Circular dna overlap region

[hyeong8984]

Hi,
I tried to assemble circular DNA with hifi reads, and made it.

I found overlapping region of contig with width about 22 kbp using nucmer. (total contig length = 878,387bp)

My question is that, why this overlapping region is made in specific region?
How canu define the seed reads to start assembly and end..?
I have been studying canu but cannot find it well.

Thank you for your time, and I look forward to your response.

[skoren]

Circular molecules typically generate self-similar sequences at the start and end, see https://canu.readthedocs.io/en/latest/faq.html#my-circular-element-is-duplicated-has-overlap. This is because the start of the molecule is arbitrarily defined (there's no start/end on a circle) and reads are not split so the end read can repeat the start sequence again. Typically, canu will mark these as circular in the header and provide trim coordinates to remove the redundancy.