Is It Reasonable To Combine Both Genome And Transcriptome Sequences To Get A Larger Assembly ??

Question

[Lhl]

Hi All,

We sequenced the subsection of transcriptome and subsection of genome of our model plant using 454 pyrosequencing technology.

We were able to assemble both transcriptome and genome data.

However, i want to know 'is it good to combine them all to get a big assembly ?'.

I asked this is because 1) transcriptome data and genome data are different. 2) the 454 assembler -- Newbler treats genome and transcriptome data differently.

Kind Regards,

lhl

Answer 1

[Qdjm]

This is probably not a good idea because your transcriptome reads can span splice junctions. I suppose that this depends on your assembler but I suspect most assemblers would be confused by this.

[Lhl]

Thanks qsjm.

I think you are right. That's also my concern.

[Lhl]

Do you think assembling genome and transcriptome separately, and then hybridizing them would make difference ?

[Qdjm]

I'm no expert -- if you don't get more response here, try looking at a few recent genome paper. Other people must have encountered this problem.

[Leszek]

+1; assembly using genomic reads only. then you can align RNA-Seq reads to estimate: how many transcripts you miss in your assembly or what fraction of your genes is fragmented in separate contigs and so on. or create transcripts de novo (oases) and then align on your genome assembly.

[Lhl]

Thanks Leszek. Will definitely stick to separate assembly.
If i am going to used oases, then i will try out multiple kmers, should i choose only the kmer that producing best assembly or should i merge multiple kmer assembly? Maybe i should open a new ticket. But hope you can give me a hint.