Issues with bacterial genome assembly (mapping vs. de novo)

[SDH]

I am attempting to assemble a genome from an organism I am working on. Initial BLAST analysis of some assembled contigs has led me to believe I am dealing with previously described bacterium “X”. Read mapping in geneious supports this as the majority of my reads are used and I get good coverage (bar a few regions of no coverage).

My de novo assembly however is giving me mixed messages…Though the assembled contigs have a high ANI, >95% when compare to the reference genome, the contigs will not map to the reference. I have a handful of contigs between 300 – 600 kb however, none will map back to the reference. I initially thought this was due to a poor assembly however my read quality is good and I can more-or-less (give or take a few base pairs) replicate the large contigs across two different assemblers (Spades and Abyss) and two different data sets from illumina and ion torrent reads generated from the same organism. DNA was extracted from a known pure culture so it is unlikely to be due to contaminating reads.

Has anyone come across this before or have any idea of where I might have gone wrong?

[Bukowski]

What are you mapping the contigs back to the reference with?

You could perhaps try reordering the contigs with progressive Mauve against the 'reference' genome, which might not be that good a reference though..

[SDH]

Good point - I was lazy and just ran the mapping within geneious though I did adjust the thresholds.

I will try bbmap - any other suggestions for bacterial genomes?

The data was trimmed/filtered prior to assembly and the fact I get very similar large contigs from two different assemblers I would like to think they are reasonably accurate.