Dear all

I expect this is more a question for the developers, but perhaps other users have relevant experiences to share.

With 9v5 we have the very useful Z scores from a normalized DOPE assessment. The manual says that +ve scores are bad, and that anything less than -1 or so is 'native-like'. Presumably the best achievable Z score will depend very much on the available templates - we should be able to do much better with 70% identical templates than with 30% identical templates. Has anyone got a feel for what kind of figures we should be happy with (and be able to quote in a paper as supporting the quality of our model) in different ranges of template % sequence identity?

Thanks in advance

Dan