Assembly of Large Genomes using Cloud Computing by Contrail

[Gangcai]

I have found a new software (the code source is not available yet) for large genomes assembly(http://sourceforge.net/projects/contrail-bio/). It relies on Hadoop to iteratively transform an on-disk representation of the assembly graph, allowing an in depth analysis even for large genomes, which can reduce the memory requirement. Contrails also use de Brujin graph strategy to do the short reads assembly.
for more see the wiki: http://sourceforge.net/apps/mediawik...title=Contrail

[Zigster]

does anyone know when the first release of Contrail is scheduled?

The quality of software coming from the Salzberg and Pop labs has been very high. Despite the inelegant name, I am really looking forward to seeing how Contrail compares with Velvet, Abyss, SOAPdenovo, etc..

[Zigster]

I am also wondering if this assembler is written in entirely in Java. Isn't that a Hadoop requirement?

[Gangcai]

I am also looking forward to the first release of Contrail....

[jjv5]

The source code has been released but it does not look like one could just run it. No documentation, no hints. Any word on when a usable version might be available?

[seb567]

The source code has been released but it does not look like one could just run it. No documentation, no hints. Any word on when a usable version might be available?

Meanwhile ->

Michael Schatz (Cold Spring Harbor Laboratory)
"Cloud Computing and the DNA Data Race: Theory and Practice."

http://schatzlab.cshl.edu/presentations/2010-10-22.CSHL.Sequence.Cloud.Computing.pdf

[FractalExpression]

has anyone gotten this up and running? does it read raw sequence files? so many questions!

[samanta]

Hi all,

I managed to get contrail up and running.

Here is how to run the program on the test case provided by Schatz.

http://www.homolog.us/blogs/2011/09/08/contrail-a-de-bruijn-genome-assembler-that-uses-hadoop/

[johawashin49]

Such an amazing post! This is an interesting thread. I wonder how effective this cloud computing is.






_____________________________________________________________________________________________________

"Defect-free software does not exist."
~ Wietse Venema ~
Hosting Dallas

[samanta]

Effectiveness of which one are you asking about - cloud computing in general or hadoop?

Cloud computing basically allows you to rent computer time from some company managing the hardware. To an user, it is nothing different from using a local supercomputing facility at the university or other place.

Hadoop is a different paradigm and has been useful for large data. You can even set it locally and do not need cloud computing for it. I have few posts on hadoop for bioinformatics.

http://www.homolog.us/blogs/category/hadoop/