IDBA-UD assembler

[Mark]

Hi All

Has anyone given the idba-ud assembler a try? My first go at it produced a bacterial genome assembly quite inferior to what I had achieved with Velvetoptimiser, but proper parameterization of idba is not exactly clear to me. What has your experience been?

Mark

[lanzen]

Hi Mark,

I am about to try it now with some mixed (by mistake) metagenome and metatranscriptome data. I found the algorithm very appealing for this purpose, since the main problem will be extremely variable coverage. I will let you know how it turns out.

Did you try it for metagenome data? I don't think there is much point in using for a normal single genome assembly..

[luperci]

lanzen,

How did it go? I'm currently trying to use idba-ud to do a metagenomic assembly of a goat gut. Do you know if there is a way to limit the memory usage? We run all our analysis using a cluster, and my job got killed because it went over the requested memory.

[lanzen]

For me it worked quite fine, but most contigs were short, so it is unclear to me what was the (contaminating) metagenomic data that I tried to remove or what was from shorter transcripts. This may not be the fault of the assembly algorithm though. I will try a more normal dataset and let you know.

Otherwise, it did not use that much memory or calculation time. A 32Gb machine seemed with 8 cores was enough and the job finished in 30 min.

Let me know if there is something else I can do to help you.

[luperci]

I allocated 30 cores and 300 GB and it ran for around 7 hours before being stopped. I have a feeling the sample was really over sequenced. Right now I looking into doing some preprocessing to clean up the reads a little. Hopefully that will help limit the resource usage. Thanks.

[mstagliamonte]

Hi,


sorry to revive an old thread. I did two idba runs, one using only one fasta file, the other using two files (two separate illumina runs for the same sample).

In the first idba run, the file contig.fa had no read counts, while they were present in the scaffold.fa.
At the second attempt (using two fasta files), it was the opposite!

Any idea what's happening?
Max