genome assembler on cDNA data

[geschickten]

Why can't we use with genome assemblers for assembly of cDNA?

[westerman]

You can use genome-base assemblers for cDNA assembly. However, rather obviously, cDNA is different looking than a genome (e.g., size; composition) and the depth-of-coverage of the sequencing is often different as well. Thus using an assembly program 'tuned' to detect cDNA is better than using a genomic assembly program.

The above is rather generic. Undoubtedly someone somewhere has written a in-depth article about the differences. However if you can answer "what is the difference between cDNA and genomic DNA" then you should have a rough idea of why assembly on one is not the same as the assembly of the other.

[jnfass]

Thanks for asking the question, geschickten.

I have indeed done this, but not for any published analysis. I don't know of any reason why most assemblers would not do just as well on transcriptome assembly as genome assembly. It seems that in some ways, the source sequence has reduced complexity (no non-coding, highly repetetive sequence), but may still have a lot of difficult complexity (similar domains, splice variants, gene families).

But really I'm just handwaving. ... I'd love to hear from someone who's done some real benchmarking on this issue.