t i=0; i<stList_length(mA->chosenPairwiseAlignments); i++) {
stIntTuple *pairwiseAlignment = stList_get(mA->chosenPairwiseAlignments, i);
int64_t seq1 = stIntTuple_get(pairwiseAlignment, 1);
int64_t seq2 = stIntTuple_get(pairwiseAlignment, 2);
assert(seq1 != seq2);
SeqFrag *seqFrag1 = stList_get(seqFrags, seq1);
SeqFrag *seqFrag2 = stList_get(seqFrags, seq2);
int64_t *pairwiseAlignmentsPerSequence = seqFrag1->rightEndId == seqFrag2->rightEndId
? pairwiseAlignmentsPerSequenceCommonEnds : pairwiseAlignmentsPerSequenceNonCommonEnds;
pairwiseAlignmentsPerSequence[seq1]++;
pairwiseAlignmentsPerSequence[seq2]++;
}
//Now calculate score adjustments.
double *scoreAdjustmentsNonCommonEnds = st_malloc(stList_length(seqFrags) * sizeof(double));
double *scoreAdjustmentsCommonEnds = st_malloc(stList_length(seqFrags) * sizeof(double));
for(int64_t i=0; i<stList_length(seqFrags); i++) {
SeqFrag *seqFrag = stList_get(seqFrags, i);
End *otherEnd = flower_getEnd(end_getFlower(end), seqFrag->rightEndId);
assert(otherEnd != NULL);
assert(stHash_search(endInstanceNumbers, otherEnd) != NULL);
int64_t commonInstanceNumber = *(int64_t *)stHash_search(endInstanceNumbers, otherEnd);
int64_t nonCommonInstanceNumber = stList_length(seqFrags) - commonInstanceNumber;
assert(commonInstanceNumber > 0 && nonCommonInstanceNumber >= 0);
assert(pairwiseAlignmentsPerSequenceNonCommonEnds[i] <= nonCommonInstanceNumber);
assert(pairwiseAlignmentsPerSequenceNonCommonEnds[i] >= 0);
assert(pairwiseAlignmentsPerSequenceCommonEnds[i] < commonInstanceNumber);
assert(pairwiseAlignmentsPerSequenceCommonEnds[i] >= 0);
//scoreAdjustmentsNonCommonEnds[i] = ((double)nonCommonInstanceNumber + commonInstanceNumber - 1)/(pairwiseAlignmentsPerSequenceNonCommonEnds[i] + pairwiseAlignmentsPerSequenceCommonEnds[i]);
//scoreAdjustmentsCommonEnds[i] = scoreAdjustmentsNonCommonEnds[i];
if(pairwiseAlignmentsPerSequenceNonCommonEnds[i] > 0) {
scoreAdjustmentsNonCommonEnds[i] = ((double)nonCommonInstanceNumber)/pairwiseAlignmentsPerSequenceNonCommonEnds[i];
assert(scoreAdjustmentsNonCommonEnds[i] >= 1.0);
assert(scoreAdjustmentsNonCommonEnds[i] <= nonCommonInstanceNumber);
}
else {
scoreAdjustmentsNonCommonEnds[i] = INT64_MIN;
}
if(pairwiseAlignmentsPerSequenceCommonEnds[i] > 0) {
scoreAdjustmentsCommonEnds[i] = ((double)commonInstanceNumber-1)/pairwiseAlignmentsPerSequenceCommonEnds[i];
assert(scoreAdjustmentsCommonEnds[i] >= 1.0);
assert(scoreAdjustmentsCommonEnds[i] <= commonInstanceNumber-1);
}
else {
scoreAdjustmentsCommonEnds[i] = INT64_MIN;
}
}
//Convert the alignment pairs to an alignment of the caps..
