New port: biology/stringtie: Transcript assembly and quantification for RNA-seq

PR:		229480
Submitted by:	Motomichi Matsuzaki <mzaki@niid.go.jp>

7 files changed, 156 insertions, 0 deletions

diff --git a/biology/stringtie/Makefile b/biology/stringtie/Makefile
new file mode 100644
index 000000000000..57c2fdb096ed
--- /dev/null
+++ b/biology/stringtie/Makefile
@@ -0,0 +1,24 @@
+# $FreeBSD$
+
+PORTNAME=	stringtie
+DISTVERSION=	1.3.4d
+CATEGORIES=	biology
+MASTER_SITES=	http://ccb.jhu.edu/software/stringtie/dl/
+
+MAINTAINER=	mzaki@niid.go.jp
+COMMENT=	Transcript assembly and quantification for RNA-seq
+
+LICENSE=	ART20 MIT
+LICENSE_COMB=	multi
+LICENSE_FILE_ART20=	${WRKSRC}/LICENSE
+LICENSE_FILE_MIT=	${WRKSRC}/samtools-0.1.18/COPYING
+
+USES=		gmake
+ALL_TARGET=	release
+
+PLIST_FILES=	bin/stringtie
+
+do-install:
+	${INSTALL_PROGRAM} ${WRKSRC}/stringtie ${STAGEDIR}${PREFIX}/bin
+
+.include <bsd.port.mk>
diff --git a/biology/stringtie/distinfo b/biology/stringtie/distinfo
new file mode 100644
index 000000000000..020a9bf31fc4
--- /dev/null
+++ b/biology/stringtie/distinfo
@@ -0,0 +1,3 @@
+TIMESTAMP = 1530582079
+SHA256 (stringtie-1.3.4d.tar.gz) = b1962d0108146ce7fea39d069b5e5de918e0e21daef9e1425ec9b778094d6ae6
+SIZE (stringtie-1.3.4d.tar.gz) = 438025
diff --git a/biology/stringtie/files/patch-Makefile b/biology/stringtie/files/patch-Makefile
new file mode 100644
index 000000000000..52df9a78c090
--- /dev/null
+++ b/biology/stringtie/files/patch-Makefile
@@ -0,0 +1,68 @@
+--- Makefile.orig	2018-03-06 16:29:49 UTC
++++ Makefile
+@@ -6,12 +6,12 @@ GDIR      :=./gclib
+ 
+ INCDIRS   := -I. -I${GDIR} -I${BAM}
+ 
+-CC        := g++
++CXX       ?= g++
+ 
+ BASEFLAGS := -Wall -Wextra ${INCDIRS} -fsigned-char -D_FILE_OFFSET_BITS=64 \
+ -D_LARGEFILE_SOURCE -fno-strict-aliasing -fno-exceptions -fno-rtti
+ 
+-LINKER    := g++
++LINKER    = ${CXX}
+ 
+ LDFLAGS   := -g -L${BAM} $(LDFLAGS)
+ 
+@@ -22,13 +22,13 @@ ifneq (,$(findstring nothreads,$(MAKECMD
+ endif
+ 
+ #detect MinGW (Windows environment)
+-ifneq (,$(findstring mingw,$(shell ${CC} -dumpmachine)))
++ifneq (,$(findstring mingw,$(shell ${CXX} -dumpmachine)))
+  WINDOWS=1
+ endif
+ 
+ # MinGW32 GCC 4.5 link problem fix
+ #ifdef WINDOWS
+-ifneq (,$(findstring 4.5.,$(shell g++ -dumpversion)))
++ifneq (,$(findstring 4.5.,$(shell ${CXX} -dumpversion)))
+  STATIC_CLIB=1
+ endif
+ #endif
+@@ -63,7 +63,7 @@ endif
+ ifneq (,$(filter %release %static, $(MAKECMDGOALS)))
+   # -- release build
+   RELEASE_BUILD=1
+-  CFLAGS := -DNDEBUG -g $(BASEFLAGS) $(CFLAGS) -O3
++  CFLAGS := -DNDEBUG -g $(BASEFLAGS) $(CFLAGS)
+ else
+   ifneq (,$(filter %memcheck %memdebug, $(MAKECMDGOALS)))
+      #use sanitizer in gcc 4.9+
+@@ -122,7 +122,7 @@ endif
+ 
+ 
+ %.o : %.cpp
+-	${CC} ${CFLAGS} -c $< -o $@
++	${CXX} ${CFLAGS} -c $< -o $@
+ 
+ OBJS += rlink.o tablemaker.o tmerge.o
+ 
+@@ -137,7 +137,7 @@ rlink.o : rlink.h tablemaker.h $(GDIR)/G
+ tmerge.o : rlink.h tmerge.h
+ tablemaker.o : tablemaker.h rlink.h
+ ${BAM}/libbam.a: 
+-	cd ${BAM} && make lib
++	cd ${BAM} && ${MAKE} lib
+ stringtie: ${BAM}/libbam.a $(OBJS) stringtie.o
+ 	${LINKER} ${LDFLAGS} -o $@ ${filter-out %.a %.so, $^} ${LIBS}
+ 	@echo
+@@ -151,6 +151,6 @@ clean:
+ 	${RM} stringtie${EXE} stringtie.o*  $(OBJS)
