long repeated elements in a bacterial genome

[RudyS]

if there are say 20 (almost) identical transposons of say 800 bases in a 6 Mbase genome, is there any way an assembler can properly position them using only single reads of 36 bases? and what to think when an assembler gives a bunch of contigs that all finish with a few hundred bases of that transposon?

RudyS

[swbarnes2]

if there are say 20 (almost) identical transposons of say 800 bases in a 6 Mbase genome, is there any way an assembler can properly position them using only single reads of 36 bases?

I don't think so.

and what to think when an assembler gives a bunch of contigs that all finish with a few hundred bases of that transposon?

Then you've got unique sequence flanked by repeats. You'll never know in what order they all fall without paired ends, or far longer reads.

[mchaisso]

I don't think so.

More directly, unequivocally no.

This will require mate pairs with a span greater than 800 bases to resolve.

-mark

[RudyS]

maybe its not possible? any other de novo assembler on the horizon?

so basically its hopeless you think ... because even with paired-end reads 800-1000 base repeated elements are a long stretch? it would seem that the assembler might actually connect inappropriate flanks for the repeats ... in my case it didnt ... but maybe it gets lost in the middle ... what assemblers are people using?