It would be nice if you included a feature that generates a report highlighting regions with no transcription activity. When your model identifies regions with high transcription activity, it should also predict which mutations could shut down or slow transcription in those areas. This would make the data more useful and help users design constructs more effectively, with reasonable implications.