biology/freebayes: Bayesian haplotype-based polymorphism discovery and genotyping

Approved by: jrm (mentor) Differential Revision: https://reviews.freebsd.org/D15116
svn path=/head/; revision=467806
2018-04-20 00:07:27 +00:00 · 2018-04-20 00:07:27 +00:00 · 73a92ce760 · 2021-03-31 03:12:20 +00:00
commit 73a92ce760
parent 42b89ed8c9
14 changed files with 293 additions and 0 deletions
--- a/biology/Makefile
+++ b/biology/Makefile
@ -33,6 +33,7 @@
    SUBDIR += fasttree
    SUBDIR += fastx-toolkit
    SUBDIR += fluctuate
+    SUBDIR += freebayes
    SUBDIR += garlic
    SUBDIR += gff2ps
    SUBDIR += gmap
--- a/biology/freebayes/Makefile
+++ b/biology/freebayes/Makefile
@ -0,0 +1,50 @@
+# $FreeBSD$
+
+PORTNAME=		freebayes
+DISTVERSIONPREFIX=	v
+DISTVERSION=		1.2.0
+CATEGORIES=		biology
+
+MAINTAINER=	jwb@FreeBSD.org
+COMMENT=	Bayesian haplotype-based polymorphism discovery and genotyping
+
+LICENSE=	MIT
+
+# Submodules use older versions than the ports, so we cannot safely unbundle
+# No shared libs are built by the submodules so there will be no conflicts
+USES=		gmake tar:xz
+USE_GITHUB=	yes
+GH_ACCOUNT=	ekg
+GH_TUPLE=	walaj:SeqLib:5941c68:SeqLib/SeqLib-temp \
+		jwalabroad:bwa:c02766e:bwa/SeqLib/bwa \
+		jwalabroad:fermi-lite:5bc90f8:fermi/SeqLib/fermi-lite \
+		samtools:htslib:49fdfbd:htslib/SeqLib/htslib \
+		ekg:bamtools:e77a43f:bamtools/bamtools \
+		ekg:intervaltree:dbb4c51:interval1/intervaltree \
+		vcflib:vcflib:5e3ce04:vcflib/vcflib-temp \
+		ekg:fastahack:c68cebb:fastahack/vcflib/fastahack \
+		ekg:filevercmp:1a9b779:filevercmp/vcflib/filevercmp \
+		ekg:fsom:a6ef318:fsom/vcflib/fsom \
+		google:googletest:d225acc:google/vcflib/googletest \
+		ekg:intervaltree:b704f19:interval2/vcflib/intervaltree \
+		ekg:multichoose:73d35da:multichoose/vcflib/multichoose \
+		ekg:smithwaterman:84c08d7:smithwaterman/vcflib/smithwaterman \
+		ekg:tabixpp:80012f8:tabixpp/vcflib/tabixpp \
+		samtools:htslib:0f298ce:hts2/vcflib/tabixpp/htslib
+
+# GCC disables sse2 by default on i386, but it's required for vcflib
+CXXFLAGS+=		--std=c++11
+CFLAGS_i386=		-msse2
+MAKE_JOBS_UNSAFE=	yes
+PLIST_FILES=		bin/bamleftalign bin/freebayes
+
+# GH_TUPLEs cannot seem to share a top-level subdirectory, so we work around
+# by dumping the parent in a temp directory and then moving the contents
+post-extract:
+	${MV} -n ${WRKSRC}/SeqLib-temp/* ${WRKSRC}/SeqLib
+	${MV} -n ${WRKSRC}/vcflib-temp/* ${WRKSRC}/vcflib
+
+do-install:
+	${INSTALL_PROGRAM} ${WRKSRC}/bin/* ${STAGEDIR}${PREFIX}/bin
+
+.include <bsd.port.mk>
--- a/biology/freebayes/distinfo
+++ b/biology/freebayes/distinfo
@ -0,0 +1,35 @@
+TIMESTAMP = 1523932688
+SHA256 (ekg-freebayes-v1.2.0_GH0.tar.gz) = a5626636112966cac87095365986722d1c871759f03c5a143646da65c416ba8a
+SIZE (ekg-freebayes-v1.2.0_GH0.tar.gz) = 2639725
+SHA256 (walaj-SeqLib-5941c68_GH0.tar.gz) = 0ac2b63f5fd7b1cec20a18ebb6799bf1f9b88d6c7cd95f7dd7410b1b82c8703f