stSortedSet *sortedAlignment =
stSortedSet_construct3((int (*)(const void *, const void *))alignedPair_cmpFn,
(void (*)(void *))alignedPair_destruct);
while(stList_length(mA->alignedPairs) > 0) {
stIntTuple *alignedPair = stList_pop(mA->alignedPairs);
assert(stIntTuple_length(alignedPair) == 5);
int64_t seqIndex1 = stIntTuple_get(alignedPair, 1);
int64_t seqIndex2 = stIntTuple_get(alignedPair, 3);
AdjacencySequence *i = stList_get(sequences, seqIndex1);
AdjacencySequence *j = stList_get(sequences, seqIndex2);
assert(i != j);
int64_t offset1 = stIntTuple_get(alignedPair, 2);
int64_t offset2 = stIntTuple_get(alignedPair, 4);
int64_t score = stIntTuple_get(alignedPair, 0);
if(score <= 0) { //Happens when indel probs are included
score = 1; //This is the minimum
}
assert(score > 0 && score <= PAIR_ALIGNMENT_PROB_1);
SeqFrag *seqFrag1 = stList_get(seqFrags, seqIndex1);
SeqFrag *seqFrag2 = stList_get(seqFrags, seqIndex2);
assert(seqFrag1 != seqFrag2);
double *scoreAdjustments = seqFrag1->rightEndId == seqFrag2->rightEndId ? scoreAdjustmentsCommonEnds : scoreAdjustmentsNonCommonEnds;
assert(scoreAdjustments[seqIndex1] != INT64_MIN);
assert(scoreAdjustments[seqIndex2] != INT64_MIN);
AlignedPair *alignedPair2 = alignedPair_construct(
i->subsequenceIdentifier, i->start + (i->strand ? offset1 : -offset1), i->strand,
j->subsequenceIdentifier, j->start + (j->strand ? offset2 : -offset2), j->strand,
score*scoreAdjustments[seqIndex1], score*scoreAdjustments[seqIndex2]); //Do the reweighting here.
assert(stSortedSet_search(sortedAlignment, alignedPair2) == NULL);
assert(stSortedSet_search(sortedAlignment, alignedPair2->reverse) == NULL);
stSortedSet_insert(sortedAlignment, alignedPair2);
stSortedSet_insert(sortedAlignment, alignedPair2->reverse);
stIntTuple_destruct(alignedPair);
}
//Cleanup
stList_destruct(seqFrags);
stList_destruct(sequences);
free(pairwiseAlignmentsPerSequenceNonCommonEnds);
free(pairwiseAlignmentsPerSequenceCommonEnds);
free(scoreAdjustmentsNonCommonEnds);
free(scoreAdjustmentsCommonEnds);
multipleAlignment_destruct(mA);
stHash_destruct(endInstanceNumbers);
return sortedAlignment;
}
void testGroup_getEnd(CuTest* testCase) {
cactusGroupTestSetup();
CuAssertTrue(testCase, group_getEnd(group, end_getName(end1)) == end1);
CuAssertTrue(testCase, group_getEnd(group, end_getName(end2)) == end2);
cactusGroupTestTeardown();
}
static int flower_constructEndsP(const void *o1, const void *o2) {
return cactusMisc_nameCompare(end_getName((End *) o1), end_getName((End *) o2));
}
void testFlower_removeIfRedundant(CuTest *testCase) {
/*
* Do a simple test to see if function can remove a redundant flower.
*/
cactusFlowerTestSetup();
endsSetup();
//First construct a redundant flower from the root.
Flower *flower2 = flower_construct(cactusDisk);
Group *group = group_construct(flower, flower2);
end_setGroup(end, group);
end_setGroup(end2, group);
//Now hang another couple of flowers of that.
Flower *flower3 = flower_construct(cactusDisk);
group_construct(flower2, flower3);
//Now hang another flower of that.
Group *group3b = group_construct2(flower2);
//Finally hang one more flower on the end..
Flower *flower4 = flower_construct(cactusDisk);
group_construct(flower3, flower4);
//Copy the ends into the flowers.
end_copyConstruct(end, flower2);
end_copyConstruct(end2, flower2);
end_copyConstruct(end, flower3);
end_setGroup(flower_getEnd(flower2, end_getName(end2)), group3b);
end_copyConstruct(end, flower4);
//st_uglyf("I got %" PRIi64 " %" PRIi64 " %" PRIi64 " %" PRIi64 "\n", flower_getName(flower), flower_getName(flower2), flower_getName(flower3), flower_getName(flower4));
//Write the mess to disk.
cactusDisk_write(cactusDisk);
//Now test the removal function (check we get a negative on this leaf).
CuAssertTrue(testCase, !flower_removeIfRedundant(flower4));
//Check we can't remove the root..
