question re: sequencing past lesion

[anzo]

I used miseq for a funky purpose: I wanted to know what a particular polymerase is incorporating opposite a defined lesion in a synthetic construct.

First I search for sequence that was incorporated upstream of the lesion.

When the polymerase puts what it "should have" placed across from the lesion, the sequence downstream of the lesion looks normal. All good.

However, when it puts what it "shouldn't have" past that lesion, I see seemingly random runs of A's and T's, sometimes C's that would have to be nontemplated incorporation. Are those real, or is this an artifact of the platform?

For instance, suppose my polymerase never put anything across from the lesion (I do know that it frequently stalls), and I basically fed the machine a long template hybridized to a shorter extended primer that is stalled at that lesion. I was just expecting the sequence to end, but it never seems to. Is there some step during library prep that would make it look like random A's and T's after that lesion (like using a polymerase that can do nontemplated extension)?

I would really appreciate some insight, thanks!