demo_metassembly

demo_metassembly is a de novo sequence metassembler intended for short paired-end reads and large genomes. you can install this tool simply by using ./install.sh (only provided now) it consist of abyss1 , abyss2 , SOAPdenovo2 , Metassembly and another related tool(nvBowtie,kind of machine learning -not packaged yet)

Contents

• Quick Start
  • Install demo_metassembly on Debian or Ubuntu : currently not provided
  • Install demo_metassembly on Mac OS X : currently not provided
• Dependencies
• Installation
• Usage

Quick Start

Install demo_metassembly on Debian or Ubuntu

Run the command

sudo apt-get install demo_metassembly  : currently not provided

or download and install the Debian package.

Dependencies

demo_metassembly requires the following libraries:

• Boost
• Open MPI
• sparsehash
• SOAPdenovo2
• Metassember
• MUMmer
• bowtie2
• samtool

Option

• sratoolkit
• Quast

Installation

This pipeline is executed in Linux system. so i recommend to use in linux and use AWS instance becuast it has many depedency between program, so it makes many error. you just copy install.sh and make file to run. it will automatically install pipeline needed. or see install.sh and follow the script.

Usage

If you have a data, skip this part. using sratoolkit, download genome data. install sratoolkit in ~/tool:

tar -xzf sratoolkit.current-centos_linux64.tar.gz
cd ~/tool/sratoolkit.2.8.2-1-centos_linux64/bin
./fastq-dump -I --split-files SRR800798		## it make paired end sequencing data

and it need to change format of data to use. you can change format using changePairEndFormat.py in Format folder.

• Abyss

If pipeline is installed well, you can use abyss 1 & abyss 2. basic commands to run abyss 1 is :

abyss-pe name=**** k=** in="/home/ubuntu/data/genome_1 /home/ubuntu/data/genome_2"

and if you want to run it parallel, add 'np' command 'np=16' :

abyss-pe name=**** k=** np=16 in="/home/ubuntu/data/genome_1 /home/ubuntu/data/genome_2"

It will be parallelized automatically using mpi and multithreading. also you can run abyss2 by adding commands "B=100M H=3 kc=3":

abyss-pe name=**** k=** np=16 in="/home/ubuntu/data/genome_1 /home/ubuntu/data/genome_2" B=100M H=3 kc=3 v=-v

then you can get result of scaffolds. check details of parameter in Abyss github

• SOAPdenovo2

If pipeline is installed well, you can use SOAPdenovo63mer & SOAPdenovo127mer. 127mer need more memory than 63mer because of detailed execution. before you run SOAPdenovo2, You make your configure file like this:

max_rd_len=110 
[LIB]
avg_ins=200
asm_flags=3
reverse_seq=0
rank=1
q1=/home/ubuntu/data/SRR800798_1.fastq
q2=/home/ubuntu/data/SRR800798_2.fastq

and then, run using SOAPdenovo***mer :

./SOAPdenovo***mer all -s configureName.config -o outputName 1>ass.log 2>ass.err

if you want to run it parllel, add command '-p' :

./SOAPdenovo***mer all -s configureName.config -p 16 -o outputName 1>ass.log 2>ass.err

then you can get result of scaffolds. check details of parameter in SOAPdenovo2 github

• Metassembler

if pipeline is installed well, you can use Metassembler. it can be executed by shell script file. See file 'executeMetaseembly.sh'. you just change some value related with your data repository and tool repository.
If you want to run it parallel on bottlenecks, open file and add this command "-p 16" in bowtie2 :

bowtie2 -x MergePipeline/A.CEstat/BWTaln/A -1 jump2k.1.fq -2 jump2k.2.fq -p 16 --minins 200 ~~.....

then you can get result of Metassembler. check details of code in Metasembler Sourceforge