X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=PairedEndModel.h;h=70d14c9de9a97680eb9a683715f887a1d2a1e16f;hb=97554bbac838f2ed578d81f98e421dac0669e74e;hp=e73c03dd779aec9574a20bcb314c3db0eb1d2975;hpb=1794b68d34625b39404d5aba9e72646805b710ce;p=rsem.git

diff --git a/PairedEndModel.h b/PairedEndModel.h
index e73c03d..70d14c9 100644
--- a/PairedEndModel.h
+++ b/PairedEndModel.h
@@ -59,11 +59,11 @@ public:
 		mw = NULL;
 
 		if (isMaster) {
-			ori = new Orientation(params.probF);
 			if (!estRSPD) rspd = new RSPD(estRSPD);
 			mld = new LenDist(params.mate_minL, params.mate_maxL);
 		}
 
+		ori = new Orientation(params.probF);
 		gld = new LenDist(params.minL, params.maxL);
 		if (estRSPD) rspd = new RSPD(estRSPD, params.B);
 		pro = new Profile(params.maxL);
@@ -230,19 +230,24 @@ void PairedEndModel::estimateFromReads(const char* readFN) {
     for (int i = 0; i < 3; i++)
     	if (N[i] > 0) {
     		genReadFileNames(readFN, i, read_type, s, readFs);
-    		ReadReader<PairedEndRead> reader(s, readFs);
+    		ReadReader<PairedEndRead> reader(s, readFs, refs->hasPolyA(), seedLen); // allow calculation of calc_lq() function
 
     		int cnt = 0;
     		while (reader.next(read)) {
     			SingleRead mate1 = read.getMate1();
     			SingleRead mate2 = read.getMate2();
-
-    			mld->update(mate1.getReadLength(), 1.0);
-    			mld->update(mate2.getReadLength(), 1.0);
-
-    			if (i == 0) {
-    				npro->updateC(mate1.getReadSeq());
-    				npro->updateC(mate2.getReadSeq());
+			
+    			if (!read.isLowQuality()) {
+    				mld->update(mate1.getReadLength(), 1.0);
+    				mld->update(mate2.getReadLength(), 1.0);
+			  
+    				if (i == 0) {
+    					npro->updateC(mate1.getReadSeq());
+    					npro->updateC(mate2.getReadSeq());
+    				}
+    			}
+    			else if (verbose && (mate1.getReadLength() < seedLen || mate2.getReadLength() < seedLen)) {
+    				printf("Warning: Read %s is ignored due to at least one of the mates' length < seed length %d!\n", read.getName().c_str(), seedLen);
     			}
 
     			++cnt;
@@ -288,7 +293,7 @@ void PairedEndModel::read(const char* inpF) {
 	FILE *fi = fopen(inpF, "r");
 	if (fi == NULL) { fprintf(stderr, "Cannot open %s! It may not exist.\n", inpF); exit(-1); }
 
-	fscanf(fi, "%d", &val);
+	assert(fscanf(fi, "%d", &val) == 1);
 	assert(val == model_type);
 
 	ori->read(fi);
@@ -298,15 +303,18 @@ void PairedEndModel::read(const char* inpF) {
 	pro->read(fi);
 	npro->read(fi);
 
-	if (fscanf(fi, "%d", &M) == 1) {
-	  mw = new double[M + 1];
-	  for (int i = 0; i <= M; i++) fscanf(fi, "%lf", &mw[i]);
+	if (fscanf(fi, "%d", &val) == 1) {
+		if (M == 0) M = val;
+		if (M == val) {
+			mw = new double[M + 1];
+			for (int i = 0; i <= M; i++) assert(fscanf(fi, "%lf", &mw[i]) == 1);
+		}
 	}
 
 	fclose(fi);
 }
 
-//Only master node can call
+//Only master node can call. Only be called at EM.cpp
 void PairedEndModel::write(const char* outF) {
 	FILE *fo = fopen(outF, "w");
 
@@ -382,9 +390,8 @@ bool PairedEndModel::simulate(int rid, PairedEndRead& read, int& sid) {
 		readseq2 = pro->simulate(sampler, mateL2, m2pos, m2dir, ref);
 	}
 
-	std::ostringstream stdout;
-	stdout<<rid<<"_"<<dir<<"_"<<sid<<"_"<<pos<<"_"<<insertL;
-	name = stdout.str();
+	strout<<rid<<"_"<<dir<<"_"<<sid<<"_"<<pos<<"_"<<insertL;
+	name = strout.str();
 
 	read = PairedEndRead(SingleRead(name + "/1", readseq1), SingleRead(name + "/2", readseq2));
 
@@ -400,39 +407,39 @@ void PairedEndModel::finishSimulation() {
 }
 
 void PairedEndModel::calcMW() {
-  assert(seedLen >= OLEN && mld->getMinL() >= seedLen);
-
-  memset(mw, 0, sizeof(double) * (M + 1));
-  mw[0] = 1.0;
-
-  for (int i = 1; i <= M; i++) { 
-    RefSeq& ref = refs->getRef(i);
-    int totLen = ref.getTotLen();
-    int fullLen = ref.getFullLen();
-    int end = std::min(fullLen, totLen - gld->getMinL() + 1);
-    double value = 0.0;
-    int minL, maxL;
-    int effL, pfpos;
-
-    //seedPos is fpos here
-    for (int seedPos = 0; seedPos < end; seedPos++) 
-      if (ref.getMask(seedPos)) {
-	minL = gld->getMinL();
-	maxL = std::min(gld->getMaxL(), totLen - seedPos);
-	pfpos = seedPos;
-	for (int fragLen = minL; fragLen <= maxL; fragLen++) {
-	  effL = std::min(fullLen, totLen - fragLen + 1); 
-	  value += gld->getAdjustedProb(fragLen, totLen) * rspd->getAdjustedProb(pfpos, effL, fullLen); 
+	assert(mld->getMinL() >= seedLen);
+
+	memset(mw, 0, sizeof(double) * (M + 1));
+	mw[0] = 1.0;
+
+	for (int i = 1; i <= M; i++) {
+		RefSeq& ref = refs->getRef(i);
+		int totLen = ref.getTotLen();
+		int fullLen = ref.getFullLen();
+		int end = std::min(fullLen, totLen - gld->getMinL() + 1);
+		double value = 0.0;
+		int minL, maxL;
+		int effL, pfpos;
+
+		//seedPos is fpos here
+		for (int seedPos = 0; seedPos < end; seedPos++)
+			if (ref.getMask(seedPos)) {
+				minL = gld->getMinL();
+				maxL = std::min(gld->getMaxL(), totLen - seedPos);
+				pfpos = seedPos;
+				for (int fragLen = minL; fragLen <= maxL; fragLen++) {
+					effL = std::min(fullLen, totLen - fragLen + 1);
+					value += gld->getAdjustedProb(fragLen, totLen) * rspd->getAdjustedProb(pfpos, effL, fullLen);
+				}
+			}
+
+		mw[i] = 1.0 - value;
+
+		if (mw[i] < 1e-8) {
+			//fprintf(stderr, "Warning: %dth reference sequence is masked for almost all positions!\n", i);
+			mw[i] = 0.0;
+		}
 	}
-      }
-    
-    mw[i] = 1.0 - value;
-
-    if (mw[i] < 1e-8) { 
-      //fprintf(stderr, "Warning: %dth reference sequence is masked for almost all positions!\n", i);
-      mw[i] = 0.0;
-    }
-  }
 }
 
 #endif /* PAIREDENDMODEL_H_ */