Instructor's Note

Start with the 01-covariate.jl file. In this file you'll guide learners on how to build covariate models. Start with the DataFrame of the nlme_sample dataset from PharmaDatasets. You'll see that this dataset has the standard PK dataset columns such as :ID, :TIME, :DV, :AMT, :EVID and :CMT. Spend some time to remind learners about these NM-TRAN standard columns, as well how to pass them as values to the keyword arguments of the read_pumas function. The dataset also contains the following list of covariates:

• :WT: subject weight in kilograms
• :SEX: subject sex, either "F" or "M"
• :CRCL: subject creatinine clearance
• :GROUP: subject dosing group, either "500 mg", "750 mg", or "1000 mg"

If your audience is data-centric and time allowing, you can spend some time doing some data exploratory analysis. If you do, don't forget to load the DataFrames/DataFramesMeta package(s). The first step in our covariate model building workflow is to parse data into a Population. This is accomplished with the read_pumas function. Here we are to use the covariates keyword argument to pass a vector of column names to be parsed as covariates. The second step of our covariate model building workflow is to develop a base model, i.e., a model without any covariate effects on its parameters. This represents the null model against which covariate models can be tested after checking if covariate inclusion is helpful in our model. The third step of our covariate model building workflow is to actually develop one or more covariate models. Let's develop two covariate models:

1. allometric scaling based on weight
2. clearance effect based on creatinine clearance

To include covariates in a Pumas model we need to first include them in the @covariates block. Then, we are free to use them inside the @pre block As you showcase the base and covariates models highlight the differences amongst them. Take note that the second covariate model needs a different set of initial parameters estimates due to having extra parameters. For the first covariate model, the one that does allometric scaling based on weight, once we included the WT covariate in the @covariates block we can use the WT values inside the @pre block. For both clearance (CL) and volume of the central compartment (Vc), we are allometric scaling by the WT value by the mean weight 70 and, in the case of CL using an allometric exponent with value 0.75. For the second covariate model, the covariate_model_wt_crcl model, we are keeping our allometric scaling on WT from before. But we are also adding a new covariate creatinine clearance (CRCL), dividing clearance (CL) into hepatic (hepCL) and renal clearance (renCL), along with a new parameter dCRCL. dCRCL is the exponent of the power function for the effect of creatinine clearance on renal clearance. In some models this parameter is fixed, however we'll allow the model to estimate it. This is a good example on how to add covariate coefficients such as dCRCL in any Pumas covariate model. Note that we need a new initial parameters values' list since the previous one we used doesn't include dCRCL, tvcl_hep or tvcl_ren. Now that we've fitted all of our models we need to compare them and choose one for our final model. We begin by analyzing the model metrics. Highlight the AIC values between the models, prefer the lowest value. Additionally, we should inspect the goodness of fit of the model. This is done with the plotting function goodness_of_fit, which should be given a result from a inspect function. Go over the plots comparing the three models. Finally, we also perform VPCs to help the model comparison task.

Now, proceed to the dose control parameters (DCP) workflow with the 02-dose_control.jl. Here the idea is to showcase the new @dosecontrol block. Explain that this block allows for four special variables:

• lags: the lag of the dose
• bioav: the bioavailability of the dose
• duration: the duration of the dose
• rate: the rate of the dose

These are specified with the syntax dcp = (; Cmt=value) where:

• dcp: the dose control parameter (lags, bioav, duration or rate)
• Cmt: the compartment where the DCP will be applied
• value: value to use the for the DCP

Here you will have three models to show learners. First, the lags model will have the lags as a DCP in the @dosecontrol block. Here you will explain how the learners can add compartments to which the lags will have an effect, and which value it will take. This is done with a NamedTuple. Second, the bioav model has a different DCP, bioav, but the logic is the same. Finally, the lags_bioav has two DCPs, lags and bioav. Here the idea is to showcase that you can not only have multiple compartments but also multiple DCPs. Be careful with random-effects (η) on the DCPs since those can include discontinuities in the objective function, and may give unstable estimates during the fitting procedure.

Finally, we'll move into indirect response models, and also into PD models. Proceed to the 03-indirect_response.jl file. Here the pkdata has two columns of observations: the :dv and the :resp columns. Explain that the :dv is for the PK measurements and the :resp for the PD measurements. This is a sequential PD model. That means, that we'll fit first the PK part of the model. The PK model is not different from the models that learners have experienced in the pre-requisites of this workshop. Then, we proceed by extracting the PK individual parameters from the fit and merging to the original data. We read the updated data into a Population and pass the PK individual parameters as covariates in the read_pumas function. Now, walkthrough the PD model definition. Make sure that learners understand the different PD components in the @param, @random, @pre and @derived block. There are two new model blocks that learners might not have been exposed yet: @init and @vars. The @init block has the purpose of defining initial values for the subjects compartments. The compartment always has an initial value of 0 or the dosing event at time 0 if not specified with @init. In the @init in the model we are only specifying initial values for the PD compartment Resp. The @vars block is used to define aliases that can be used in the @dynamics and @derived blocks. These aliases can use any variable defined in the @pre and @covariates block and any Julia variable or value, user-defined or not. This is a good way to declutter your ODE equations in the @dynamics block. Finally, showcase a GoF plot for the PD inspect result. The important part here is to showcase the flexibility and ease-of-use of Pumas modeling syntax to define complex models without overloading the user.

Get in touch

If you have any suggestions or want to get in touch with our education team, please send an email to training@pumas.ai.

License

This content is licensed under Creative Commons Attribution-ShareAlike 4.0 International.