+SIZE (walaj-SeqLib-5941c68_GH0.tar.gz) = 314968
+SHA256 (jwalabroad-bwa-c02766e_GH0.tar.gz) = cce547db0efef8572d63b47d2065f07e36f770fc8a6f15f5cd03d6fa9ab9b1f7
+SIZE (jwalabroad-bwa-c02766e_GH0.tar.gz) = 205546
+SHA256 (jwalabroad-fermi-lite-5bc90f8_GH0.tar.gz) = 2b3b9b637f48dc3239618efd6b55183eef6b6232013c217eb1c319f1dd3da224
+SIZE (jwalabroad-fermi-lite-5bc90f8_GH0.tar.gz) = 248454
+SHA256 (samtools-htslib-49fdfbd_GH0.tar.gz) = 0bc9c7d48e2a439badf5990113ae3e4de9acd05e4c090938236ee40588d115b9
+SIZE (samtools-htslib-49fdfbd_GH0.tar.gz) = 1145082
+SHA256 (ekg-bamtools-e77a43f_GH0.tar.gz) = d3a143636ccedcc54417a2d6d88dd726c7d2c046318620b88c7ba2fb30b01419
+SIZE (ekg-bamtools-e77a43f_GH0.tar.gz) = 540118
+SHA256 (ekg-intervaltree-dbb4c51_GH0.tar.gz) = 1f3ba92a83cb782cf22dfb383b8c93c11c1c88e0b2e5fa434e24668260ecd1f8
+SIZE (ekg-intervaltree-dbb4c51_GH0.tar.gz) = 3780
+SHA256 (vcflib-vcflib-5e3ce04_GH0.tar.gz) = d084c1df7bb0004df18e013da3cbf896b642d853c855aea58ae029abe1d2097a
+SIZE (vcflib-vcflib-5e3ce04_GH0.tar.gz) = 20119782
+SHA256 (ekg-fastahack-c68cebb_GH0.tar.gz) = c6f7186bbe733ff78c25b2afd7e7ba79daa3ecd14a5876ea0b033a43744c876d
+SIZE (ekg-fastahack-c68cebb_GH0.tar.gz) = 17312
+SHA256 (ekg-filevercmp-1a9b779_GH0.tar.gz) = 34c21d1b58295bca72b92c1cd832a4171b07ec97c19ffc0d4d8ec67e266c6070
+SIZE (ekg-filevercmp-1a9b779_GH0.tar.gz) = 3053
+SHA256 (ekg-fsom-a6ef318_GH0.tar.gz) = 68351c16d5c4b96142409e38c6828b6199eae44eaa7579bde165734d4832565b
+SIZE (ekg-fsom-a6ef318_GH0.tar.gz) = 9686
+SHA256 (google-googletest-d225acc_GH0.tar.gz) = 6880e0d51fa80a508ad5674ce483c1c468315faddac10d580b5505e8e7ef6bfa
+SIZE (google-googletest-d225acc_GH0.tar.gz) = 1279815
+SHA256 (ekg-intervaltree-b704f19_GH0.tar.gz) = 7e38b5f5873790afbcfdf6adbb621cdf4a1d9f22e28934a8edfe62cd65918f0d
+SIZE (ekg-intervaltree-b704f19_GH0.tar.gz) = 3768
+SHA256 (ekg-multichoose-73d35da_GH0.tar.gz) = af52475d7568863ad8173aebfb6b8b87690118bad8945876b629a2b3c05577cf
+SIZE (ekg-multichoose-73d35da_GH0.tar.gz) = 4697
+SHA256 (ekg-smithwaterman-84c08d7_GH0.tar.gz) = 299c9e28e50977642d1538f8940852976b6db3241619eb1f40b78484b1cde58e
+SIZE (ekg-smithwaterman-84c08d7_GH0.tar.gz) = 38382
+SHA256 (ekg-tabixpp-80012f8_GH0.tar.gz) = 6aecc865673776bec4e2ddb14e629a4798ad36bf8f7639968812a6262675fb3f
+SIZE (ekg-tabixpp-80012f8_GH0.tar.gz) = 3530
+SHA256 (samtools-htslib-0f298ce_GH0.tar.gz) = 829bb54383745fb923ed5d8dabbb6aa8b5d9501601c7a4163276d86aca1196ee
+SIZE (samtools-htslib-0f298ce_GH0.tar.gz) = 1065854
--- a/biology/freebayes/files/patch-Makefile
+++ b/biology/freebayes/files/patch-Makefile
@ -0,0 +1,20 @@
+--- Makefile.orig	2018-03-23 14:01:01 UTC
+++ Makefile
+@@ -5,7 +5,7 @@ wbamtools: vcflib/Makefile log
+ 	cd src && $(MAKE) -f Makefile.bamtools
+ 
+ log: src/version_git.h
+-	wget -q http://hypervolu.me/freebayes/build/$(shell cat src/version_git.h | grep v | cut -f 3 -d\  | sed s/\"//g) &
