Fixed a bug with genome generation for annotations without spliece ju…

…nctions.

CHANGES

@@ -1,3 +1,5 @@
+Fixed a bug with genome generation for annotations without spliece junctions.
+For STARlong, increased compilation-time max read length to 500000 and max number of exons to 1000
 Implemented on the fly insertion of the extra sequences into the genome indexes.
 Implemented --alignEndsType Extend3pOfRead1 option for full extension of the 3' end of read 1.
 Fixed a bug in the --alignEndsType Extend5pOfRead1 option.

extras/scripts/sjFromSAMcollapseUandM_inclOverlaps.awk

@@ -0,0 +1,55 @@
+BEGIN {
+   OFS="\t";
+   mapqU=255;
+}
+{
+if (substr($1,1,1)!="@") {
+
+    m=and($2,0x80)/0x80+1;
+
+    if ($1!=readNameOld) delete readSJs;
+    readNameOld=$1;
+
+    n=split($6,L,/[A-Z]/)-1;
+    split($6,C,/[0-9]*/);
+    t=1;g=$4;
+    for (k=1;k<=n;k++) {#scan through CIGAR operations
+        if (C[k+1]=="S" || C[k+1]=="I") {
+           t+=L[k];
+        } else if (C[k+1]=="D") {
+           g+=L[k];
+        } else if (C[k+1]=="N") {
+           sj1=$3 "\t" g "\t" g+L[k]-1;
+           readSJs[sj1]++;
+
+           if (readSJs[sj1]==1) {#only count this junction if it has nto been counted for the same read
+               SJ[sj1]=1;
+               if ($5>=mapqU) {
+                   SJu[sj1]++;
+               } else {
+                   SJm[sj1]++;
+               };
+           };
+
+           if ($5>=mapqU) {
+               SJu1[sj1]++;
+           } else {
+               SJm1[sj1]++;
+           };
+
+           g+=L[k];
+
+        } else { # M operation
+           g+=L[k];
+           t+=L[k];
+        };
+    };
+};
+};
+END {
+
+for (ii in SJ) {
+    print ii, SJu[ii]+0, SJm[ii]+0, SJu1[ii]+0, SJm1[ii]+0;
+};
+
+};

source/IncludeDefine.h

@@ -104,7 +104,7 @@ typedef uint8_t uint8;
 #define MAX_N_MATES 2
 #define DEF_readNameLengthMax 50000
 #if defined COMPILE_FOR_LONG_READS
-    #define DEF_readSeqLengthMax 50000
+    #define DEF_readSeqLengthMax 500000
 #else
     #define DEF_readSeqLengthMax 500
 #endif

source/genomeGenerate.cpp

@@ -405,6 +405,7 @@ void genomeGenerate(Parameters *P) {
 
         //write an extra 0 at the end of the array, filling the last bytes that otherwise are not accessible, but will be written to disk
         //this is - to avoid valgrind complaints. Note that SA2 is allocated with plenty of space to spare.
+        P->nSAbyte=mainGenome.SA.lengthByte;
         SA2.writePacked(P->nSA,0);
     };