This is a serious tragesty -- there is no reasonable use for such a function as anybody who insists on using known-inaccurate input data basically deserves to get a ton of errors. Except ...
... except we have an already-published, already-referenced genome that we need to functionally annotate, which means we have to sneak it past tbl2asn, which means we have to reproduce whatever painstaking edits and fixes we made last February.
I propose a command that looks at every single mRNA and performs a six-frame translation on its CDS (maybe just the first segment is enough?), then decides the phase of the CDS based on which translation doesn't have a million internal stops.
I guess we don't want to change strand, so maybe I should've said "does a three-frame translation". Anyway, this function should get us back to an error-free .tbl file.
This is a serious tragesty -- there is no reasonable use for such a function as anybody who insists on using known-inaccurate input data basically deserves to get a ton of errors. Except ...
... except we have an already-published, already-referenced genome that we need to functionally annotate, which means we have to sneak it past tbl2asn, which means we have to reproduce whatever painstaking edits and fixes we made last February.
I propose a command that looks at every single mRNA and performs a six-frame translation on its CDS (maybe just the first segment is enough?), then decides the phase of the CDS based on which translation doesn't have a million internal stops.
I guess we don't want to change strand, so maybe I should've said "does a three-frame translation". Anyway, this function should get us back to an error-free .tbl file.