> If it's an contamination or an error, it will have very low read coverage (typical sequencing project be 50-200x meaning if you realign and pile up the reads

> If it's an contamination or an error, it will have very low read coverage (typical sequencing project be 50-200x meaning if you realign and pile up the reads back to the main assembly, you get a normal distribution centered around 50-200 reads supporting a particular base or region. If you have a very low coverage region or a sharp drop off in coverage, then it's most likely an error. Wouldn’t sequence-specific biases(capture efficiency, amplification) result in distortions?