I have paired reads from sequencing bacterial genome with Illumina MiSeq. I assembled them into contigs and now I have 4022 contigs with N50=72749, length of the longest contig 229883, number of mapped contigs 41 and reference genome (1.8 Mb chromosome + 3 plasmids) coverage 91.35%. And I don't know what to do with 3981 unmapped contigs. Do they contain any useful information or they can be removed from the further analysis? And what about plasmids? How to separate plasmid DNA from a bacterial chromosome? And how can I improve this assembly by sequencing by Sanger?
Thank in advance for an answer!
•
written
by
a.a.kechin •
Might try to check for contamination first
Perform a blastn of ncbi then use MEGAN to identify contigs that are contamination i.e. human or mouse etc. Then delete those contigs.
You say contigs, have they not been scaffolded too? What assembler did you use and parameters?
Try using a higher k-mer value in your assembly and although you are likely to get smaller contigs they may extend your current contigs and allow you to join some together by aligning them. Or if not tried use a multi-kmer assembly approach like with soapdenovo2 usually 61-85 can do a good job. If plasmids are AT rich then high kmer value may assemble these specifically better and may have the coverage to make this more successful versus the genome.
Finally:
Align your genome against another similar using lastz to allow you to scaffold the contigs you have so can do some sangar to fill the gaps. Some use Mauve to do this automatically but not found it that good.
•
written
by
rob234king
Thank you! I'll try it.
I used a5-miseq pipeline. It produced scaffolds, too. But they were the same as contigs. How to separate contamination from horizontal gene transfer?
•
written
by
a.a.kechin •
Log in to answer this question.
