Run classification on ~500 sequences that are randomly assigned ATCG on each position.
Purpose
From @iamciera on Slack
The main purpose is to test a negative control: I’m hoping that the model classifies the random sequences as non-functional. A real sequence, that has unknown classification, could have function, but it has never been measured. Random sequences likely have no functional structure that is functional, or at least are more likely to have to function then a random sequence.
Run classification on ~500 sequences that are randomly assigned ATCG on each position.
Purpose From @iamciera on Slack