DISCOVAR de novo

海宝建议用这个拼接软件

http://www.broadinstitute.org/software/discovar/blog/?page_id=98

DISCOVAR – variant caller 适合于call variant 和拼接小基因组

DISCOVAR de novo 适合拼接大基因组

下载：

ftp://ftp.broadinstitute.org/pub/crd/DiscovarDeNovo/latest_source_code/LATEST_VERSION.tar.gz

安装：

General Instructions for Building Our Software

System Requirements

Software released from the CRD group at the Broad Institute is built and tested on a modern version of Linux for the x86_64 architecture. Our software does not run on 32-bit machines: you must have a 64-bit Linux system. Our users have successfully built and executed our software using a variety of Linux distributions including Ubuntu, RedHat, and SUSE. We expect that any flavor of x86_64 Linux will work fine, as long as it provides the necessary software prerequisites, as described below.

Basic Compiler and Library Requirements

We rely on reasonably up-to-date versions of these software packages:

• GCC, with its associated g++ compiler for the C++ language, version 4.7 or later.  We're now using C++11 features, and require a modern GCC.
• For ARACHNE only, the Xerces-C library. The source can be downloaded from the Xerces-c download page.  Supply the argument --with-xerces=/path/to/xercesc/installation when you run./configure.

The Build Procedure

• Move the package that you downloaded from our FTP site to a location on you system where you'd like to build the software.
• Extract the contents by typing: tar xzf package.tgz
• You'll have a new subdirectory in the current directory named after the package and revision. cd package Some older releases spill all the source into the current directory, rather than creating a new subdirectory. If that's the case, ignore this step.
• Execute the configuration script by typing: ./configure This assumes that you can copy executables to /usr/local/bin.  If you cannot, you should instead execute: ./configure --prefix=/path/to/install/directory Note: Some older packages lack a configuration script. Consider returning to the FTP site for a more recent version, or just skip this step.
• Build the software by typing: make all
• Install the software by typing: make install
• The executables will be in /usr/local/bin or in /path/to/install/directory/bin.  Make sure this is on your path.

If this goes well, you're ready to go. Consult the manual for the package to learn how to set up your data and what programs to execute.

对于数据的要求：

Sequencing data requirements summary
● Illumina MiSeq or HiSeq 2500 genome sequencers
● PCR-free library preparation
● 250 base paired end reads (or longer)
● ~450 base pair fragment size
● ~60x coverage

Input files
DISCOVAR requires a BAM file containing the raw reads from the sequencer. For variant calling it also
requires a matching reference FASTA file.

call variant  命令：

DISCOVAR can currently generate variants for small regions, and not the entire genome at once. To
generate variants for a 100 kb region for example, use:
Discovar \
READS=reads.bam \
OUT_HEAD=assembly \
REGIONS=1:50000150000
\
REFERENCE=genome.fasta
The complete set of variant calls for this region is given in the text file:
assembly.final.variant

Input files

DISCOVAR requires a BAM file containing the raw reads from the sequencer. For variant calling it also
requires a matching reference FASTA file.
BAM files
The reads to assemble must be in a BAM file or files. The name of the BAM file is specified with the
required argument READS :
READS= filename
Multiple BAM files are specified using a comma separated list:
READS= filename1,filename2,...
Alternatively, the BAM files can be specified in a separate file contain a list of BAM filenames, one per
line:
READS= @listfilename

DISCOVAR calls SAMtools internally to extract reads from the BAM.

Reference file (optional)

This is only required if you are using DISCOVAR as a variant caller. The reference information is used
only for variant calling and not in the assembly process. Specifying a valid FASTA reference file is all
that is required to cause DISCOVAR to generate variants.
To specify a reference FASTA file use the optional argument REFERENCE :
REFERENCE= filename

It should be the same file that was used to generate the alignments in the input BAM file(s), or at least
should share the same coordinate system. The FASTA record names should match those in the BAM
file. Ns are allowed.In addition to the reference FASTA file, DISCOVAR also requires the associated index file ( .fai

DISCOVAR can currently de novo assemble small genomes (up to 50 Mb), with larger genome support

to come soon.

The syntax for DISCOVAR de novo assembly is:
Discovar READS= bamfilenames \
OUT_HEAD= outputfilename \
REGIONS=all

This will take as input all the reads in the BAM file reads.bam , generate an assembly, then write the
output to a set of files prefixed with assembly

安装：

安装命令（在安装包里执行）：

$ CC=/opt/centos/devtoolset-1.1/root/usr/bin/gcc CXX=/opt/centos/devtoolset-1.1/root/usr/bin/c++ ./configure

$ make -j 32

$ sudo make install

 

GCC 4.7安装在/opt/centos/devtoolset-1.1，安装办法参考：

http://superuser.com/questions/381160/how-to-install-gcc-4-7-x-4-8-x-on-centos

 

如果是双端测序, 运行命令：

DiscovarDeNovo

NUM_THREADS=30（线程数）

MAX_MEM_GB=300G（最大内存）

MEMORY_CHECK=False （加上这个没报错）

READS=Project_TongJi_DNAseq_THB/Sample_TongJi-DNA-1/TongJi-DNA-1_CTTGTA_L000_R1.fastq.gz,Project_TongJi_DNAseq_THB/Sample_TongJi-DNA-1/TongJi-DNA-1_CTTGTA_L000_R2.fastq.gz（要用都好隔开）

OUT_DIR=wortdic（结果所在目录）