I'm trying to assemble a genome of halobacteria from the DNA sample obtained from enrichment culture (culture with media, optimized for growth of 1 species, but it obviously contains a mix of different species). Amount of microbe of interest should be at least 80%.
Coverage was quite high - about 1000х (ILMN MiSeq 2x250 reads). I trimmed reads and ran with CLC assembler. I've got quite fine contigs, but:
- total length of assembly is about 12 Mb which is far too much
- read coverage of contigs is not clustered, but distributed continuously, so I cannot easily divide them into groups...
Does anybody have some tips for such assembly or maybe somebody can suggest an appropriate assembler for this?
Hi, I presume you have paired reads. The genome assembly quality is highly dependent on the size of the fragments between your paired reads. With 700 bp fragments I get nice genome drafts with less than 100 contigs. One way to reduce contig numbers is to apply a filter on the read cover and contig size. Usually I remove contigs
Dear Dr. Stepan,
I suppose your reads in the raw data are all about 250 bp, right?. Well, 250 bp is considered a long read. Therefore, the OLC (Overlap-Layout-Consensus) algorithm is the most apropriate for this case. We have some good OLC based assemblers that you could try use, for exemple, Newbler, Mira or Celera. Particulary, i like the Newbler because it has a friendly GUI interface.
I hope this helps.
Have a good nigth Dr. Stepan
Hi!
Based on my experience, I would try the Newbler assembler (v2.3 and later) after reformatting the FASTQ header so that Newbler identifies the sequences as paired reads.
But it might be a good idea to actually reduce the amount data to
- Badges
- Science method