void DoMuscle()
{
SetOutputFileName(g_pstrOutFileName.get());
SetInputFileName(g_pstrInFileName.get());
SetMaxIters(g_uMaxIters.get());
SetSeqWeightMethod(g_SeqWeight1.get());
TextFile fileIn(g_pstrInFileName.get());
SeqVect v;
v.FromFASTAFile(fileIn);
const unsigned uSeqCount = v.Length();
if (0 == uSeqCount)
Quit("No sequences in input file");
ALPHA Alpha = ALPHA_Undefined;
switch (g_SeqType.get())
{
case SEQTYPE_Auto:
Alpha = v.GuessAlpha();
break;
case SEQTYPE_Protein:
Alpha = ALPHA_Amino;
break;
case SEQTYPE_DNA:
Alpha = ALPHA_DNA;
break;
case SEQTYPE_RNA:
Alpha = ALPHA_RNA;
break;
default:
Quit("Invalid seq type");
}
SetAlpha(Alpha);
v.FixAlpha();
//
// AED 21/12/06: Moved matrix loading code inside the PP param function so it gets called for all alignment types
//
SetPPScore();
unsigned uMaxL = 0;
unsigned uTotL = 0;
for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex)
{
unsigned L = v.GetSeq(uSeqIndex).Length();
uTotL += L;
if (L > uMaxL)
uMaxL = L;
}
SetIter(1);
g_bDiags.get() = g_bDiags1.get();
SetSeqStats(uSeqCount, uMaxL, uTotL/uSeqCount);
SetMuscleSeqVect(v);
MSA::SetIdCount(uSeqCount);
// Initialize sequence ids.
// From this point on, ids must somehow propogate from here.
for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex)
v.SetSeqId(uSeqIndex, uSeqIndex);
if (0 == uSeqCount)
Quit("Input file '%s' has no sequences", g_pstrInFileName.get());
if (1 == uSeqCount)
{
TextFile fileOut(g_pstrOutFileName.get(), true);
v.ToFile(fileOut);
return;
}
if (uSeqCount > 1)
MHackStart(v);
// First iteration
Tree GuideTree;
if (0 != g_pstrUseTreeFileName.get())
{
// Discourage users...
if (!g_bUseTreeNoWarn.get())
fprintf(stderr, g_strUseTreeWarning);
// Read tree from file
TextFile TreeFile(g_pstrUseTreeFileName.get());
GuideTree.FromFile(TreeFile);
// Make sure tree is rooted
if (!GuideTree.IsRooted())
Quit("User tree must be rooted");
if (GuideTree.GetLeafCount() != uSeqCount)
Quit("User tree does not match input sequences");
const unsigned uNodeCount = GuideTree.GetNodeCount();
for (unsigned uNodeIndex = 0; uNodeIndex < uNodeCount; ++uNodeIndex)
{
if (!GuideTree.IsLeaf(uNodeIndex))
continue;
const char *LeafName = GuideTree.GetLeafName(uNodeIndex);
unsigned uSeqIndex;
bool SeqFound = v.FindName(LeafName, &uSeqIndex);
if (!SeqFound)
Quit("Label %s in tree does not match sequences", LeafName);
unsigned uId = v.GetSeqIdFromName(LeafName);
GuideTree.SetLeafId(uNodeIndex, uId);
}
}
else
TreeFromSeqVect(v, GuideTree, g_Cluster1.get(), g_Distance1.get(), g_Root1.get(),
g_pstrDistMxFileName1.get());
const char *Tree1 = ValueOpt("Tree1");
if (0 != Tree1)
{
TextFile f(Tree1, true);
GuideTree.ToFile(f);
if (g_bClusterOnly.get())
return;
}
SetMuscleTree(GuideTree);
ValidateMuscleIds(GuideTree);
MSA msa;
ProgNode *ProgNodes = 0;
if (g_bLow.get())
ProgNodes = ProgressiveAlignE(v, GuideTree, msa);
else
ProgressiveAlign(v, GuideTree, msa);
SetCurrentAlignment(msa);
if (0 != g_pstrComputeWeightsFileName.get())
{
extern void OutWeights(const char *FileName, const MSA &msa);
SetMSAWeightsMuscle(msa);
OutWeights(g_pstrComputeWeightsFileName.get(), msa);
return;
}
ValidateMuscleIds(msa);
if (1 == g_uMaxIters.get() || 2 == uSeqCount)
{
//TextFile fileOut(g_pstrOutFileName.get(), true);
//MHackEnd(msa);
//msa.ToFile(fileOut);
MuscleOutput(msa);
return;
}
if (0 == g_pstrUseTreeFileName.get())
{
g_bDiags.get() = g_bDiags2.get();
SetIter(2);
if (g_bLow.get())
{
if (0 != g_uMaxTreeRefineIters.get())
RefineTreeE(msa, v, GuideTree, ProgNodes);
}
else
RefineTree(msa, GuideTree);
const char *Tree2 = ValueOpt("Tree2");
if (0 != Tree2)
{
TextFile f(Tree2, true);
GuideTree.ToFile(f);
}
}
SetSeqWeightMethod(g_SeqWeight2.get());
SetMuscleTree(GuideTree);
if (g_bAnchors.get())
RefineVert(msa, GuideTree, g_uMaxIters.get() - 2);
else
RefineHoriz(msa, GuideTree, g_uMaxIters.get() - 2, false, false);
#if 0
// Refining by subfamilies is disabled as it didn't give better
// results. I tried doing this before and after RefineHoriz.
