Hi Logambiga,

That's actually not a trivial question! So far people said that a good assembly had 1/10000 bases wrong. It was possible be calculated those qualities back in the capillary data, looking into the alignment and checking the phred score of the quality of the bases.

This would just give you base correctness. There are now tools that correct (check) assemblies automatically using NGS, like icorn - http://bioinformatics.oxfordjournals.org/content/26/14/1704. Basically reads get mapped, SNP called and the bases corrected based on some test. This is than repeated repetitively.

But to look into regions that are wrongly joint, latest tools, like reapr (http://genomebiology.com/2013/14/5/R47) are based on mapping read pairs... and you might not have those...

I would be pragmatic to check for latter errors: If you have a closely related reference, do a genome alignment, check for breaks in synteny, and revisit the alignments at the breakpoints. You could also do some PCR over those regions to check the correctness of the assembly.

Hope this helps a bit,

Thomas

Hi Thomas Otto,

Thanks for your suggestions.