Query on Simulation Setup for In-Vitro PD Model in Pumas

Simulation
1

Hej Pumas Community,

I am currently working on a semi-mechanistic PD model based on in vitro OD data and have successfully translated my model from NONMEM to Pumas.

Since this is in vitro data, there are no dose records. Instead, I have multiple fixed concentrations, with observations collected from 0 to 24 hours at 10-minute intervals.

Additionally, each concentration has multiple experimental replicates. For example:

  • ID = 1–10

  • CONC = 0.001

  • Observation times = 0:0.1:24 hours

I am seeking clarification on the following:

  1. How can I simulate multiple concentrations where each ID corresponds to a single fixed concentration?

  2. How should I structure the dataset to account for multiple replicates per concentration?

  3. How do I specify the total number of individuals in the simulation under this setup?

I would greatly appreciate any guidance on how to properly structure and simulate this scenario in Pumas.

Thank you in advance.

2

Hi Sabariakilesh!

It seems to me that most of what you are asking is straight forward applications of constructing Subjects one at a time. But let me ask some clarifying questions to be completely sure.

What if we focus on one ID one concentration. Then you can construct a subject as follows

sub = Subject(id = “1”, covariates = (; conc = 0.001))

Then you want to simulate at every 10 minutes, so you can use simobs with your model and parameters to simulate this trajectory for this one subject.

simobs(model, sub, param; obstimes = 0.0:1/6:24.0)

Does this correspond to what you want to do conditional on a specific ID and specific concentration?

3

Hej Patrick!

Thank you for the clarification. I was able to successfully construct a single subject without any issues.

However, I am not entirely sure how to simulate multiple replicates per concentration, especially since I am working with multiple concentration levels. Additionally, how would I specify the total number of individuals in the simulation under this setup?

I would appreciate your guidance on how to structure this properly.

4

If you need to build a solid understanding of the concepts you could just build them individually and combine them. Are you familiar with array comprehensions or the `map` function? For example, if you just focus on one concentration level, would you be able to construct a vector with all the individuals you need?