Running 454 assembly on ec2 (using gsAssembler/newbler/runAssembly)

gistfile1.md

• make sure we map the ephemeral storage to the instance on the instance start. This uses a 64-bit Amazon Linux AMI, which is modeled on Fedora. So, the typical Fedora install commands work (e.g. yum install whatever):

    ec2-run-instances --key keypair ami-74f0061d \
    --instance-type=m2.2xlarge \
    --block-device-mapping '/dev/sda2=ephemeral0' \
    --block-device-mapping '/dev/sda3=ephemeral1'
  

• dont think we actually need them, but i like to have them around, in case we do - they also install glibc.x86_64

    yum groupinstall "Development Tools"
  

• install additional libraries needed by 454 (if you ran the above)

    yum install zlib.x86_64 libXi.x86_64 libXtst.x86_64 libXaw.x86_64 screen
  

• if you didn't install the "Development Tools"

    yum install glibc.x86_64 zlib.x86_64 libXi.x86_64 libXtst.x86_64 libXaw.x86_64 screen
  

• get 2.5.3 package, unzip, install.

    cd /usr/local/src
    wget http://some.fancy.url/DataAnalysis_2.5.3.tgz
    tar -xzvf DataAnalysis_2.5.3.tgz && cd Data*
  

• setup software by following prompts. Disregard i386 warnings and keep choosing "Next" - you've installed what you need during the steps above:

    ./setup.sh
  

• get sff/fasta files from somewhere to the instance ephemeral store

    mkdir -p /mnt/ephemeral0/data && cd /mnt/ephemeral0/data
    wget http://some.fancy.url/fasta.tar.gz
    cd ../
  

• start screen, so we can logout or in case we get disconnected the analysis will still run

    screen
  

• run the assembly (m2.2xlarge has 4 processing cores) - use default options (data are trimmed)

    /opt/454/bin/runAssembly -o my-organism -cpu 4 fasta/my-organisms.454reads.combined.fsa
  

• detach from the screen session by typing ctrl-a+d. reattach to the session using

    screen -r
  

• download data from instance (using SFTP or something)