I am not sure the form checks this (and whether this is related to the issue here), but it seems weird to have an IMP for the root MF. Does it accept ND?

I am not sure the form checks this (and whether this is related to the issue here), but it seems weird to have an IMP for the root MF. Does it accept ND?

These aren't root annotations indicating "unknown". These are Process only annotations where the root annotation is something Paul insisted on to indicate that some function is involved in this process. I always use the same evidence code for all the lines in a Process only annotation since this is one statement based on the mutant phenotype. ND would not be appropriate as this annotation is not being made on the basis of "no data", but on the mutant phenotype of the affected process indicating that the gene product must have some function.

I also don't think this is relevant since ALL of the annotations I was trying to change were IMP annotations and the one blank that couldn't be completed wasn't the evidence line on the anonymous root MF.

Here's the graph view of what these annotations are: