inconsistency deriving from definition of entity of organismal origin

[Alan Ruttenberg]

This was tricky.

The manifestation was that two assays were inconsistent, PCR-SSCP
assay, and radioactivity detection.
Upon fiddling, I found that removing the has_output restriction on
those classes fixed the problem, but I couldn't see why.

In browsing in LSW, I looked at the definition of has_output, which
computes an effective domain and range, and saw entity of organismal
origin mentioned. (surprising)

I went to look at the definition of entity of organismal origin and
it was:

necessary and sufficient:
is output of some protocol application
necessary:
material entity
has_role some (reagent role or specimen role)

The problem is that the necessary and sufficient condition made it
equivalent to anything that was the output of some protocol application.
That meant that any output of a protocol application necessarily was
an entity of organismal origin, and hence material entity, and
has_role some (reagent role or specimen role) .

The two assays had outputs which were images, which are dependent
continuants. Since dependent continuants are disjoint from material,
we arrive at the contradiction.

I've cced Matthew, because neither the debugger in p3 or p4 was able
to explain this.

The repair which will be checked in in a moment was to make the
necessary conditions part of the n & s conditions, i.e. an EOO is
anything that

is output of some protocol application AND material entity AND
has_role some (reagent role or specimen role)

I believe this was the intention.

-Alan