I believe one of the Majiq developers did respond in private, but as he has since left the lab, I've been having trouble contacting him about this specific message. What I do know is that this is data dependent and is only generated on a specific unexpected read case. It would help to be able to see the minimum set of data which generates the error, so that we can analyze exactly what's happening with these reads. Would you be opposed to sharing your input data over private channels to discuss this issue? You can reach out to
sje...@pennmedicine.upenn.edu for private communications and we can resume here once there is a conclusive answer.