Implications of high-coverage at contig ends

[rhinoceros]

So, I have this contig that I presume represents a complete phage genome. For some reason, especially the contig ends have very high coverage. Could anyone here offer any insight as to why this might be. Also, some of the reads near the 5'-end of the contig have their pair near the 3'-end of the contig. Does this imply a circular conformation (I'm thinking rolling circle) or could it be due to the presence of perhaps 100s of copies of the phage genome around the time of sequencing?

[maubp]

Some linear phage do (at least temporarily) form circles, so that might be why.

Another possibility is the enriched ends might reflect PCR primers used in any selective amplification step?

[ctseto]

What's the insert size distribution for the reads at the ends? If the insert size is equal to the length of the contig, it would be the ends pairing with each other across a rolling circle, perhaps.

[rhinoceros]

Some linear phage do (at least temporarily) form circles, so that might be why.

Another possibility is the enriched ends might reflect PCR primers used in any selective amplification step?

This is from an environmental total DNA sample that was MDA WGA'd prior to sequencing. As far as I can tell from the protein content, this particular phage has a dsDNA genome, so rolling circle is possible. I still don't quite understand why it would result in higher coverage of the putative origin of replication region though.

What's the insert size distribution for the reads at the ends? If the insert size is equal to the length of the contig, it would be the ends pairing with each other across a rolling circle, perhaps.

I have to look into that in more detail. Can this be parsed from a sam or a bam file (I'm not really familiar with either format) with awk or something?