Observations vs. Population predictions plot

Diagnostics
1

Hello,
I ran a 2 compartment model with first order absorption in Pumas and in NONMEM. I used the same initial estimates for both. However, the parameter estimates generated in Pumas are slightly different than in NONMEM and the observed concentrations vs. population prediction plot generated by pumas looks different than what I got through NONMEM (NONMEM results were more evenly distributed above and below the line of unity). I would appreciate any insight or suggestion as to what could be causing these differences.

GOF results from NONMEM
NONMEM_obs_v_pred

GOF results from Pumas
pumas_obs_v_pred

Parameter estimates of pumas vs. NONMEM
image

Thank you!

2

Can you please check it your NONMEM → Pumas model translation is correct?

Please run:

-2 * loglikelihood(pumas_model, pop, nonmen_results, FOCE())

This should be equal to the OFV with constant from NONMEN results.

If that differs, then either your Pumas model or Population are not equivalent from NONMEN.

3

Thank you Jose for your response. Yes, it was the population. Got that resolved.