// Should get back to this as it seems like this should work.
RefineSubfams(msa, GuideTree, g_uMaxIters.get() - 2);
#endif
ValidateMuscleIds(msa);
ValidateMuscleIds(GuideTree);
//TextFile fileOut(g_pstrOutFileName.get(), true);
//MHackEnd(msa);
//msa.ToFile(fileOut);
MuscleOutput(msa);
}
void RefineW(const MSA &msaIn, MSA &msaOut)
{
const unsigned uSeqCount = msaIn.GetSeqCount();
const unsigned uColCount = msaIn.GetColCount();
// Reserve same nr seqs, 20% more cols
const unsigned uReserveColCount = (uColCount*120)/100;
msaOut.SetSize(uSeqCount, uReserveColCount);
for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex)
{
msaOut.SetSeqName(uSeqIndex, msaIn.GetSeqName(uSeqIndex));
msaOut.SetSeqId(uSeqIndex, msaIn.GetSeqId(uSeqIndex));
}
const unsigned uWindowCount = (uColCount + g_uRefineWindow.get() - 1)/g_uRefineWindow.get();
if (0 == g_uWindowTo.get())
g_uWindowTo.get() = uWindowCount - 1;
#if MEMDEBUG
_CrtSetBreakAlloc(1560);
#endif
if (g_uWindowOffset.get() > 0)
{
MSA msaTmp;
MSAFromColRange(msaIn, 0, g_uWindowOffset.get(), msaOut);
}
if (!g_bQuiet.get())
fprintf(stderr, "\n");
for (unsigned uWindowIndex = g_uWindowFrom.get(); uWindowIndex <= g_uWindowTo.get(); ++uWindowIndex)
{
if (!g_bQuiet.get())
fprintf(stderr, "Window %d of %d \r", uWindowIndex, uWindowCount);
const unsigned uColFrom = g_uWindowOffset.get() + uWindowIndex*g_uRefineWindow.get();
unsigned uColTo = uColFrom + g_uRefineWindow.get() - 1;
if (uColTo >= uColCount)
uColTo = uColCount - 1;
assert(uColTo >= uColFrom);
SeqVect v;
SeqVectFromMSACols(msaIn, uColFrom, uColTo, v);
#if MEMDEBUG
_CrtMemState s1;
_CrtMemCheckpoint(&s1);
#endif
// Begin AED 5/20/06
// remove any empty seqs in this window
std::vector< size_t > empty_seqs;
SeqVect vr;
for( size_t seqI = 0; seqI < v.size(); ++seqI )
{
if( v[seqI]->size() == 0 )
empty_seqs.push_back(seqI);
else
vr.push_back(v[seqI]);
}
std::vector< unsigned > seqid_map( vr.size() );
for( size_t seqI = 0; seqI < vr.size(); ++seqI )
{
seqid_map[seqI] = vr[seqI]->GetId();
vr[seqI]->SetId(seqI);
}
MSA msaTmp;
if( vr.size() > 1 )
MUSCLE(vr, msaTmp);
// remap the seqids to their original state
