Linear regression in genetics, interactions

#1
Dear All,

I am searching for the association between a phenotype (PHEN) and the additive genetic model (ADD, with the values 0,1 and 2) with two adjustors, Year of Birth (numeric) and COUNTY (catergory). The data has been weighted for the year of birth. In R I use the next command:

summary(lm(PHEN ~ YoB * COUNTY * ADD, weights = COUNT_WEIGHTS, data = 2_F))

The result of the above command is:
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.764813 4.503829 -0.392 0.695
YoB 0.069811 0.074147 0.942 0.347
COUNTY 0.006501 0.468143 0.014 0.989
ADD 2.169632 2.566560 0.845 0.398
YoB:COUNTY -0.001149 0.007722 -0.149 0.882
YoB:ADD -0.036325 0.042164 -0.862 0.389
COUNTY:ADD -0.074915 0.264138 -0.284 0.777
YoB:COUNTY:ADD 0.001680 0.004346 0.387 0.699

All the p-values are above 5 %, which means that none of the adjustors nor the genetic model have effect on the phenotype, also .

Next, should I change the command into:
summary(lm(PHEN ~ ADD, weights = COUNT_WEIGHTS, data = 2_F))

The result is :
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.95981 0.16061 12.202 <2e-16 ***
ADD 0.18061 0.09182 1.967 0.0496 *

or should I change it into:
summary(lm(PHEN ~ YoB + COUNTY + ADD, weights = COUNT_WEIGHTS, data = 2_F))

The result is:
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.68157 0.63565 1.072 0.2840
YoB 0.02355 0.01005 2.342 0.0194 *
COUNTY -0.01870 0.01036 -1.806 0.0713 .
ADD 0.18605 0.09155 2.032 0.0425 *

Which of the options would be the appropriate one.

Thank you very much.

Thierry