void MHackEnd(MSA &msa) { if (ALPHA_Amino != g_Alpha) return; if (0 == M) return; const unsigned uSeqCount = msa.GetSeqCount(); const unsigned uColCount = msa.GetColCount(); for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex) { unsigned uId = msa.GetSeqId(uSeqIndex); if (M[uId]) { for (unsigned uColIndex = 0; uColIndex < uColCount; ++uColIndex) { if (!msa.IsGap(uSeqIndex, uColIndex)) { msa.SetChar(uSeqIndex, uColIndex, 'M'); break; } } } } delete[] M; M = 0; }
void AssertMSAEq(const MSA &msa1, const MSA &msa2) { const unsigned uSeqCount1 = msa1.GetSeqCount(); const unsigned uSeqCount2 = msa2.GetSeqCount(); if (uSeqCount1 != uSeqCount2) Quit("Seq count differs"); for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount1; ++uSeqIndex) { Seq seq1; msa1.GetSeq(uSeqIndex, seq1); unsigned uId = msa1.GetSeqId(uSeqIndex); unsigned uSeqIndex2 = msa2.GetSeqIndex(uId); Seq seq2; msa2.GetSeq(uSeqIndex2, seq2); if (!seq1.Eq(seq2)) { Log("Input:\n"); seq1.LogMe(); Log("Output:\n"); seq2.LogMe(); Quit("Seq %s differ ", msa1.GetSeqName(uSeqIndex)); } } }
void MSAFromColRange(const MSA &msaIn, unsigned uFromColIndex, unsigned uColCount, MSA &msaOut) { const unsigned uSeqCount = msaIn.GetSeqCount(); const unsigned uInColCount = msaIn.GetColCount(); if (uFromColIndex + uColCount - 1 > uInColCount) Quit("MSAFromColRange, out of bounds"); msaOut.SetSize(uSeqCount, uColCount); for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex) { const char *ptrName = msaIn.GetSeqName(uSeqIndex); unsigned uId = msaIn.GetSeqId(uSeqIndex); msaOut.SetSeqName(uSeqIndex, ptrName); msaOut.SetSeqId(uSeqIndex, uId); for (unsigned uColIndex = 0; uColIndex < uColCount; ++uColIndex) { const char c = msaIn.GetChar(uSeqIndex, uFromColIndex + uColIndex); msaOut.SetChar(uSeqIndex, uColIndex, c); } } }
// Append msa2 at the end of msa1 void AppendMSA(MSA &msa1, const MSA &msa2) { const unsigned uSeqCount = msa1.GetSeqCount(); const unsigned uColCount1 = msa1.GetColCount(); const unsigned uColCount2 = msa2.GetColCount(); const unsigned uColCountCat = uColCount1 + uColCount2; for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex) { unsigned uId = msa1.GetSeqId(uSeqIndex); unsigned uSeqIndex2; bool bFound = msa2.GetSeqIndex(uId, &uSeqIndex2); if (bFound) { for (unsigned uColIndex = 0; uColIndex < uColCount2; ++uColIndex) { const char c = msa2.GetChar(uSeqIndex2, uColIndex); msa1.SetChar(uSeqIndex, uColCount1 + uColIndex, c); } } else { for (unsigned uColIndex = 0; uColIndex < uColCount2; ++uColIndex) msa1.SetChar(uSeqIndex, uColCount1 + uColIndex, '-'); } } }
