Time series analysis

[amralasra]

Hi
I have a critical question
i am trying to use a multiple regression of two independent variables and one dependent variable.
however, the model is time series so i have to test for stationarity. i found that the variables are not stationary so i have to test cointegration.
i read that in case of multiple regression i should use johansen cointegration test, but when i tried it there were alot of options to specifiy and i don't really understand all of them? so how can i run the test in my case?

also, i am trying to use the log transformation for one of the varaibles, so in this case, to test its stationarity, should i test the original variable or the transformed one?

in case i am trying to use time lagged data, that is yn and x(n-1), how can i estimate the multiple regression? is there special command that will process the original data to run the lagged model?

for one of the varaibles x, the scatter plot with y is positively sloping, however, the multiple regression shows negative coefficient? is this normal?

plus the model shows insignifiacnt coefficients although R2 is really high so what could be the interpretation in this case?

too many questions, but i really need help urgently and as soon as possible.
thx