Domain error while fitting

Estimation
1

Hi, can someone help me understand the error while fitting. I am not able to interpret where the mistake is

ERROR: DomainError with x > shift must hold. Got
x => 0
shift => 0:

Stacktrace:
 [1] macro expansion at C:\Users\julia\.julia\packages\ArgCheck\3oH4Y\src\checks.jl:180 [inlined]
 [2] inverse at C:\Users\julia\.julia\packages\TransformVariables\a4AMY\src\scalar.jl:75 [inlined]
 [3] inverse!(::SubArray{Float64,1,Array{Float64,1},Tuple{UnitRange{Int64}},true}, ::TransformVariables.ShiftedExp{true,Int64}, ::Int64) at C:\Users\julia\.julia\packages\TransformVariables\a4AMY\src\scalar.jl:23
 [4] _inverse!_tuple(::Array{Float64,1}, ::Tuple{TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},Pumas.PDiagTransform,TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64}}, ::Tuple{Float64,Int64,Int64,Int64,PDMats.PDiagMat{Float64,Array{Float64,1}},Float64,Float64}) at C:\Users\julia\.julia\packages\TransformVariables\a4AMY\src\aggregation.jl:196
 [5] inverse!(::Array{Float64,1}, ::TransformVariables.TransformTuple{NamedTuple{(:tvcl, :tvvd, :tvvmax, :tvkm, :Ω, :σ_add, :σ_prop),Tuple{TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},Pumas.PDiagTransform,TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64}}}}, ::NamedTuple{(:tvcl, :tvvd, :tvvmax, :tvkm, :Ω, :σ_add, :σ_prop),Tuple{Float64,Int64,Int64,Int64,PDMats.PDiagMat{Float64,Array{Float64,1}},Float64,Float64}}) at C:\Users\julia\.julia\packages\TransformVariables\a4AMY\src\aggregation.jl:237
 [6] inverse(::TransformVariables.TransformTuple{NamedTuple{(:tvcl, :tvvd, :tvvmax, :tvkm, :Ω, :σ_add, :σ_prop),Tuple{TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64},Pumas.PDiagTransform,TransformVariables.ShiftedExp{true,Int64},TransformVariables.ShiftedExp{true,Int64}}}}, ::NamedTuple{(:tvcl, :tvvd, :tvvmax, :tvkm, :Ω, :σ_add, :σ_prop),Tuple{Float64,Int64,Int64,Int64,PDMats.PDiagMat{Float64,Array{Float64,1}},Float64,Float64}}) at C:\Users\julia\.julia\packages\TransformVariables\a4AMY\src\generic.jl:206
 [7] fit(::PumasModel{ParamSet{NamedTuple{(:tvcl, :tvvd, :tvvmax, :tvkm, :Ω, :σ_add, :σ_prop),Tuple{RealDomain{Int64,TransformVariables.Infinity{true},Float64},RealDomain{Int64,TransformVariables.Infinity{true},Int64},RealDomain{Int64,TransformVariables.Infinity{true},Int64},RealDomain{Int64,TransformVariables.Infinity{true},Int64},PDiagDomain{PDMats.PDiagMat{Float64,Array{Float64,1}}},RealDomain{Int64,TransformVariables.Infinity{true},Float64},RealDomain{Int64,TransformVariables.Infinity{true},Float64}}}},var"#69#84",var"#70#85",var"#72#87",ODEProblem{Nothing,Tuple{Nothing,Nothing},false,Nothing,ODEFunction{false,var"#73#88",LinearAlgebra.UniformScaling{Bool},Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing},Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{,Tuple{}}},DiffEqBase.StandardODEProblem},var"#74#89",var"#79#94"}, ::Array{Subject{NamedTuple{(:CONC,),Tuple{Array{Union{Missing, Float64},1}}},Pumas.ConstantCovar{NamedTuple{,Tuple{}}},Array{Pumas.Event{Float64,Float64,Float64,Float64,Float64,Float64,Nothing,Int64},1},Array{Float64,1}},1}, ::NamedTuple{(:tvcl, :tvvd, :tvvmax, :tvkm, :Ω, :σ_add, :σ_prop),Tuple{Float64,Int64,Int64,Int64,PDMats.PDiagMat{Float64,Array{Float64,1}},Float64,Float64}}, ::Pumas.FOCEI; optimize_fn::Pumas.DefaultOptimizeFN{Nothing,NamedTuple{(:show_trace, :store_trace, :extended_trace, :g_tol, :allow_f_increases),Tuple{Bool,Bool,Bool,Float64,Bool}}}, constantcoef::NamedTuple{,Tuple{}}, omegas::Tuple{}, ensemblealg::EnsembleSerial, checkidentification::Bool, kwargs::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{,Tuple{}}}) at C:\User88",LinearAlgebra.UniformScaling{Bool},Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing,Nothing},Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{,Tuple{}}},DiffEqBase.StandardODEProblem},var"#74#89",var"#79#94"}, ::Array{Subject{NamedTuple{(:CONC,),Tuple{Array{Union{Missing, Float64},1}}},Pumas.ConstantCovar{NamedTuple{,Tuple{}}},Array{Pumas.Event{Float64,Float64,Float64,Float64,Float64,Float64,Nothing,Int64},1},Array{Float64,1}},1}, ::NamedTuple{(:tvcl, :tvvd, :tvvmax, :tvkm, :Ω, :σ_add, :σ_prop),Tuple{Float64,Int64,Int64,Int64,PDMats.PDiagMat{Float64,Array{Float64,1}},Float64,Float64}}, ::Pumas.FOCEI) at C:\Users\julia\.julia\packages\Pumas\iIBBr\src\estimation\likelihoods.jl:1995
 [9] top-level scope at none:1```
2

Dear Arun,

Two things to check.

  1. Do you have any zero concentrations in your dataset.
  2. Do you have any time zero in your dataset after the dose was given?

Could you share the model?

Best,
Vijay

3

Hi Vijay, I checked the dataset for zero concentrations and time zero after dose was given. There are none. Below is the model

    @param begin
        tvcl    ∈   RealDomain(lower=0, init=0.05)
        tvvd    ∈   RealDomain(lower=0, init=1)
        tvvmax  ∈   RealDomain(lower=0, init=1)
        tvkm    ∈   RealDomain(lower=0, init=0)
        Ω       ∈   PDiagDomain(init=[0.01,0.01,0.01,0.01])
        σ_add   ∈   RealDomain(lower=0, init=0.01)
        σ_prop  ∈   RealDomain(lower=0, init=0.01)
    end
    @random begin
        η       ~ MvNormal(Ω)
    end
    @pre begin
        Cl      =   tvcl*exp(η[1])
        Vd      =   tvvd*exp(η[2])
        Vmax    =   tvvmax*exp(η[3])
        Km      =   tvkm*exp(η[4])
    end
    @dynamics begin
        Central'=   - Cl*(Central/Vd) - ((Vmax/(Km+(Central/Vd)))*(Central/Vd))
    end
    @derived begin
        cp      = @.(Central/Vd)*1000
        dv      ~ @.Normal(cp,sqrt((cp*σ_prop)^2+σ_add^2))
    end
end```
4

Could you please share the starting values as well? This error typically happens when you start the fit outside the parameters bounds. However, in most cases we should now be able to give a more informative error message in these cases so could you please let us know which version of Pumas that you are using.

5

Hi Andreas,
The error is sorted now, thanks .
As you had mentioned, its the issue with the parameter bounds where i had given same value (zero) to both the lower bound and initial for one of the parameter.

FYI, I’m running on Pumas v1.1.0