For fixed
, the IBD probabilities lie on a line going
through 
, and given by
Hence, the trinomial probabilities may be re-parameterized as
, where 
yields the strict-dominant case, and
t=1 corresponds to the case of no DS at the candidate locus .
Quasi-dominant probabilities are very close to the additive probabilities, i.e. to the line 
. Also, there is a
small overlap between the IBD probabilities of quasi-recessive and
quasi-dominant models, and a large region of the ASP triangle is not covered by either model.
Figure 3: ASP quasi-recessive and quasi-dominant IBD probabilities.
Strict-recessive model: the IBD probabilities lie on the
Hardy-Weinberg curve joining (0,0,1) to
. Quasi-recessive model: the
lines under the Hardy-Weinberg curve are the IBD probabilities for
fixed p and varying r. For fixed p, as r increases from 1 to
, the IBD probabilities move along a line from
to a point on the
Hardy-Weinberg curve. Strict-dominant model: the IBD probabilities
are on the curve joining 
to
. Quasi-dominant model: the
lines above the strict-dominant curve are the IBD probabilities for
fixed p and varying r.