void SetThreeWayWeightsMuscle(MSA &msa) { MuscleContext *ctx =getMuscleContext(); const Tree* &g_ptrMuscleTree = ctx->msa2.g_ptrMuscleTree; unsigned &g_uTreeSplitNode1 = ctx->msa2.g_uTreeSplitNode1; unsigned &g_uTreeSplitNode2 = ctx->msa2.g_uTreeSplitNode2; if (NULL_NEIGHBOR == g_uTreeSplitNode1 || NULL_NEIGHBOR == g_uTreeSplitNode2) { msa.SetHenikoffWeightsPB(); return; } const unsigned uMuscleSeqCount = g_ptrMuscleTree->GetLeafCount(); WEIGHT *Weights = new WEIGHT[uMuscleSeqCount]; CalcThreeWayWeights(*g_ptrMuscleTree, g_uTreeSplitNode1, g_uTreeSplitNode2, Weights); const unsigned uMSASeqCount = msa.GetSeqCount(); for (unsigned uSeqIndex = 0; uSeqIndex < uMSASeqCount; ++uSeqIndex) { const unsigned uId = msa.GetSeqId(uSeqIndex); if (uId >= uMuscleSeqCount) Quit("SetThreeWayWeightsMuscle: id out of range"); msa.SetSeqWeight(uSeqIndex, Weights[uId]); } #if LOCAL_VERBOSE { Log("SetThreeWayWeightsMuscle\n"); for (unsigned n = 0; n < uMSASeqCount; ++n) { const unsigned uId = msa.GetSeqId(n); Log("%20.20s %6.3f\n", msa.GetSeqName(n), Weights[uId]); } } #endif msa.NormalizeWeights((WEIGHT) 1.0); delete[] Weights; }
static void SeqFromMSACols(const MSA &msa, unsigned uSeqIndex, unsigned uColFrom, unsigned uColTo, Seq &s) { s.Clear(); s.SetName(msa.GetSeqName(uSeqIndex)); s.SetId(msa.GetSeqId(uSeqIndex)); for (unsigned uColIndex = uColFrom; uColIndex <= uColTo; ++uColIndex) { char c = msa.GetChar(uSeqIndex, uColIndex); if (!IsGapChar(c)) s.AppendChar(c); } }
void SetClustalWWeightsMuscle(MSA &msa) { if (0 == g_MuscleWeights) Quit("g_MuscleWeights = 0"); const unsigned uSeqCount = msa.GetSeqCount(); for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex) { const unsigned uId = msa.GetSeqId(uSeqIndex); if (uId >= g_uMuscleIdCount) Quit("SetClustalWWeightsMuscle: id out of range"); msa.SetSeqWeight(uSeqIndex, g_MuscleWeights[uId]); } msa.NormalizeWeights((WEIGHT) 1.0); }
void convertMSA2MAlignment(MSA& msa, const DNAAlphabet* al, MultipleSequenceAlignment& res) { assert(res->isEmpty()); MuscleContext *ctx = getMuscleContext(); res->setAlphabet(al); ctx->output_uIds.clear(); for(int i=0, n = msa.GetSeqCount(); i < n; i++) { QString name = msa.GetSeqName(i); QByteArray seq; seq.reserve(msa.GetColCount()); for (int j = 0, m = msa.GetColCount(); j < m ; j++) { char c = msa.GetChar(i, j); seq.append(c); } ctx->output_uIds.append(ctx->tmp_uIds[msa.GetSeqId(i)]); res->addRow(name, seq); } }
// Append msa2 at the end of msa1 void MSAAppend(MSA &msa1, const MSA &msa2) { const unsigned uSeqCount = msa1.GetSeqCount(); const unsigned uColCount1 = msa1.GetColCount(); const unsigned uColCount2 = msa2.GetColCount(); for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCount; ++uSeqIndex) { unsigned uId = msa1.GetSeqId(uSeqIndex); unsigned uSeqIndex2 = msa2.GetSeqIndex(uId); for (unsigned uColIndex = 0; uColIndex < uColCount2; ++uColIndex) { const char c = msa2.GetChar(uSeqIndex2, uColIndex); msa1.SetChar(uSeqIndex, uColCount1 + uColIndex, c); } } }
void MSAFromSeqSubset(const MSA &msaIn, const unsigned uSeqIndexes[], unsigned uSeqCount, MSA &msaOut) { const unsigned uColCount = msaIn.GetColCount(); msaOut.SetSize(uSeqCount, uColCount); for (unsigned uSeqIndexOut = 0; uSeqIndexOut < uSeqCount; ++uSeqIndexOut) { unsigned uSeqIndexIn = uSeqIndexes[uSeqIndexOut]; const char *ptrName = msaIn.GetSeqName(uSeqIndexIn); unsigned uId = msaIn.GetSeqId(uSeqIndexIn); msaOut.SetSeqName(uSeqIndexOut, ptrName); msaOut.SetSeqId(uSeqIndexOut, uId); for (unsigned uColIndex = 0; uColIndex < uColCount; ++uColIndex) { const char c = msaIn.GetChar(uSeqIndexIn, uColIndex); msaOut.SetChar(uSeqIndexOut, uColIndex, c); } } }