+	# wget -q http://hypervolu.me/freebayes/build/$(shell cat src/version_git.h | grep v | cut -f 3 -d\  | sed s/\"//g) &
+ 
+ src/version_git.h:
+ 	cd src && $(MAKE) autoversion
+@@ -27,7 +27,7 @@ uninstall:
+ 	rm /usr/local/bin/freebayes /usr/local/bin/bamleftalign
+ 
+ test:
+-	cd test && make test
+	cd test && ${MAKE} test
+ 
+ clean:
+ 	cd src && $(MAKE) clean
--- a/biology/freebayes/files/patch-SeqLib_bwa_Makefile
+++ b/biology/freebayes/files/patch-SeqLib_bwa_Makefile
@ -0,0 +1,14 @@
+--- SeqLib/bwa/Makefile.orig	2018-03-23 18:47:04 UTC
+++ SeqLib/bwa/Makefile
+@@ -1,8 +1,8 @@
+-CC=			gcc
+CC ?=			gcc
+ #CC=			clang --analyze
+-CFLAGS=		-g -Wall -Wno-unused-function -O2
+CFLAGS ?=		-g -Wall -Wno-unused-function -O2
+ WRAP_MALLOC=-DUSE_MALLOC_WRAPPERS
+-AR=			ar
+AR ?=			ar
+ DFLAGS=		-DHAVE_PTHREAD $(WRAP_MALLOC)
+ ## moved is.o bwtiindex.o rope.o to LOBJS
+ LOBJS=		utils.o kthread.o ksw.o bwt.o bntseq.o bwa.o bwamem.o bwamem_pair.o \
--- a/biology/freebayes/files/patch-SeqLib_bwa_bwt__lite.c
+++ b/biology/freebayes/files/patch-SeqLib_bwa_bwt__lite.c
@ -0,0 +1,10 @@
+--- SeqLib/bwa/bwt_lite.c.orig	2018-03-23 18:49:48 UTC
+++ SeqLib/bwa/bwt_lite.c
+@@ -1,6 +1,7 @@
+ #include <stdlib.h>
+ #include <string.h>
+ #include <stdio.h>
+#include <sys/types.h>
+ #include "bwt_lite.h"
+ 
+ #ifdef USE_MALLOC_WRAPPERS
--- a/biology/freebayes/files/patch-SeqLib_bwa_bwtgap.h
+++ b/biology/freebayes/files/patch-SeqLib_bwa_bwtgap.h
@ -0,0 +1,10 @@
+--- SeqLib/bwa/bwtgap.h.orig	2018-03-23 18:49:18 UTC
+++ SeqLib/bwa/bwtgap.h
+@@ -3,6 +3,7 @@
+ 
+ #include "bwt.h"
+ #include "bwtaln.h"
+#include <sys/types.h>
+ 
+ typedef struct { // recursion stack
+ 	u_int32_t info; // score<<21 | i
--- a/biology/freebayes/files/patch-SeqLib_bwa_kthread.c
+++ b/biology/freebayes/files/patch-SeqLib_bwa_kthread.c
@ -0,0 +1,10 @@
+--- SeqLib/bwa/kthread.c.orig	2018-03-23 18:48:35 UTC
+++ SeqLib/bwa/kthread.c
+@@ -1,6 +1,7 @@
+ #include <pthread.h>
+ #include <stdlib.h>
+ #include <limits.h>
+#include <inttypes.h>
+ 
+ /************
+  * kt_for() *
--- a/biology/freebayes/files/patch-SeqLib_fermi-lite_Makefile
+++ b/biology/freebayes/files/patch-SeqLib_fermi-lite_Makefile
@ -0,0 +1,12 @@
+--- SeqLib/fermi-lite/Makefile.orig	2018-03-23 18:50:25 UTC
+++ SeqLib/fermi-lite/Makefile
+@@ -1,6 +1,6 @@
+-CC=			gcc
+-CFLAGS=		-g -Wall -O2 -Wno-unused-function #-fno-inline-functions -fno-inline-functions-called-once
+-CPPFLAGS=
+CC ?=			gcc
+CFLAGS ?=		-g -Wall -O2 -Wno-unused-function #-fno-inline-functions -fno-inline-functions-called-once
+CPPFLAGS ?=
+ INCLUDES=	
+ OBJS=		kthread.o misc.o \
+ 			bseq.o htab.o bfc.o \
--- a/biology/freebayes/files/patch-SeqLib_htslib_Makefile
+++ b/biology/freebayes/files/patch-SeqLib_htslib_Makefile
@ -0,0 +1,38 @@
+--- SeqLib/htslib/Makefile.orig	2017-06-20 12:40:28 UTC
+++ SeqLib/htslib/Makefile
+@@ -22,20 +22,20 @@
+ # FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ # DEALINGS IN THE SOFTWARE.
