[Velvet-users] A strategy for scaling genome and transcriptome contig assembly (digital normalization)

---------- Forwarded message ----------
From: "C. Titus Brown"
Date: Mar 25, 2012 12:48 PM
Subject: [Velvet-users] A strategy for scaling genome and transcriptome contig assembly (digital normalization)
To: "velvet-users"
Hi all,

last week I posted a preprint of a paper discussing a strategy for coverage normalization, data reduction, and error elimination:

http://ivory.idyll.org/blog/mar-12/diginorm-paper-posted.html

we call this strategy 'digital normalization' and it can yield good to spectacular reductions in data size and memory usage for assembly.  In the paper we test it with Velvet, Oases, and Trinity on a variety of data sets.

---

On the paper site,

http://ged.msu.edu/papers/2012-diginorm/

I just posted a tutorial for running it on microbial genomes prior to Velvet assembly, and on the Trinity paper's yeast mRNAseq data set prior to Oases or Trinity assembly.

http://ged.msu.edu/angus/diginorm-2012/tutorial.html

The tutorial uses an Amazon EC2 instance for reproducibility, but with a bit of hopefully obvious tweaking the commands should work on any Linux system.  Note, you'll need about 15 gb of RAM for the yeast Oases & Trinity assemblies.

Let me know if you have any questions (but be please to ask just on the relevant mailing list -- I'm sending this to velvet, oases, and trinity lists).

cheers,
--titus

_______________________________________________
Velvet-users mailing list
http://listserver.ebi.ac.uk/mailman/listinfo/velvet-users