CuAssertTrue(testCase, !flower_removeIfRedundant(flower));
//We will remove flower2
//Before
CuAssertTrue(testCase, flower_getGroupNumber(flower) == 1);
CuAssertTrue(testCase, group_getFlower(flower_getParentGroup(flower2)) == flower);
CuAssertTrue(testCase, flower_removeIfRedundant(flower2));
//After, check the flower/group connections
CuAssertTrue(testCase, flower_getGroupNumber(flower) == 2);
CuAssertTrue(testCase, !flower_isLeaf(flower));
CuAssertTrue(testCase, group_getFlower(flower_getParentGroup(flower3)) == flower);
group3b = end_getGroup(end2);
CuAssertTrue(testCase, group_getFlower(group3b) == flower);
CuAssertTrue(testCase, group_isLeaf(group3b));
CuAssertTrue(testCase, flower_getGroup(flower, flower_getName(flower3)) == flower_getParentGroup(flower3));
//Check the ends..
CuAssertTrue(testCase, flower_getEndNumber(flower) == 2);
CuAssertTrue(testCase, flower_getEndNumber(flower3) == 1);
CuAssertTrue(testCase, group_getEndNumber(group3b) == 1);
CuAssertTrue(testCase, end_getGroup(end) == flower_getParentGroup(flower3));
CuAssertTrue(testCase, end_getGroup(end2) == group3b);
CuAssertTrue(testCase, flower_getEnd(flower3, end_getName(end)) != NULL);
//Check the child of 3 is still okay..
CuAssertTrue(testCase, group_getFlower(flower_getParentGroup(flower4)) == flower3);
//Now do removal of flower3
CuAssertTrue(testCase, !flower_removeIfRedundant(flower));
CuAssertTrue(testCase, !flower_removeIfRedundant(flower4));
CuAssertTrue(testCase, flower_removeIfRedundant(flower3));
//Check groups again
CuAssertTrue(testCase, flower_getGroupNumber(flower) == 2);
CuAssertTrue(testCase, !flower_isLeaf(flower));
CuAssertTrue(testCase, group_getFlower(flower_getParentGroup(flower4)) == flower);
CuAssertTrue(testCase, group_getFlower(group3b) == flower);
CuAssertTrue(testCase, flower_getGroup(flower, flower_getName(flower4)) == flower_getParentGroup(flower4));
//Check the ends again..
CuAssertTrue(testCase, flower_getEndNumber(flower) == 2);
CuAssertTrue(testCase, flower_getEndNumber(flower4) == 1);
CuAssertTrue(testCase, group_getEndNumber(group3b) == 1);
CuAssertTrue(testCase, end_getGroup(end) == flower_getParentGroup(flower4));
CuAssertTrue(testCase, end_getGroup(end2) == group3b);
CuAssertTrue(testCase, flower_getEnd(flower4, end_getName(end)) != NULL);
cactusFlowerTestTeardown();
}
int main(int argc, char *argv[]) {
char * logLevelString = NULL;
char * cactusDiskDatabaseString = NULL;
int64_t i, j;
int64_t spanningTrees = 10;
int64_t maximumLength = 1500;
bool useProgressiveMerging = 0;
float matchGamma = 0.5;
bool useBanding = 0;
int64_t k;
stList *listOfEndAlignmentFiles = NULL;
char *endAlignmentsToPrecomputeOutputFile = NULL;
bool calculateWhichEndsToComputeSeparately = 0;
int64_t largeEndSize = 1000000;
int64_t chainLengthForBigFlower = 1000000;
int64_t longChain = 2;
char *ingroupCoverageFilePath = NULL;
int64_t minimumSizeToRescue = 1;
double minimumCoverageToRescue = 0.0;
PairwiseAlignmentParameters *pairwiseAlignmentBandingParameters = pairwiseAlignmentBandingParameters_construct();
/*
* Setup the input parameters for cactus core.
*/
bool pruneOutStubAlignments = 0;
/*
* Parse the options.
