Exemple #1
0
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);
	}
Exemple #2
0
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);
	}
Exemple #3
0
void ProgAlignSubFams()
	{
	MSA msaOut;

	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();

	PTR_SCOREMATRIX UserMatrix = 0;
	if (0 != g_pstrMatrixFileName.get())
		{
		const char *FileName = g_pstrMatrixFileName.get();
		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;

	if (ALPHA_DNA == Alpha || ALPHA_RNA == Alpha)
		{
		SetPPScore(PPSCORE_SPN);
		g_Distance1.get() = DISTANCE_Kmer4_6;
		}

	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 (uSeqCount > 1)
		MHackStart(v);

	if (0 == uSeqCount)
		{
		msaOut.Clear();
		return;
		}

	if (1 == uSeqCount && ALPHA_Amino == Alpha)
		{
		const Seq &s = v.GetSeq(0);
		msaOut.FromSeq(s);
		return;
		}

	Tree GuideTree;
	TreeFromSeqVect(v, GuideTree, g_Cluster1.get(), g_Distance1.get(), g_Root1.get());
	SetMuscleTree(GuideTree);

	MSA msa;
	if (g_bLow.get())
		{
		ProgNode *ProgNodes = 0;
		ProgNodes = ProgressiveAlignE(v, GuideTree, msa);
		delete[] ProgNodes;
		}
	else
		ProgressiveAlign(v, GuideTree, msa);
	SetCurrentAlignment(msa);
	TreeFromMSA(msa, GuideTree, g_Cluster2.get(), g_Distance2.get(), g_Root2.get());
	SetMuscleTree(GuideTree);

	unsigned *SubFams = new unsigned[uSeqCount];
	unsigned uSubFamCount;
	SubFam(GuideTree, g_uMaxSubFamCount.get(), SubFams, &uSubFamCount);

	SetProgressDesc("Align node");
	const unsigned uNodeCount = 2*uSeqCount - 1;

	ProgNode *ProgNodes = new ProgNode[uNodeCount];
	bool *NodeIsSubFam = new bool[uNodeCount];
	bool *NodeInSubFam = new bool[uNodeCount];

	for (unsigned i = 0; i < uNodeCount; ++i)
		{
		NodeIsSubFam[i] = false;
		NodeInSubFam[i] = false;
		}

	for (unsigned i = 0; i < uSubFamCount; ++i)
		{
		unsigned uNodeIndex = SubFams[i];
		assert(uNodeIndex < uNodeCount);
		NodeIsSubFam[uNodeIndex] = true;
		SetInFam(GuideTree, uNodeIndex, NodeInSubFam);
		}

	unsigned uJoin = 0;
	unsigned uTreeNodeIndex = GuideTree.FirstDepthFirstNode();
	do
		{
		if (NodeIsSubFam[uTreeNodeIndex])
			{
#if	TRACE
			Log("Node %d: align subfam\n", uTreeNodeIndex);
#endif
			ProgNode &Node = ProgNodes[uTreeNodeIndex];
			AlignSubFam(v, GuideTree, uTreeNodeIndex, Node.m_MSA);
			Node.m_uLength = Node.m_MSA.GetColCount();
			}
		else if (!NodeInSubFam[uTreeNodeIndex])
			{
#if	TRACE
			Log("Node %d: align two subfams\n", uTreeNodeIndex);
#endif
			Progress(uJoin, uSubFamCount - 1);
			++uJoin;

			const unsigned uMergeNodeIndex = uTreeNodeIndex;
			ProgNode &Parent = ProgNodes[uMergeNodeIndex];

			const unsigned uLeft = GuideTree.GetLeft(uTreeNodeIndex);
			const unsigned uRight = GuideTree.GetRight(uTreeNodeIndex);

			ProgNode &Node1 = ProgNodes[uLeft];
			ProgNode &Node2 = ProgNodes[uRight];

			PWPath Path;
			AlignTwoMSAs(Node1.m_MSA, Node2.m_MSA, Parent.m_MSA, Path);
			Parent.m_uLength = Parent.m_MSA.GetColCount();

