VSEARCH 1.0.14: Fix for ignored strand in multiple sequence alignment…

… after clustering

README.md

@@ -32,22 +32,22 @@ If you can't find an answer in the VSEARCH documentation, please visit the [VSEA
 
 In the example below, VSEARCH will identify sequences in the file database.fsa that are at least 90% identical on the plus strand to the query sequences in the file queries.fsa and write the results to the file alnout.txt.
 
-`./vsearch-1.0.13-linux-x86_64 --usearch_global queries.fsa --db database.fsa --id 0.9 --alnout alnout.txt`
+`./vsearch-1.0.14-linux-x86_64 --usearch_global queries.fsa --db database.fsa --id 0.9 --alnout alnout.txt`
 
 ## Download and install
 
 The latest releases of VSEARCH are available [here](https://github.com/torognes/vsearch/releases).
 
-Binary executables of VSEARCH are available in the `bin` folder for [GNU/Linux on x86-64 systems](https://github.com/torognes/vsearch/blob/master/bin/vsearch-1.0.13-linux-x86_64) and [Apple Mac OS X on x86-64 systems](https://github.com/torognes/vsearch/blob/master/bin/vsearch-1.0.13-osx-x86_64). These executables include support for  input files compressed by zlib and bzip2 (with files usually ending in .gz or .bz2).
+Binary executables of VSEARCH are available in the `bin` folder for [GNU/Linux on x86-64 systems](https://github.com/torognes/vsearch/blob/master/bin/vsearch-1.0.14-linux-x86_64) and [Apple Mac OS X on x86-64 systems](https://github.com/torognes/vsearch/blob/master/bin/vsearch-1.0.14-osx-x86_64). These executables include support for  input files compressed by zlib and bzip2 (with files usually ending in .gz or .bz2).
 
 Download the appropriate executable and make a symbolic link in a folder included in your `$PATH` from `vsearch` to the appropriate binary. You may use the following commands (assuming `~/bin` is in your `$PATH`):
 
 ```sh
 cd ~
 mkdir -p bin
 cd bin
-wget https://github.com/torognes/vsearch/releases/download/v1.0.13/vsearch-1.0.13-linux-x86_64
-ln -s vsearch-1.0.13-linux-x86_64 vsearch
+wget https://github.com/torognes/vsearch/releases/download/v1.0.14/vsearch-1.0.14-linux-x86_64
+ln -s vsearch-1.0.14-linux-x86_64 vsearch
 ```
 
 Substitute `linux` with `osx` in those lines if you're on a Mac.

doc/vsearch.1

@@ -1,5 +1,5 @@
 .\" ============================================================================
-.TH vsearch 1 "February 17, 2015" "version 1.0.13" "USER COMMANDS"
+.TH vsearch 1 "February 17, 2015" "version 1.0.14" "USER COMMANDS"
 .\" ============================================================================
 .SH NAME
 vsearch \(em chimera detection, clustering, dereplication, masking, pairwise alignment, searching, shuffling and sorting of amplicons from metagenomic projects.
@@ -1524,10 +1524,16 @@ clustering, pairwise alignment and searching).
 .BR v1.0.12\~ "released February 6th, 2015"
 Temporarily fixes a problem with long headers in FASTA files.
 .TP
-.BR v1.0.13\~ "released February 6th, 2015"
+.BR v1.0.13\~ "released February 17th, 2015"
 Fix a memory allocation problem when computing multiple sequence alignments
 with the --msaout and --consout options, as well as a memory leak.
 Also increased line buffer for reading FASTA files to 4MB.
+.TP
+.BR v1.0.14\~ "released February 17th, 2015"
+Fix a bug where the multiple alignment and consensus sequence computed
+after clustering ignored the strand of the sequences.
+Also decreased size of line buffer for reading FASTA files to 1MB again
+due to excessive stack memory usage.
 .LP
 .\" ============================================================================
 .\" TODO:

src/cluster.cc

@@ -29,6 +29,7 @@ typedef struct clusterinfo_s
   int seqno;
   int clusterno;
   char * cigar;
+  int strand;
 } clusterinfo_t;
 
 static clusterinfo_t * clusterinfo = 0;
@@ -711,6 +712,7 @@ void cluster_core_parallel()
               clusterinfo[seqno].seqno = seqno;
               clusterinfo[seqno].clusterno = clusterinfo[target].clusterno;
               clusterinfo[seqno].cigar = best->nwalignment;
+              clusterinfo[seqno].strand = best->strand;
               best->nwalignment = 0;
             }
           else
@@ -724,6 +726,7 @@ void cluster_core_parallel()
               clusterinfo[seqno].seqno = seqno;
               clusterinfo[seqno].clusterno = clusters;
               clusterinfo[seqno].cigar = 0;
+              clusterinfo[seqno].strand = 0;
 
