De novo or mapped (reference guide) assembly

[fegunsu]

Hi everyone.
I want to ask about de novo assembly.
If I mapped whole genome reads from Illumina to reference genome for extracting chloropast genome then using ABySS or Velvet for assembly. Can you say that my job is de novo assembly?
If yes, it is still de novo, so how is mapped assembly done?

Thanks

[westerman]

Extracting reads using a reference as a guideline and then using the extracted reads for a assembly would indeed be a de-novo assembly. You are merely reducing the number of reads used in assembly and not fundamentally changing assembly. In other words if you sequence chloroplast only instead of the whole genome then you should arrive at approximately the same number of reads to start the assembly from.

Assuming that you use fairly loose parameters for the extraction then approaching a project in this manner might pick up indels and other genome rearrangements that a stricter reference-mapping assembly would miss. On the other hand de-novo, due to a lack of coverage or uneven coverage, could miss portions of the genome.

Notes:
1) mira-bait is a good program to extract reads

2) Some people will do a "reference-guided" assembly which tries to combine the best of both approaches.

[fegunsu]

Thank you for the answer. I am using BWA to extract and using Velvet to assemble it.