Mismatch in read quality between input and simulated reads

[ezherman]

Hi,

We have been using NanoSim v3.2.3, and comparing the read qualities in input and simulated reads.

Input

We are characterising on approx 300k reads with the following quality distribution:

Output

We are simulating 100k reads. The simulated reads have a narrower quality score distribution, without many reads below Q20 or above Q30:

Questions

Is this expected behaviour? And are there parameters we could tweak?

Thanks!