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 AlignTwoMSAsGivenPathSW(const PWPath &Path, const MSA &msaA, const MSA &msaB, MSA &msaCombined) { msaCombined.Clear(); #if TRACE Log("AlignTwoMSAsGivenPathSW\n"); Log("Template A:\n"); msaA.LogMe(); Log("Template B:\n"); msaB.LogMe(); #endif const unsigned uColCountA = msaA.GetColCount(); const unsigned uColCountB = msaB.GetColCount(); const unsigned uSeqCountA = msaA.GetSeqCount(); const unsigned uSeqCountB = msaB.GetSeqCount(); msaCombined.SetSeqCount(uSeqCountA + uSeqCountB); // Copy sequence names into combined MSA for (unsigned uSeqIndexA = 0; uSeqIndexA < uSeqCountA; ++uSeqIndexA) { msaCombined.SetSeqName(uSeqIndexA, msaA.GetSeqName(uSeqIndexA)); msaCombined.SetSeqId(uSeqIndexA, msaA.GetSeqId(uSeqIndexA)); } for (unsigned uSeqIndexB = 0; uSeqIndexB < uSeqCountB; ++uSeqIndexB) { msaCombined.SetSeqName(uSeqCountA + uSeqIndexB, msaB.GetSeqName(uSeqIndexB)); msaCombined.SetSeqId(uSeqCountA + uSeqIndexB, msaB.GetSeqId(uSeqIndexB)); } unsigned uColIndexA = 0; unsigned uColIndexB = 0; unsigned uColIndexCombined = 0; const unsigned uEdgeCount = Path.GetEdgeCount(); for (unsigned uEdgeIndex = 0; uEdgeIndex < uEdgeCount; ++uEdgeIndex) { const PWEdge &Edge = Path.GetEdge(uEdgeIndex); #if TRACE Log("\nEdge %u %c%u.%u\n", uEdgeIndex, Edge.cType, Edge.uPrefixLengthA, Edge.uPrefixLengthB); #endif const char cType = Edge.cType; const unsigned uPrefixLengthA = Edge.uPrefixLengthA; unsigned uColCountA = 0; if (uPrefixLengthA > 0) { const unsigned uNodeIndexA = uPrefixLengthA - 1; const unsigned uTplColIndexA = uNodeIndexA; if (uTplColIndexA > uColIndexA) uColCountA = uTplColIndexA - uColIndexA; } const unsigned uPrefixLengthB = Edge.uPrefixLengthB; unsigned uColCountB = 0; if (uPrefixLengthB > 0) { const unsigned uNodeIndexB = uPrefixLengthB - 1; const unsigned uTplColIndexB = uNodeIndexB; if (uTplColIndexB > uColIndexB) uColCountB = uTplColIndexB - uColIndexB; } AppendUnalignedTerminals(msaA, uColIndexA, uColCountA, msaB, uColIndexB, uColCountB, uSeqCountA, uSeqCountB, msaCombined, uColIndexCombined); switch (cType) { case 'M': { assert(uPrefixLengthA > 0); assert(uPrefixLengthB > 0); const unsigned uColA = uPrefixLengthA - 1; const unsigned uColB = uPrefixLengthB - 1; assert(uColIndexA == uColA); assert(uColIndexB == uColB); AppendMatch(msaA, uColIndexA, msaB, uColIndexB, uSeqCountA, uSeqCountB, msaCombined, uColIndexCombined); break; } case 'D': { assert(uPrefixLengthA > 0); const unsigned uColA = uPrefixLengthA - 1; assert(uColIndexA == uColA); AppendDelete(msaA, uColIndexA, uSeqCountA, uSeqCountB, msaCombined, uColIndexCombined); break; } case 'I': { assert(uPrefixLengthB > 0); const unsigned uColB = uPrefixLengthB - 1; assert(uColIndexB == uColB); AppendInsert(msaB, uColIndexB, uSeqCountA, uSeqCountB, msaCombined, uColIndexCombined); break; } default: assert(false); } } unsigned uInsertColCountA = uColCountA - uColIndexA; unsigned uInsertColCountB = uColCountB - uColIndexB; AppendUnalignedTerminals(msaA, uColIndexA, uInsertColCountA, msaB, uColIndexB, uInsertColCountB, uSeqCountA, uSeqCountB, msaCombined, uColIndexCombined); }
// Return true if any changes made bool RefineVert(MSA &msaIn, const Tree &tree, unsigned uIters) { bool bAnyChanges = false; const unsigned uColCountIn = msaIn.GetColCount(); const unsigned uSeqCountIn = msaIn.GetSeqCount(); if (uColCountIn < 3 || uSeqCountIn < 3) return false; unsigned *AnchorCols = new unsigned[uColCountIn]; unsigned uAnchorColCount; SetMSAWeightsMuscle(msaIn); FindAnchorCols(msaIn, AnchorCols, &uAnchorColCount); const unsigned uRangeCount = uAnchorColCount + 1; Range *Ranges = new Range[uRangeCount]; #if TRACE Log("%u ranges\n", uRangeCount); #endif ColsToRanges(AnchorCols, uAnchorColCount, uColCountIn, Ranges); ListVertSavings(uColCountIn, uAnchorColCount, Ranges, uRangeCount); #if TRACE { Log("Anchor cols: "); for (unsigned i = 0; i < uAnchorColCount; ++i) Log(" %u", AnchorCols[i]); Log("\n"); Log("Ranges:\n"); for (unsigned i = 0; i < uRangeCount; ++i) Log("%4u - %4u\n", Ranges[i].m_uBestColLeft, Ranges[i].m_uBestColRight); } #endif delete[] AnchorCols; MSA msaOut; msaOut.SetSize(uSeqCountIn, 0); for (unsigned uSeqIndex = 0; uSeqIndex < uSeqCountIn; ++uSeqIndex) { const char *ptrName = msaIn.GetSeqName(uSeqIndex); unsigned uId = msaIn.GetSeqId(uSeqIndex); msaOut.SetSeqName(uSeqIndex, ptrName); msaOut.SetSeqId(uSeqIndex, uId); } for (unsigned uRangeIndex = 0; uRangeIndex < uRangeCount; ++uRangeIndex) { MSA msaRange; const Range &r = Ranges[uRangeIndex]; const unsigned uFromColIndex = r.m_uBestColLeft; const unsigned uRangeColCount = r.m_uBestColRight - uFromColIndex; if (0 == uRangeColCount) continue; else if (1 == uRangeColCount) { MSAFromColRange(msaIn, uFromColIndex, 1, msaRange); MSAAppend(msaOut, msaRange); continue; } MSAFromColRange(msaIn, uFromColIndex, uRangeColCount, msaRange); #if TRACE Log("\n-------------\n"); Log("Range %u - %u count=%u\n", r.m_uBestColLeft, r.m_uBestColRight, uRangeColCount); Log("Before:\n"); msaRange.LogMe(); #endif bool bLockLeft = (0 != uRangeIndex); bool bLockRight = (uRangeCount - 1 != uRangeIndex); bool bAnyChangesThisBlock = RefineHoriz(msaRange, tree, uIters, bLockLeft, bLockRight); bAnyChanges = (bAnyChanges || bAnyChangesThisBlock); #if TRACE Log("After:\n"); msaRange.LogMe(); #endif MSAAppend(msaOut, msaRange); #if TRACE Log("msaOut after Cat:\n"); msaOut.LogMe(); #endif } #if DEBUG // Sanity check AssertMSAEqIgnoreCaseAndGaps(msaIn, msaOut); #endif delete[] Ranges; if (bAnyChanges) msaIn.Copy(msaOut); return bAnyChanges; }
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! }