Merge branch 'dev'

Conflicts:
	main.c

NEWS.md

@@ -1,3 +1,44 @@
+Release 0.7.11 (16 September, 2014)
+-----------------------------------
+
+A major change to BWA-MEM is the support of mapping to ALT contigs. To use this
+feature, users need to manually create a file "<index>.alt" with each line
+giving the name of an ALT contig. In mapping, BWA-MEM considers all chromosomes
+and contigs equally when it finds seeds, constructs chains, extends seeds and
+derives the final alignments. It also uses all hits for the estimation of
+mapping quality of ALT hits. However, BWA-MEM ignores ALT hits when it
+estimates the mapping quality of hits to the primary assembly. As a result,
+having ALT contigs almost has not effect on alignments to the primary assembly
+(seeding may be affected in rare corner cases). At the same time, users may
+get a primary alignment to ALT contigs (no 0x800 flag) if there are no good
+hits to the primary assembly, or get a supplementary alignment to ALT contigs
+if it is better than hits to the primary assembly. Since this release, it is
+recommended to include ALT contigs.
+
+Users may consider to use ALT contigs from GRCh38. I am also constructing a
+non-redundant and more complete set of sequences missing from GRCh38.
+
+Other notable changes to BWA-MEM:
+
+ * Added option `-b` to `bwa index`. This option tunes the batch size used in
+   the construction of BWT. It is advised to use large `-b` for huge reference
+   sequences such as the *nt* database.
+
+ * Optimized for PacBio data. This includes a change to the scoring based on a
+   mini-study done by Aaron Quinlan and a heuristic speedup. Further speedup is
+   possible, but needs more careful investigation.
+
+ * Dropped PacBio read-to-read alignment for now. BWA-MEM is only good at
+   finding the best hit, not all hits. Option `-x pbread` is still available,
+   but not shown on the command line.
+
+ * Added new pre-setting for Oxford Nanopore 2D reads. For small genomes,
+   though, LAST is still more sensitive.
+
+(0.7.11: 16 September 2014, r845)
+
+
+
 Release 0.7.10 (13 July, 2014)
 ------------------------------
 

fastmap.c

@@ -155,7 +155,9 @@ int main_mem(int argc, char *argv[])
 		fprintf(stderr, "       -L INT[,INT]  penalty for 5'- and 3'-end clipping [%d,%d]\n", opt->pen_clip5, opt->pen_clip3);
 		fprintf(stderr, "       -U INT        penalty for an unpaired read pair [%d]\n\n", opt->pen_unpaired);
 		fprintf(stderr, "       -x STR        read type. Setting -x changes multiple parameters unless overriden [null]\n");
-		fprintf(stderr, "                     pacbio: -k17 -W40 -r10 -A1 -B1 -O1 -E1 -L0\n");
+		fprintf(stderr, "                     pacbio: -k17 -W40 -r10 -A1 -B1 -O1 -E1 -L0  (PacBio reads to ref)\n");
+		fprintf(stderr, "                     ont2d: -k14 -W20 -r10 -A1 -B1 -O1 -E1 -L0  (Oxford Nanopore 2D-reads to ref)\n");
+		fprintf(stderr, "                     intractg: -B9 -O16 -L5  (intra-species contigs to ref)\n");
 //		fprintf(stderr, "                     pbread: -k13 -W40 -c1000 -r10 -A1 -B1 -O1 -E1 -N25 -FeaD.001\n");
 		fprintf(stderr, "\nInput/output options:\n\n");
 		fprintf(stderr, "       -p            first query file consists of interleaved paired-end sequences\n");
@@ -181,23 +183,33 @@ int main_mem(int argc, char *argv[])
 	}
 
 	if (mode) {
-		if (strcmp(mode, "pacbio") == 0 || strcmp(mode, "pbref") == 0 || strcmp(mode, "pbread1") == 0 || strcmp(mode, "pbread") == 0) {
+		if (strcmp(mode, "intractg") == 0) {
-			if (!opt0.a) opt->a = 1, opt0.a = 1;
+			if (!opt0.o_del) opt->o_del = 16;
-			update_a(opt, &opt0);
+			if (!opt0.o_ins) opt->o_ins = 16;
+			if (!opt0.b) opt->b = 9;
+			if (!opt0.pen_clip5) opt->pen_clip5 = 5;
+			if (!opt0.pen_clip3) opt->pen_clip3 = 5;
+		} else if (strcmp(mode, "pacbio") == 0 || strcmp(mode, "pbref") == 0 || strcmp(mode, "pbread") == 0 || strcmp(mode, "ont2d") == 0) {
 			if (!opt0.o_del) opt->o_del = 1;
 			if (!opt0.e_del) opt->e_del = 1;
 			if (!opt0.o_ins) opt->o_ins = 1;
 			if (!opt0.e_ins) opt->e_ins = 1;
 			if (!opt0.b) opt->b = 1;
 			if (opt0.split_factor == 0.) opt->split_factor = 10.;
-			if (!opt0.min_chain_weight) opt->min_chain_weight = 40;
+			if (strcmp(mode, "pbread") == 0) { // pacbio read-to-read setting; NOT working well!
-			if (strcmp(mode, "pbread1") == 0 || strcmp(mode, "pbread") == 0) {
 				opt->flag |= MEM_F_ALL | MEM_F_SELF_OVLP | MEM_F_ALN_REG;
+				if (!opt0.min_chain_weight) opt->min_chain_weight = 40;
 				if (!opt0.max_occ) opt->max_occ = 1000;
 				if (!opt0.min_seed_len) opt->min_seed_len = 13;
 				if (!opt0.max_chain_extend) opt->max_chain_extend = 25;
 				if (opt0.drop_ratio == 0.) opt->drop_ratio = .001;
+			} else if (strcmp(mode, "ont2d") == 0) {
+				if (!opt0.min_chain_weight) opt->min_chain_weight = 20;
+				if (!opt0.min_seed_len) opt->min_seed_len = 14;
+				if (!opt0.pen_clip5) opt->pen_clip5 = 0;
+				if (!opt0.pen_clip3) opt->pen_clip3 = 0;
 			} else {
+				if (!opt0.min_chain_weight) opt->min_chain_weight = 40;
 				if (!opt0.min_seed_len) opt->min_seed_len = 17;
 				if (!opt0.pen_clip5) opt->pen_clip5 = 0;
 				if (!opt0.pen_clip3) opt->pen_clip3 = 0;

main.c

@@ -4,7 +4,7 @@
 #include "utils.h"
 
 #ifndef PACKAGE_VERSION
-#define PACKAGE_VERSION "0.7.10-r837-dirty"
+#define PACKAGE_VERSION "0.7.10-r846-dirty"
 #endif
 
 int bwa_fa2pac(int argc, char *argv[]);