+ 
+-CC     = gcc
+-AR     = ar
+-RANLIB = ranlib
+CC     ?= gcc
+AR     ?= ar
+RANLIB ?= ranlib
+ 
+ # Default libraries to link if configure is not used
+ htslib_default_libs = -lz -lm -lbz2 -llzma
+ 
+-CPPFLAGS =
+CPPFLAGS ?=
+ # TODO: probably update cram code to make it compile cleanly with -Wc++-compat
+ # For testing strict C99 support add -std=c99 -D_XOPEN_SOURCE=600
+ #CFLAGS   = -g -Wall -O2 -pedantic -std=c99 -D_XOPEN_SOURCE=600 -D__FUNCTION__=__func__
+-CFLAGS   = -g -Wall -O2
+CFLAGS   ?= -g -Wall -O2
+ EXTRA_CFLAGS_PIC = -fpic
+-LDFLAGS  =
+LDFLAGS  ?=
+ LIBS     = $(htslib_default_libs)
+ 
+ prefix      = /usr/local
+@@ -86,7 +86,7 @@ BUILT_THRASH_PROGRAMS = \
+ 	test/thrash_threads5 \
+ 	test/thrash_threads6
+ 
+-all: lib-static lib-shared $(BUILT_PROGRAMS) plugins $(BUILT_TEST_PROGRAMS)
+all: lib-static $(BUILT_PROGRAMS) plugins $(BUILT_TEST_PROGRAMS)
+ 
+ HTSPREFIX =
+ include htslib_vars.mk
--- a/biology/freebayes/files/patch-src_LargeFileSupport.h
+++ b/biology/freebayes/files/patch-src_LargeFileSupport.h
@ -0,0 +1,11 @@
+--- src/LargeFileSupport.h.orig	2018-03-23 21:15:04 UTC
+++ src/LargeFileSupport.h
+@@ -8,7 +8,7 @@ typedef __int64 off_type;
+ #define ftell64(a)     ftell(a)
+ #define fseek64(a,b,c) fseek(a,b,c)
+ typedef __int64_t off_type;
+-#elif defined(__APPLE__)
+#elif defined(__APPLE__) || defined(__FreeBSD__)
+ #define ftell64(a)     ftello(a)
+ #define fseek64(a,b,c) fseeko(a,b,c)
+ typedef off_t off_type;
--- a/biology/freebayes/files/patch-src_Makefile
+++ b/biology/freebayes/files/patch-src_Makefile
@ -0,0 +1,62 @@
+--- src/Makefile.orig	2018-03-23 14:01:01 UTC
+++ src/Makefile
+@@ -5,8 +5,9 @@
+ ################################################################################
+ 
+ # Compiler
+-CXX=g++ ${CXXFLAGS}
+-C=gcc
+CXX ?= g++
+# CXX += ${CXXFLAGS}
+CC ?= gcc
+ 
+ export CXXFLAGS
+ export C
+@@ -17,7 +18,7 @@ export LIBFLAGS
+ 
+ # Compiler flags
+ 
+-CFLAGS=-O3 -D_FILE_OFFSET_BITS=64 -g
+CFLAGS += -D_FILE_OFFSET_BITS=64
+ #CFLAGS=-O3 -static -D VERBOSE_DEBUG  # enables verbose debugging via --debug2
+ 
+ SEQLIB_ROOT=../SeqLib
+@@ -25,7 +26,7 @@ VCFLIB_ROOT=../vcflib
+ TABIX_ROOT=$(VCFLIB_ROOT)/tabixpp
+ HTSLIB_ROOT=$(SEQLIB_ROOT)/htslib
+ 
+-LIBS = -lz -llzma -lbz2 -lm -lpthread
+LIBS = -lz -llzma -lbz2 -lm -lpthread -lexecinfo
+ INCLUDE = -I../src -I../ttmath -I$(VCFLIB_ROOT)/src/ -I$(VCFLIB_ROOT)/smithwaterman/ -I$(VCFLIB_ROOT)/multichoose/ -I$(VCFLIB_ROOT)/filevercmp/ -I$(VCFLIB_ROOT)/fastahack/ -I$(HTSLIB_ROOT) -I$(SEQLIB_ROOT) 