+ 	${RM} core.*
+ allclean cleanAll cleanall:
+-	cd ${BAM} && make clean
++	cd ${BAM} && ${MAKE} clean
+ 	${RM} stringtie${EXE} stringtie.o* $(OBJS)
+ 	${RM} core.*
diff --git a/biology/stringtie/files/patch-gclib_GIntHash.hh b/biology/stringtie/files/patch-gclib_GIntHash.hh
new file mode 100644
index 000000000000..a29977a4db84
--- /dev/null
+++ b/biology/stringtie/files/patch-gclib_GIntHash.hh
@@ -0,0 +1,20 @@
+--- gclib/GIntHash.hh.orig	2018-03-06 16:29:49 UTC
++++ gclib/GIntHash.hh
+@@ -122,7 +122,7 @@ public:
+   uint32 NextKey() {
+      Cell* cell=ghash.NextCell();
+      if (cell) return cell->key;
+-       else return NULL;
++       else return 0;
+   }
+   OBJ* NextValue() {
+      Cell* cell=ghash.NextCell();
+@@ -201,7 +201,7 @@ public:
+   uint32 NextKey() {
+      Cell* cell=ghash.NextCell();
+      if (cell) return cell->key;
+-       else return NULL;
++       else return 0;
+   }
+   OBJ* NextValue() {
+      Cell* cell=ghash.NextCell();
diff --git a/biology/stringtie/files/patch-samtools-0.1.18_Makefile b/biology/stringtie/files/patch-samtools-0.1.18_Makefile
new file mode 100644
index 000000000000..86d0fa80d3fc
--- /dev/null
+++ b/biology/stringtie/files/patch-samtools-0.1.18_Makefile
@@ -0,0 +1,10 @@
+--- samtools-0.1.18/Makefile.orig	2018-03-06 16:29:49 UTC
++++ samtools-0.1.18/Makefile
+@@ -1,5 +1,5 @@
+-CC=			gcc
+-CFLAGS=		-g -Wall -fsigned-char -O2 #-m64 #-arch ppc
++CC?=			gcc
++CFLAGS?=		-g -Wall -fsigned-char -O2 #-m64 #-arch ppc
+ DFLAGS=		-D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -D_CURSES_LIB=0 #-D_USE_KNETFILE 
+ KNETFILE_O=	knetfile.o
+ LOBJS=		bgzf.o kstring.o bam_aux.o bam.o bam_import.o sam.o bam_index.o	\
diff --git a/biology/stringtie/files/patch-samtools-0.1.18_bam__aux.c b/biology/stringtie/files/patch-samtools-0.1.18_bam__aux.c
new file mode 100644
index 000000000000..fb7dcf5c19f2
--- /dev/null
+++ b/biology/stringtie/files/patch-samtools-0.1.18_bam__aux.c
@@ -0,0 +1,20 @@
+--- samtools-0.1.18/bam_aux.c.orig	2018-03-06 16:29:49 UTC
++++ samtools-0.1.18/bam_aux.c
+@@ -25,7 +25,7 @@ uint8_t *bam_aux_get_core(bam1_t *b, con
+ 	return bam_aux_get(b, tag);
+ }
+ 
+-inline int aux_type2size(uint8_t type)
++int aux_type2size(uint8_t type)
+ {
+     switch (type) {
+     case 'A': case 'c': case 'C':
+@@ -43,7 +43,7 @@ inline int aux_type2size(uint8_t type)
+     }
+ }
+ 
+-inline uint8_t* skip_aux(uint8_t* s) {
++uint8_t* skip_aux(uint8_t* s) {
+ 	int size = aux_type2size(*s); ++s; // skip type
+ 	uint32_t n;
+ 	switch (size) {
diff --git a/biology/stringtie/pkg-descr b/biology/stringtie/pkg-descr
new file mode 100644
index 000000000000..246cda936a4b
--- /dev/null
+++ b/biology/stringtie/pkg-descr
@@ -0,0 +1,11 @@
+StringTie is a fast and highly efficient assembler of RNA-seq alignments into
+potential transcripts. The input can be the alignments of raw reads as usual;
+it also accepts longer assembled sequences. The output can be processed by
+downstream software including Ballgown, Cuffdiff, etc.
+
+Citation:
+Pertea et al. (2015) StringTie enables improved reconstruction of a
+transcriptome from RNA-seq reads.
+Nature Biotechnology 33: 290-295 doi:10.1038/nbt.3122
+
+WWW: http://www.ccb.jhu.edu/software/stringtie/