I have assembled four bacterial genomes derived from Miseq 2x300 pairedend data using SPADES and A5 pipeline. I also utilized CLC Workbench and CISA (to join assemblies), but they performed poorly.

After checking the assemblies with reapr and checkm, I got contradictory statistics.

Checkm demonstrated that the completeness of all genomes are over 99%.

However, the results of reapr are frustrating. One of the genomes has a proportion of error free bases of 22%. In fact, looking closely, all genome assemblies are full of N's after the reapr processing, even though some of the genomes present a proportion of error free bases over 80%... All genomes present FCD errors (one with more than 60!)

I tried to use Gapfiller, but it did not filled any gap of the reapr processed contigs...

I noted that all assemblies do not have high coverage, most are around 15X.

My main objective is to evaluate the presence of some genes of interest (less than a hundred) and to compare the genomes to other related ones. Therefore, I do not need a finished genome.

What do you recommend that I should do before submitting the genomes to Genbank? Should I sequence more? Isn't REAPR too much stringent? Is there any other software that could improve my draft genome using the actual sequencing data?

Thank you