#19 ECO:0000200 sequence alignment evidence + children

open
nobody
None
5
2011-09-05
2011-09-05
alan_bridge
No

Hi

I just had a query about the child terms of "ECO:0000200 sequence alignment evidence states". The definition for ECO:0000200 states.that the term is

"Used when a sequence alignment is the basis for making an annotation, but only when the alignment has been manually reviewed"..

yet the child terms of this allow for use in manual or automatic assertion, e.g.

ECO:0000248 sequence alignment evidence used in automatic assertion
ECO:0000247 sequence alignment evidence used in manual assertion

This seems to preclude the use of sequence alignments for fully automated annotations, i.e. you could not use this evidence code for automatic assignment of function to a protein based on a pairwise alignment. You could use "ECO:0000248 sequence alignment evidence used in automatic assertion" though if you made an annotation pipeline that spits out BLAST results that are subsequently checked by a human, then placed back in the pipeline, but that seems unrealistic.

I wondered why the necessity for alignments to be reviewed? Would you maybe consider making evidence codes to cover sequence alignment evidence from unreviewed alignments, like this:

1. "Used when a sequence alignment is the basis for making an annotation, irrespective of whether the alignment has been manually reviewed or not".
2. "Used when a manually reviewed sequence alignment is the basis for making an annotation".
3. "Used when an unreviewed sequence alignment is the basis for making an annotation".

Or something similar?

Thanks, Alan

Discussion