			Node1.m_MSA.Clear();
			Node2.m_MSA.Clear();
			}
		else
			{
#if	TRACE
			Log("Node %d: in subfam\n", uTreeNodeIndex);
#endif
			;
			}
		uTreeNodeIndex = GuideTree.NextDepthFirstNode(uTreeNodeIndex);
		}
	while (NULL_NEIGHBOR != uTreeNodeIndex);
	ProgressStepsDone();

	unsigned uRootNodeIndex = GuideTree.GetRootNodeIndex();
	ProgNode &RootProgNode = ProgNodes[uRootNodeIndex];

	TextFile fOut(g_pstrOutFileName.get(), true);
	MHackEnd(RootProgNode.m_MSA);
	RootProgNode.m_MSA.ToFile(fOut);

	delete[] NodeInSubFam;
	delete[] NodeIsSubFam;
	delete[] ProgNodes;
	delete[] SubFams;

	ProgNodes = 0;
	NodeInSubFam = 0;
	NodeIsSubFam = 0;
	SubFams = 0;
	}
Exemple #4
0
void Refine()
	{
	SetOutputFileName(g_pstrOutFileName.get());
	SetInputFileName(g_pstrInFileName.get());
	SetStartTime();

	SetMaxIters(g_uMaxIters.get());
	SetSeqWeightMethod(g_SeqWeight1.get());

	TextFile fileIn(g_pstrInFileName.get());
	MSA msa;
	msa.FromFile(fileIn);

	const unsigned uSeqCount = msa.GetSeqCount();
	if (0 == uSeqCount)
		Quit("No sequences in input file");

	ALPHA Alpha = ALPHA_Undefined;
	switch (g_SeqType.get())
		{
	case SEQTYPE_Auto:
		Alpha = msa.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 SeqType");
		}
	SetAlpha(Alpha);
	msa.FixAlpha();

	SetPPScore();
	if (ALPHA_DNA == Alpha || ALPHA_RNA == Alpha)
		SetPPScore(PPSCORE_SPN);

	MSA::SetIdCount(uSeqCount);

// Initialize sequence ids.
// From this point on, ids must somehow propogate from here.
	for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex)
		msa.SetSeqId(uSeqIndex, uSeqIndex);
	SetMuscleInputMSA(msa);

	Tree GuideTree;
	TreeFromMSA(msa, GuideTree, g_Cluster2.get(), g_Distance2.get(), g_Root2.get());
	SetMuscleTree(GuideTree);

	if (g_bAnchors.get())
		RefineVert(msa, GuideTree, g_uMaxIters.get());
	else
		RefineHoriz(msa, GuideTree, g_uMaxIters.get(), false, false);

	ValidateMuscleIds(msa);
	ValidateMuscleIds(GuideTree);

//	TextFile fileOut(g_pstrOutFileName.get(), true);
//	msa.ToFile(fileOut);
	MuscleOutput(msa);
	}
Exemple #5
0
void DoRefineW()
	{
	SetOutputFileName(g_pstrOutFileName);
	SetInputFileName(g_pstrInFileName);
	SetStartTime();

	SetMaxIters(g_uMaxIters);
	SetSeqWeightMethod(g_SeqWeight1);

	TextFile fileIn(g_pstrInFileName);
	MSA msa;
	msa.FromFile(fileIn);

	const unsigned uSeqCount = msa.GetSeqCount();
	if (0 == uSeqCount)
		Quit("No sequences in input file");

	MSA::SetIdCount(uSeqCount);

// Initialize sequence ids.
// From this point on, ids must somehow propogate from here.
	for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex)
		msa.SetSeqId(uSeqIndex, uSeqIndex);
	SetMuscleInputMSA(msa);

	ALPHA Alpha = ALPHA_Undefined;
	switch (g_SeqType)
		{
	case SEQTYPE_Auto:
		Alpha = msa.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 SeqType");
		}
	SetAlpha(Alpha);
	msa.FixAlpha();

	if (ALPHA_DNA == Alpha || ALPHA_RNA == Alpha)
		SetPPScore(PPSCORE_SPN);

	MSA msaOut;
	RefineW(msa, msaOut);

//	ValidateMuscleIds(msa);

//	TextFile fileOut(g_pstrOutFileName, true);
//	msaOut.ToFile(fileOut);
	MuscleOutput(msaOut);
	}