*/
while (1) {
static struct option long_options[] = { { "logLevel", required_argument, 0, 'a' }, { "cactusDisk", required_argument, 0, 'b' }, {
"help", no_argument, 0, 'h' }, { "spanningTrees", required_argument, 0, 'i' },
{ "maximumLength", required_argument, 0, 'j' }, { "useBanding", no_argument, 0, 'k' },
{ "gapGamma", required_argument, 0, 'l' }, { "matchGamma", required_argument, 0, 'L' },
{ "splitMatrixBiggerThanThis", required_argument, 0, 'o' }, { "anchorMatrixBiggerThanThis",
required_argument, 0, 'p' }, { "repeatMaskMatrixBiggerThanThis", required_argument, 0, 'q' }, {
"diagonalExpansion", required_argument, 0, 'r' }, { "constraintDiagonalTrim", required_argument, 0, 't' }, {
"minimumDegree", required_argument, 0, 'u' }, { "alignAmbiguityCharacters", no_argument, 0, 'w' }, {
"pruneOutStubAlignments", no_argument, 0, 'y' }, {
"minimumIngroupDegree", required_argument, 0, 'A' }, { "minimumOutgroupDegree", required_argument, 0, 'B' },
{ "precomputedAlignments", required_argument, 0, 'D' }, {
"endAlignmentsToPrecomputeOutputFile", required_argument, 0, 'E' }, { "useProgressiveMerging",
no_argument, 0, 'F' }, { "calculateWhichEndsToComputeSeparately", no_argument, 0, 'G' }, { "largeEndSize",
required_argument, 0, 'I' },
{"ingroupCoverageFile", required_argument, 0, 'J'},
{"minimumSizeToRescue", required_argument, 0, 'K'},
{"minimumCoverageToRescue", required_argument, 0, 'M'},
{ "minimumNumberOfSpecies", required_argument, 0, 'N' },
{ 0, 0, 0, 0 } };
int option_index = 0;
int key = getopt_long(argc, argv, "a:b:hi:j:kl:o:p:q:r:t:u:wy:A:B:D:E:FGI:J:K:L:M:N:", long_options, &option_index);
if (key == -1) {
break;
}
switch (key) {
case 'a':
logLevelString = stString_copy(optarg);
st_setLogLevelFromString(logLevelString);
break;
case 'b':
cactusDiskDatabaseString = stString_copy(optarg);
break;
case 'h':
usage();
return 0;
case 'i':
i = sscanf(optarg, "%" PRIi64 "", &spanningTrees);
(void) i;
assert(i == 1);
assert(spanningTrees >= 0);
break;
case 'j':
i = sscanf(optarg, "%" PRIi64 "", &maximumLength);
assert(i == 1);
assert(maximumLength >= 0);
break;
case 'k':
useBanding = !useBanding;
break;
case 'l':
i = sscanf(optarg, "%f", &pairwiseAlignmentBandingParameters->gapGamma);
assert(i == 1);
assert(pairwiseAlignmentBandingParameters->gapGamma >= 0.0);
break;
case 'L':
i = sscanf(optarg, "%f", &matchGamma);
assert(i == 1);
assert(matchGamma >= 0.0);
break;
case 'o':
i = sscanf(optarg, "%" PRIi64 "", &k);
assert(i == 1);
assert(k >= 0);
pairwiseAlignmentBandingParameters->splitMatrixBiggerThanThis = (int64_t) k * k;
break;
case 'p':
i = sscanf(optarg, "%" PRIi64 "", &k);
assert(i == 1);
assert(k >= 0);
pairwiseAlignmentBandingParameters->anchorMatrixBiggerThanThis = (int64_t) k * k;
break;
case 'q':
i = sscanf(optarg, "%" PRIi64 "", &k);
assert(i == 1);
assert(k >= 0);
pairwiseAlignmentBandingParameters->repeatMaskMatrixBiggerThanThis = (int64_t) k * k;
break;
case 'r':
i = sscanf(optarg, "%" PRIi64 "", &pairwiseAlignmentBandingParameters->diagonalExpansion);
assert(i == 1);
assert(pairwiseAlignmentBandingParameters->diagonalExpansion >= 0);
assert(pairwiseAlignmentBandingParameters->diagonalExpansion % 2 == 0);
break;
case 't':
i = sscanf(optarg, "%" PRIi64 "", &pairwiseAlignmentBandingParameters->constraintDiagonalTrim);
assert(i == 1);
assert(pairwiseAlignmentBandingParameters->constraintDiagonalTrim >= 0);
break;
case 'u':
i = sscanf(optarg, "%" PRIi64 "", &minimumDegree);
assert(i == 1);
break;
case 'w':
