Recommendations for RNA-seq mapping for ABI/Illumina data-sets

[thedavid]

I'm trying to compare transcriptomes of developing zebrafish and have tracked down RNA-seq data-sets for several of the time-points I'm interested in.

The problem is that there are three different platforms - ABI Solid 3.0, Illumina Genome Analyzer II, and Illumina Hi-Seq. The people I work with are intimately familiar with handling Illumina data, but not with Solid data-sets.

I've downloaded the SRA files for the data sets and have converted them to FASTQ (using abi-dump for the abi data, and fastq-dump for the others).

So, what's the "standard" aligner for Solid 3.0 RNA-seq data? I know Tophat v1 can do it....is that the "best?"

Thanks,
D

[mbblack]

Optimally, one would map SOLiD reads in color space, and for that, in my experience, ABI's own software (older=BioScope, current=LifeScope) does perhaps the best job of any mapping tool. LifeScope is free with version 2.5.1 (only requires registration to download). SHRiMP would be an alternative aligner and handles color space mapping.

I've never mapped my SOLiD reads in anything but color space, so cannot speak to aligning them in base space, but would think Tophat/Bowtie would be a logical choice.

[thedavid]

Thank mbblack. It looks like it's best to use ABI's mapper.

So, what's a recommended tool to convert from colorspace to basespace after mapping? I can't figure out if ABI provides one.

[mbblack]

Thank mbblack. It looks like it's best to use ABI's mapper.

So, what's a recommended tool to convert from colorspace to basespace after mapping? I can't figure out if ABI provides one.

LifeScope takes color space reads, converts your reference genome to color space, and then maps reads in color space. But your output is in base space with color space based mapping quality scores - just regular old BAM files as output (and wig files to view coverage). So there is nothing to convert in terms of output.

[thedavid]

Well. LifeScope is a real PITA to install. We have a nice SGE-based cluster here and support staff to install software for us. Unfortunately LifeScope appears to need it's own computer to run a server off of. Which seems sort of insane. Is this true?

Anyways, looks like I'm stuck using Tophat v1.

[mbblack]

Well. LifeScope is a real PITA to install. We have a nice SGE-based cluster here and support staff to install software for us. Unfortunately LifeScope appears to need it's own computer to run a server off of. Which seems sort of insane. Is this true?

Anyways, looks like I'm stuck using Tophat v1.

All I've ever done is download the installer, run it, fill in the appropriate information and the script installs everything just fine. While I've used UNIX/Linux for years, I am no sysadmin by any stretch (my formal education was all marine biology and population genetics back in the 1990's), but it was a pretty seemless installation shell script as far as I have experienced. Our cluster is just an 8-core head node and 4x8-core compute nodes running CentOS 5.5 (we do not have a Linux sysadmin on staff, so we make do with the couple of us who have an Linux-IT background and we've managed to figure it out so trust me, it cannot be that hard for any actually qualified sysadmin to get working).

The most challenging thing for me was getting a new queue for PBS to initialize properly, but a few minutes reading online TORQUE & MAUI documentation solved that for me.

I do not remember all the system requirements off the top of my head, but I know it supports Redhat and CentOS up to v.5.7, PBS/Torque 2.3 and up, and SGE although I do not remember the minimum version number. Also needs python 2.3 or up, and PERL 5.8.something or up.

The lifescope gui just runs as a simple Apache server on localhost. Your sysadmin will have to set up the appropriate queue for lifescope's jobs, and the shared directories for user projects, reference genomes and such.