+ #INCLUDE = -I../ttmath -I$(BAMTOOLS_ROOT)/src/ -I$(VCFLIB_ROOT)/src/ -I$(TABIX_ROOT)/ -I$(VCFLIB_ROOT)/smithwaterman/ -I$(VCFLIB_ROOT)/multichoose/ -I$(VCFLIB_ROOT)/filevercmp/ -I$(VCFLIB_ROOT)/fastahack/ -I$(HTSLIB_ROOT) -I$(SEQLIB_ROOT) -I$(SEQLIB_ROOT)/htslib
+ 
+@@ -46,10 +47,10 @@ gprof:
+ .PHONY: all static debug profiling gprof
+ 
+ $(HTSLIB_ROOT)/libhts.a:
+-	cd $(HTSLIB_ROOT) && make
+	cd $(HTSLIB_ROOT) && ${MAKE}
+ 
+ $(SEQLIB_ROOT)/src/libseqlib.a:
+-	cd $(SEQLIB_ROOT) && ./configure && make 
+	cd $(SEQLIB_ROOT) && ./configure && ${MAKE} 
+ 
+ OBJECTS=BedReader.o \
+ 		CNV.o \
+@@ -122,7 +123,7 @@ freebayes.o: freebayes.cpp TryCatch.h $(
+ 	$(CXX) $(CXXFLAGS) $(INCLUDE) -c freebayes.cpp
+ 
+ fastlz.o: fastlz.c fastlz.h
+-	$(C) $(CFLAGS) $(INCLUDE) -c fastlz.c	
+	$(CC) $(CFLAGS) $(INCLUDE) -c fastlz.c	
+ 
+ Parameters.o: Parameters.cpp Parameters.h Version.h
+ 	$(CXX) $(CXXFLAGS) $(INCLUDE) -c Parameters.cpp
+@@ -277,5 +278,5 @@ autoversion:
+ 
+ clean:
+ 	rm -rf *.o *.cgh *~ freebayes alleles ../bin/freebayes ../bin/alleles ../vcflib/*.o ../vcflib/tabixpp/*.{o,a} tabix.hpp
+-	if [ -d $(BAMTOOLS_ROOT)/build ]; then make -C $(BAMTOOLS_ROOT)/build clean; fi
+-	make -C $(VCFLIB_ROOT)/smithwaterman clean
+	if [ -d $(BAMTOOLS_ROOT)/build ]; then ${MAKE} -C $(BAMTOOLS_ROOT)/build clean; fi
+	${MAKE} -C $(VCFLIB_ROOT)/smithwaterman clean
--- a/biology/freebayes/files/patch-vcflib_Makefile
+++ b/biology/freebayes/files/patch-vcflib_Makefile
@ -0,0 +1,13 @@
+--- vcflib/Makefile.orig	2018-03-23 18:36:07 UTC
+++ vcflib/Makefile
+@@ -136,8 +136,8 @@ all: $(OBJECTS) $(BINS)
+ 
+ GIT_VERSION := $(shell git describe --abbrev=4 --dirty --always)
+ 
+-CXX = g++
+-CXXFLAGS = -O3 -D_FILE_OFFSET_BITS=64 -std=c++0x 
+CXX ?= g++
+CXXFLAGS += -D_FILE_OFFSET_BITS=64 -std=c++0x 
+ #CXXFLAGS = -O2
+ #CXXFLAGS = -pedantic -Wall -Wshadow -Wpointer-arith -Wcast-qual
+ 
--- a/biology/freebayes/pkg-descr
+++ b/biology/freebayes/pkg-descr
@ -0,0 +1,7 @@
+FreeBayes is a Bayesian genetic variant detector designed to find small
+polymorphisms, specifically SNPs (single-nucleotide polymorphisms), indels
+(insertions and deletions), MNPs (multi-nucleotide polymorphisms), and complex
+events (composite insertion and substitution events) smaller than the length of
+a short-read sequencing alignment.
+
+WWW: https://github.com/ekg/freebayes