               /* add current sequence to database */
               dbindex_addsequence(seqno);
@@ -826,13 +829,15 @@ void cluster_core_serial()
           clusterinfo[seqno].seqno = seqno;
           clusterinfo[seqno].clusterno = clusterinfo[target].clusterno;
           clusterinfo[seqno].cigar = best->nwalignment;
+          clusterinfo[seqno].strand = best->strand;
           best->nwalignment = 0;
         }
       else
         {
           clusterinfo[seqno].seqno = seqno;
           clusterinfo[seqno].clusterno = clusters;
           clusterinfo[seqno].cigar = 0;
+          clusterinfo[seqno].strand = 0;
           dbindex_addsequence(seqno);
           cluster_core_results_nohit(clusters,
                                      si_p->query_head,
@@ -1130,6 +1135,7 @@ void cluster(char * dbname,
           int clusterno = clusterinfo[i].clusterno;
           int seqno = clusterinfo[i].seqno;
           char * cigar = clusterinfo[i].cigar;
+          int strand = clusterinfo[i].strand;
 
           if (clusterno != lastcluster)
             {
@@ -1145,6 +1151,7 @@ void cluster(char * dbname,
           /* add current sequence to the cluster */
           msa_target_list[msa_target_count].seqno = seqno;
           msa_target_list[msa_target_count].cigar = cigar;
+          msa_target_list[msa_target_count].strand = strand;
           msa_target_count++;
 
           progress_update(i);

src/db.cc

@@ -34,7 +34,7 @@ seqinfo_t * seqindex = 0;
 static char * datap = 0;
 
 #define MEMCHUNK 10485760
-#define LINEALLOC 4194304
+#define LINEALLOC 1048576
 
 regex_t db_regexp;
 

src/msa.cc

@@ -151,6 +151,19 @@ void msa(FILE * fp_msaout, FILE * fp_consout,
   memset(profile, 0, 4 * sizeof(int) * alnlen);
   aln = (char *) xmalloc(alnlen+1);
   char * cons = (char *) xmalloc(alnlen+1);
+
+  /* Find longest target sequence on reverse strand and allocate buffer */
+  long longest_reversed = 0;
+  for(int i=0; i < target_count; i++)
+    if (target_list[i].strand)
+      {
+        long len = db_getsequencelen(target_list[i].seqno);
+        if (len > longest_reversed)
+          longest_reversed = len;
+      }
+  char * rc_buffer = 0;
+  if (longest_reversed > 0)
+    rc_buffer = (char*) xmalloc(longest_reversed + 1);
 
   /* blank line before each msa */
   if (fp_msaout)
@@ -161,6 +174,13 @@ void msa(FILE * fp_msaout, FILE * fp_consout,
       int target_seqno = target_list[j].seqno;
       char * target_seq = db_getsequence(target_seqno);
 
+      if (target_list[j].strand)
+        {
+          reverse_complement(rc_buffer, target_seq,
+                             db_getsequencelen(target_seqno));
+          target_seq = rc_buffer;
+        }
+
       int inserted = 0;
       int qpos = 0;
       int tpos = 0;
@@ -235,6 +255,8 @@ void msa(FILE * fp_msaout, FILE * fp_consout,
         }
     }  
 
+  if (rc_buffer)
+    free(rc_buffer);
 
   /* consensus */
 

src/msa.h

@@ -1,5 +1,5 @@
 /*
-    Copyright (C) 2014 Torbjorn Rognes
+    Copyright (C) 2014-2015 Torbjorn Rognes
 
     This program is free software: you can redistribute it and/or modify
     it under the terms of the GNU Affero General Public License as
@@ -23,6 +23,7 @@ struct msa_target_s
 {
   int seqno;
   char * cigar;
+  int strand;
 };
 
 void msa(FILE * fp_msaout, FILE * fp_consout,

src/query.cc

@@ -26,7 +26,7 @@
    be overwritten on the next call of query_getnext. */
 
 #define MEMCHUNK 4096
-#define LINEALLOC 4194304
+#define LINEALLOC 1048576
 
 extern unsigned int chrstatus[256];
 

src/util.cc

@@ -183,19 +183,9 @@ void reverse_complement(char * rc, char * seq, long len)
      The memory for rc must be long enough for the rc of the sequence
      (identical to the length of seq + 1. */
 
-#if 0
-  printf("Sequence:\n");
-  fprint_fasta_seq_only(stdout, seq, len, 60);
-#endif
   for(long i=0; i<len; i++)
-    {
-      rc[i] = chrmap_complement[(int)(seq[len-1-i])];
-    }
+    rc[i] = chrmap_complement[(int)(seq[len-1-i])];
   rc[len] = 0;
-#if 0
-  printf("RC:\n");
-  fprint_fasta_seq_only(stdout, rc, len, 60);
-#endif
 }
 
 #ifdef HAVE_BZLIB

src/vsearch.h

@@ -84,7 +84,7 @@
 #include "allpairs.h"
 
 #define PROG_NAME "vsearch"
-#define PROG_VERSION "v1.0.13"
+#define PROG_VERSION "v1.0.14"
 
 #ifdef __APPLE__
 #define PROG_ARCH "osx_x86_64"