Edge effect of circular mitochondrion seed sequences being represented linearly in seed fasta file?

[ckeeling]

The way reads are being recruited based upon similarity to the seed sequences, would one expect the ends of the seed sequences to recruit fewer reads than the middle portions? My linear fasta representations of the circular seed sequences all start/end within the same repetitive region, which is also the region my assembly is struggling with. Repetitive regions are always going to be harder to assemble, but if I cannot recruit enough reads that span this break in the circle/span the repetitive region, it will be even harder. I'm trying with editing the fasta so that they are breaking in a non-repetitive region to see if it makes a difference, but was wondering if others have had this problem before. Is the subsequent extension phase sufficient to capture the appropriate reads in this region?