Combining multiple Trinity assemblies?

[yeaman]

Hello all,

I'm working on a project where we will have transcriptome sequences to about 35x for ~45 individuals from a single population, grown up in 7 different treatments. We are interested in assembling a reference transcriptome and then assessing patterns of differential expression among all of the individuals by aligning to this reference.

We are planning to use Trinity to assemble the transcriptomes, but we were uncertain whether it would be best to try to 'merge' them together afterwards (using Minimus or CAP3 or similar) or to simply paste together the .fasta files into one large file before running them through Trinity and then sort it out afterwards. The individuals are not clones (lodgepole pine), so there will be allele variants to contend with in addition to splicing and errors.

Any thoughts?

Thanks!

[Emilie]

Hi,

How about using a similar approach to Trans-ABySS 1st step, which merges together the contigs (.fasta files) obtained with ABySS using different k-values?

http://www.bcgsc.ca/platform/bioinfo/software/trans-abyss

The merging process is described in the manuscript:

http://www.ncbi.nlm.nih.gov/pubmed/20935650

Emilie

[Wallysb01]

I think this is what you're looking for. Like with the merging steps in trans-ABySS, it will only merge contains that are buried in another contig. So contags with SNPs won't collapse, which you'll probably have, since you have different individuals. I've messed around with other things that aren't as strict (mostly programs designed to assemble sanger sequences), but I'm not so sure I trust them all that much for transcriptomics with alternate splices and that. It seems to me this just isn't something that can be done all that cleanly right now.

http://groups.google.com/group/trans-abyss/browse_thread/thread/56f008afe45d93e4#