for( size_t seqI = 0; seqI < vr.size(); ++seqI )
vr[seqI]->SetId(seqid_map[seqI]);
// merge empty seqs back in
{
const unsigned uSeqCount = msaOut.GetSeqCount();
const unsigned uColCount1 = msaOut.GetColCount();
const unsigned uColCount2 = vr.size() > 1 ? msaTmp.GetColCount() : vr[0]->size();
const unsigned uColCountCat = uColCount1 + uColCount2;
for( unsigned seqI = 0; seqI < vr.size(); ++seqI )
{
unsigned uSeqIndex = msaOut.GetSeqIndex(seqid_map[seqI]);
if( vr.size() > 1 )
{
unsigned uSeqIndex2 = msaTmp.GetSeqIndex(seqI);
for (unsigned uColIndex = 0; uColIndex < uColCount2; ++uColIndex)
{
const char c = msaTmp.GetChar(uSeqIndex2, uColIndex);
msaOut.SetChar(uSeqIndex, uColCount1 + uColIndex, c);
}
}else{
for (unsigned uColIndex = 0; uColIndex < uColCount2; ++uColIndex)
{
const char c = vr[0]->GetChar(uColIndex);
msaOut.SetChar(uSeqIndex, uColCount1 + uColIndex, c);
}
}
}
for( unsigned seqI = 0; seqI < empty_seqs.size(); ++seqI )
{
unsigned uSeqId2 = v[empty_seqs[seqI]]->GetId();
unsigned uSeqIndex = msaOut.GetSeqIndex(uSeqId2);
for (unsigned uColIndex = 0; uColIndex < uColCount2; ++uColIndex)
{
msaOut.SetChar(uSeqIndex, uColCount1 + uColIndex, '-');
}
}
vr.clear();
}
// AppendMSA(msaOut, msaTmp);
// end AED 5/20/06
if (uWindowIndex == g_uSaveWindow.get())
{
MSA msaInTmp;
unsigned uOutCols = msaOut.GetColCount();
unsigned un = uColTo - uColFrom + 1;
MSAFromColRange(msaIn, uColFrom, un, msaInTmp);
char fn[256];
sprintf(fn, "win%d_inaln.tmp", uWindowIndex);
TextFile fIn(fn, true);
msaInTmp.ToFile(fIn);
sprintf(fn, "win%d_inseqs.tmp", uWindowIndex);
TextFile fv(fn, true);
v.ToFile(fv);
sprintf(fn, "win%d_outaln.tmp", uWindowIndex);
TextFile fOut(fn, true);
msaTmp.ToFile(fOut);
}
#if MEMDEBUG
void FreeDPMemSPN();
FreeDPMemSPN();
_CrtMemState s2;
_CrtMemCheckpoint(&s2);
_CrtMemState s;
_CrtMemDifference(&s, &s1, &s2);
_CrtMemDumpStatistics(&s);
_CrtMemDumpAllObjectsSince(&s1);
exit(1);
#endif
//#if DEBUG
// AssertMSAEqIgnoreCaseAndGaps(msaInTmp, msaTmp);
//#endif
}
if (!g_bQuiet.get())
fprintf(stderr, "\n");
// AssertMSAEqIgnoreCaseAndGaps(msaIn, msaOut);//@@uncomment!
