New to omics! Work with non-model lower eukaryotes... help :)

[EminentJunkie]

Hello SEQanswers!

I've been loving the wikis here for a while but didn't think to become a posting member until now. I'm learning (through almost entirely self-direction) how to do all of the bioinformatics required to assemble genomes and transcriptomes de novo (taxa is isolated such that there are no strong references available).

Can anyone please point me in the right direction? What assemblers should I be looking into if I'll be using Illumina HiSeq reads for a high-repeat genome circa 150Mb?

Thanks SEQanswers community. You guys are great.

EJ

[westerman]

Try many assemblers. :-)

Honestly none of the assemblers are right for every job. Personally I use ABySS because in general it works fine but there are some projects where I have problems with it. I suspect for your project it will be ok. Ray, SOAPdenovo, Velvet ... the list goes on. For transcriptomes I use Trinity but have heard that Oases is equally accurate and much faster.

All that said, for 'hard' yet small genomes obtaining some PacBio reads should be helpful to get over the rough spots. Bioinformatics can't solve everything. :-(

[lac302]

I had good luck with SGA and SOAPdenovo assembling highly a highly repetitive fish genome (1.1Gb)

I don't know what your computing resources are but the thing that attracted me to SGA was the ability to assemble the genome including scaffolding on a ~$2000 linux machine with 64GB ram.

[juan.isaza]

Try different assemblers and different configuration for each one (e.g. k-mer). Compare different parameters N50, total contigs, total assembled bases, total assembled reads etc.
you could do a quick evaluation of each assembly by blastx using nr (non redundant proteins) from NCBI as database.