pairwiseAlignmentBandingParameters->alignAmbiguityCharacters = 1;
break;
case 'y':
pruneOutStubAlignments = 1;
break;
case 'A':
i = sscanf(optarg, "%" PRIi64 "", &minimumIngroupDegree);
assert(i == 1);
break;
case 'B':
i = sscanf(optarg, "%" PRIi64 "", &minimumOutgroupDegree);
assert(i == 1);
break;
case 'D':
listOfEndAlignmentFiles = stString_split(optarg);
break;
case 'E':
endAlignmentsToPrecomputeOutputFile = stString_copy(optarg);
break;
case 'F':
useProgressiveMerging = 1;
break;
case 'G':
calculateWhichEndsToComputeSeparately = 1;
break;
case 'I':
i = sscanf(optarg, "%" PRIi64 "", &largeEndSize);
assert(i == 1);
break;
case 'J':
ingroupCoverageFilePath = stString_copy(optarg);
break;
case 'K':
i = sscanf(optarg, "%" PRIi64, &minimumSizeToRescue);
assert(i == 1);
break;
case 'M':
i = sscanf(optarg, "%lf", &minimumCoverageToRescue);
assert(i == 1);
break;
case 'N':
i = sscanf(optarg, "%" PRIi64, &minimumNumberOfSpecies);
if (i != 1) {
st_errAbort("Error parsing minimumNumberOfSpecies parameter");
}
break;
default:
usage();
return 1;
}
}
st_setLogLevelFromString(logLevelString);
/*
* Load the flowerdisk
*/
stKVDatabaseConf *kvDatabaseConf = stKVDatabaseConf_constructFromString(cactusDiskDatabaseString);
CactusDisk *cactusDisk = cactusDisk_construct(kvDatabaseConf, 0); //We precache the sequences
st_logInfo("Set up the flower disk\n");
/*
* Load the hmm
*/
StateMachine *sM = stateMachine5_construct(fiveState);
/*
* For each flower.
*/
if (calculateWhichEndsToComputeSeparately) {
stList *flowers = flowerWriter_parseFlowersFromStdin(cactusDisk);
if (stList_length(flowers) != 1) {
st_errAbort("We are breaking up a flower's end alignments for precomputation but we have %" PRIi64 " flowers.\n", stList_length(flowers));
}
stSortedSet *endsToAlignSeparately = getEndsToAlignSeparately(stList_get(flowers, 0), maximumLength, largeEndSize);
assert(stSortedSet_size(endsToAlignSeparately) != 1);
stSortedSetIterator *it = stSortedSet_getIterator(endsToAlignSeparately);
End *end;
while ((end = stSortedSet_getNext(it)) != NULL) {
fprintf(stdout, "%s\t%" PRIi64 "\t%" PRIi64 "\n", cactusMisc_nameToStringStatic(end_getName(end)), end_getInstanceNumber(end), getTotalAdjacencyLength(end));
}
return 0; //avoid cleanup costs
stSortedSet_destructIterator(it);
stSortedSet_destruct(endsToAlignSeparately);
} else if (endAlignmentsToPrecomputeOutputFile != NULL) {
/*
* In this case we will align a set of end and save the alignments in a file.
*/
stList *names = flowerWriter_parseNames(stdin);
Flower *flower = cactusDisk_getFlower(cactusDisk, *((Name *)stList_get(names, 0)));
FILE *fileHandle = fopen(endAlignmentsToPrecomputeOutputFile, "w");
for(int64_t i=1; i<stList_length(names); i++) {
End *end = flower_getEnd(flower, *((Name *)stList_get(names, i)));
if (end == NULL) {
st_errAbort("The end %" PRIi64 " was not found in the flower\n", *((Name *)stList_get(names, i)));
}
stSortedSet *endAlignment = makeEndAlignment(sM, end, spanningTrees, maximumLength, useProgressiveMerging,
matchGamma, pairwiseAlignmentBandingParameters);
writeEndAlignmentToDisk(end, endAlignment, fileHandle);
stSortedSet_destruct(endAlignment);
}
fclose(fileHandle);
return 0; //avoid cleanup costs
stList_destruct(names);
st_logInfo("Finished precomputing end alignments\n");
} else {
/*
* Compute complete flower alignments, possibly loading some precomputed alignments.
*/
bedRegion *bedRegions = NULL;
size_t numBeds = 0;
if (ingroupCoverageFilePath != NULL) {
// Pre-load the mmap for the coverage file.