}
void DoMuscle(CompositeVect*CVLocation)
{
SetOutputFileName(g_pstrOutFileName);
SetInputFileName(g_pstrInFileName);
SetMaxIters(g_uMaxIters);
SetSeqWeightMethod(g_SeqWeight1);
TextFile fileIn(g_pstrInFileName);
SeqVect v;
v.FromFASTAFile(fileIn);
const unsigned uSeqCount = v.Length();
if (0 == uSeqCount)
Quit("No sequences in input file");
ALPHA Alpha = ALPHA_Undefined;
switch (g_SeqType)
{
case SEQTYPE_Auto:
Alpha = v.GuessAlpha();
break;
case SEQTYPE_Protein:
Alpha = ALPHA_Amino;
break;
case SEQTYPE_DNA:
Alpha = ALPHA_DNA;
break;
case SEQTYPE_RNA:
Alpha = ALPHA_RNA;
break;
default:
Quit("Invalid seq type");
}
SetAlpha(Alpha);
v.FixAlpha();
PTR_SCOREMATRIX UserMatrix = 0;
if (0 != g_pstrMatrixFileName)
{
const char *FileName = g_pstrMatrixFileName;
const char *Path = getenv("MUSCLE_MXPATH");
if (Path != 0)
{
size_t n = strlen(Path) + 1 + strlen(FileName) + 1;
char *NewFileName = new char[n];
sprintf(NewFileName, "%s/%s", Path, FileName);
FileName = NewFileName;
}
TextFile File(FileName);
UserMatrix = ReadMx(File);
g_Alpha = ALPHA_Amino;
g_PPScore = PPSCORE_SP;
}
SetPPScore();
if (0 != UserMatrix)
g_ptrScoreMatrix = UserMatrix;
unsigned uMaxL = 0;
unsigned uTotL = 0;
for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex)
{
unsigned L = v.GetSeq(uSeqIndex).Length();
uTotL += L;
if (L > uMaxL)
uMaxL = L;
}
SetIter(1);
g_bDiags = g_bDiags1;
SetSeqStats(uSeqCount, uMaxL, uTotL/uSeqCount);
SetMuscleSeqVect(v);
MSA::SetIdCount(uSeqCount);
// Initialize sequence ids.
// From this point on, ids must somehow propogate from here.
for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex)
v.SetSeqId(uSeqIndex, uSeqIndex);
if (0 == uSeqCount)
Quit("Input file '%s' has no sequences", g_pstrInFileName);
if (1 == uSeqCount)
{
TextFile fileOut(g_pstrOutFileName, true);
v.ToFile(fileOut);
return;
}
if (uSeqCount > 1)
MHackStart(v);
// First iteration
Tree GuideTree;
if (0 != g_pstrUseTreeFileName)
{
// Discourage users...
if (!g_bUseTreeNoWarn)
fprintf(stderr, "%s", g_strUseTreeWarning);
// Read tree from file
TextFile TreeFile(g_pstrUseTreeFileName);
GuideTree.FromFile(TreeFile);
// Make sure tree is rooted
if (!GuideTree.IsRooted())
Quit("User tree must be rooted");
if (GuideTree.GetLeafCount() != uSeqCount)
Quit("User tree does not match input sequences");
const unsigned uNodeCount = GuideTree.GetNodeCount();
for (unsigned uNodeIndex = 0; uNodeIndex < uNodeCount; ++uNodeIndex)
{
if (!GuideTree.IsLeaf(uNodeIndex))
continue;
const char *LeafName = GuideTree.GetLeafName(uNodeIndex);
unsigned uSeqIndex;
bool SeqFound = v.FindName(LeafName, &uSeqIndex);
if (!SeqFound)
Quit("Label %s in tree does not match sequences", LeafName);
unsigned uId = v.GetSeqIdFromName(LeafName);
GuideTree.SetLeafId(uNodeIndex, uId);
}
}
else
TreeFromSeqVect(v, GuideTree, g_Cluster1, g_Distance1, g_Root1,
g_pstrDistMxFileName1);
const char *Tree1 = ValueOpt("Tree1");
if (0 != Tree1)
{
TextFile f(Tree1, true);
GuideTree.ToFile(f);
if (g_bClusterOnly)
return;
}
SetMuscleTree(GuideTree);
ValidateMuscleIds(GuideTree);
MSA msa;
msa.SetCompositeVector(CVLocation);
ProgNode *ProgNodes = 0;
if (g_bLow)
ProgNodes = ProgressiveAlignE(v, GuideTree, msa);
else
ProgressiveAlign(v, GuideTree, msa);
SetCurrentAlignment(msa);
if (0 != g_pstrComputeWeightsFileName)
{
extern void OutWeights(const char *FileName, const MSA &msa);
SetMSAWeightsMuscle(msa);
OutWeights(g_pstrComputeWeightsFileName, msa);
return;
}
ValidateMuscleIds(msa);
if (1 == g_uMaxIters || 2 == uSeqCount)
{
//TextFile fileOut(g_pstrOutFileName, true);
//MHackEnd(msa);
//msa.ToFile(fileOut);
MuscleOutput(msa);
return;
}
if (0 == g_pstrUseTreeFileName)
{
g_bDiags = g_bDiags2;
SetIter(2);
if (g_bLow)
{
if (0 != g_uMaxTreeRefineIters)
RefineTreeE(msa, v, GuideTree, ProgNodes);
}
else
RefineTree(msa, GuideTree);
const char *Tree2 = ValueOpt("Tree2");
if (0 != Tree2)
{
TextFile f(Tree2, true);
GuideTree.ToFile(f);
}
}
SetSeqWeightMethod(g_SeqWeight2);
SetMuscleTree(GuideTree);
if (g_bAnchors)
RefineVert(msa, GuideTree, g_uMaxIters - 2);
else
RefineHoriz(msa, GuideTree, g_uMaxIters - 2, false, false);
#if 0
// Refining by subfamilies is disabled as it didn't give better
// results. I tried doing this before and after RefineHoriz.
// Should get back to this as it seems like this should work.
RefineSubfams(msa, GuideTree, g_uMaxIters - 2);
#endif
ValidateMuscleIds(msa);
ValidateMuscleIds(GuideTree);
//TextFile fileOut(g_pstrOutFileName, true);
//MHackEnd(msa);
//msa.ToFile(fileOut);
MuscleOutput(msa);
}
void RefineW(const MSA &msaIn, MSA &msaOut)
{
const unsigned uSeqCount = msaIn.GetSeqCount();
const unsigned uColCount = msaIn.GetColCount();
// Reserve same nr seqs, 20% more cols
const unsigned uReserveColCount = (uColCount*120)/100;
msaOut.SetSize(uSeqCount, uReserveColCount);
for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex)
{
msaOut.SetSeqName(uSeqIndex, msaIn.GetSeqName(uSeqIndex));
msaOut.SetSeqId(uSeqIndex, msaIn.GetSeqId(uSeqIndex));
}
const unsigned uWindowCount = (uColCount + g_uRefineWindow - 1)/g_uRefineWindow;
if (0 == g_uWindowTo)
g_uWindowTo = uWindowCount - 1;
#if MEMDEBUG
_CrtSetBreakAlloc(1560);
#endif
if (g_uWindowOffset > 0)
{
MSA msaTmp;
MSAFromColRange(msaIn, 0, g_uWindowOffset, msaOut);
}
fprintf(stderr, "\n");
for (unsigned uWindowIndex = g_uWindowFrom; uWindowIndex <= g_uWindowTo; ++uWindowIndex)
{
fprintf(stderr, "Window %d of %d \r", uWindowIndex, uWindowCount);
const unsigned uColFrom = g_uWindowOffset + uWindowIndex*g_uRefineWindow;
unsigned uColTo = uColFrom + g_uRefineWindow - 1;
if (uColTo >= uColCount)
uColTo = uColCount - 1;
assert(uColTo >= uColFrom);
SeqVect v;
SeqVectFromMSACols(msaIn, uColFrom, uColTo, v);
#if MEMDEBUG
_CrtMemState s1;
_CrtMemCheckpoint(&s1);
#endif
MSA msaTmp;
MUSCLE(v, msaTmp);
AppendMSA(msaOut, msaTmp);
if (uWindowIndex == g_uSaveWindow)
{
MSA msaInTmp;
unsigned uOutCols = msaOut.GetColCount();
unsigned un = uColTo - uColFrom + 1;
MSAFromColRange(msaIn, uColFrom, un, msaInTmp);
char fn[256];
sprintf(fn, "win%d_inaln.tmp", uWindowIndex);
TextFile fIn(fn, true);
msaInTmp.ToFile(fIn);
sprintf(fn, "win%d_inseqs.tmp", uWindowIndex);
TextFile fv(fn, true);
v.ToFile(fv);
sprintf(fn, "win%d_outaln.tmp", uWindowIndex);
TextFile fOut(fn, true);
msaTmp.ToFile(fOut);
}
#if MEMDEBUG
void FreeDPMemSPN();
FreeDPMemSPN();
_CrtMemState s2;
_CrtMemCheckpoint(&s2);
_CrtMemState s;
_CrtMemDifference(&s, &s1, &s2);
_CrtMemDumpStatistics(&s);
_CrtMemDumpAllObjectsSince(&s1);
exit(1);
#endif
//#if DEBUG
// AssertMSAEqIgnoreCaseAndGaps(msaInTmp, msaTmp);
//#endif
}
fprintf(stderr, "\n");
// AssertMSAEqIgnoreCaseAndGaps(msaIn, msaOut);//@@uncomment!
}