FILE *coverageFile = fopen(ingroupCoverageFilePath, "rb");
if (coverageFile == NULL) {
st_errnoAbort("Opening coverage file %s failed",
ingroupCoverageFilePath);
}
fseek(coverageFile, 0, SEEK_END);
int64_t coverageFileLen = ftell(coverageFile);
assert(coverageFileLen >= 0);
assert(coverageFileLen % sizeof(bedRegion) == 0);
if (coverageFileLen == 0) {
// mmap doesn't like length-0 mappings, for obvious
// reasons. Pretend that the coverage file doesn't
// exist in this case, since it contains no data.
ingroupCoverageFilePath = NULL;
} else {
// Establish a memory mapping for the file.
bedRegions = mmap(NULL, coverageFileLen, PROT_READ, MAP_SHARED,
fileno(coverageFile), 0);
if (bedRegions == MAP_FAILED) {
st_errnoAbort("Failure mapping coverage file");
}
numBeds = coverageFileLen / sizeof(bedRegion);
}
fclose(coverageFile);
}
stList *flowers = flowerWriter_parseFlowersFromStdin(cactusDisk);
if (listOfEndAlignmentFiles != NULL && stList_length(flowers) != 1) {
st_errAbort("We have precomputed alignments but %" PRIi64 " flowers to align.\n", stList_length(flowers));
}
cactusDisk_preCacheStrings(cactusDisk, flowers);
for (j = 0; j < stList_length(flowers); j++) {
flower = stList_get(flowers, j);
st_logInfo("Processing a flower\n");
stSortedSet *alignedPairs = makeFlowerAlignment3(sM, flower, listOfEndAlignmentFiles, spanningTrees, maximumLength,
useProgressiveMerging, matchGamma, pairwiseAlignmentBandingParameters, pruneOutStubAlignments);
st_logInfo("Created the alignment: %" PRIi64 " pairs\n", stSortedSet_size(alignedPairs));
stPinchIterator *pinchIterator = stPinchIterator_constructFromAlignedPairs(alignedPairs, getNextAlignedPairAlignment);
/*
* Run the cactus caf functions to build cactus.
*/
stPinchThreadSet *threadSet = stCaf_setup(flower);
stCaf_anneal(threadSet, pinchIterator, NULL);
if (minimumDegree < 2) {
stCaf_makeDegreeOneBlocks(threadSet);
}
if (minimumIngroupDegree > 0 || minimumOutgroupDegree > 0 || minimumDegree > 1) {
stCaf_melt(flower, threadSet, blockFilterFn, 0, 0, 0, INT64_MAX);
}
if (ingroupCoverageFilePath != NULL) {
// Rescue any sequence that is covered by outgroups
// but currently unaligned into single-degree blocks.
stPinchThreadSetIt pinchIt = stPinchThreadSet_getIt(threadSet);
stPinchThread *thread;
while ((thread = stPinchThreadSetIt_getNext(&pinchIt)) != NULL) {
Cap *cap = flower_getCap(flower,
stPinchThread_getName(thread));
assert(cap != NULL);
Sequence *sequence = cap_getSequence(cap);
assert(sequence != NULL);
rescueCoveredRegions(thread, bedRegions, numBeds,
sequence_getName(sequence),
minimumSizeToRescue,
minimumCoverageToRescue);
}
stCaf_joinTrivialBoundaries(threadSet);
}
stCaf_finish(flower, threadSet, chainLengthForBigFlower, longChain, INT64_MAX, INT64_MAX); //Flower now destroyed.
stPinchThreadSet_destruct(threadSet);
st_logInfo("Ran the cactus core script.\n");
/*
* Cleanup
*/
//Clean up the sorted set after cleaning up the iterator
stPinchIterator_destruct(pinchIterator);
stSortedSet_destruct(alignedPairs);
st_logInfo("Finished filling in the alignments for the flower\n");
}
stList_destruct(flowers);
//st_errAbort("Done\n");
/*
* Write and close the cactusdisk.
*/
cactusDisk_write(cactusDisk);
return 0; //Exit without clean up is quicker, enable cleanup when doing memory leak detection.
if (bedRegions != NULL) {
// Clean up our mapping.
munmap(bedRegions, numBeds * sizeof(bedRegion));
}
}
///////////////////////////////////////////////////////////////////////////
// Cleanup
///////////////////////////////////////////////////////////////////////////
stateMachine_destruct(sM);
cactusDisk_destruct(cactusDisk);
stKVDatabaseConf_destruct(kvDatabaseConf);
//destructCactusCoreInputParameters(cCIP);
free(cactusDiskDatabaseString);
if (listOfEndAlignmentFiles != NULL) {
stList_destruct(listOfEndAlignmentFiles);
}
if (logLevelString != NULL) {
free(logLevelString);
}
st_logInfo("Finished with the flower disk for this flower.\n");
//while(1);
return 0;
}