nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Fit Curve Analysis - How to converge rate equations

Bingh94

Hello all,

I'm attempting to recreate reaction kinetics work from a publication (Link in doc) to see applicability of the Fit Curve tool on a DSD style data set. The data features substrate concs (z, supplementary), product concs (y) over residence time (x), separated by the pattern of the DoE (ID, Function).

I've successfully created a first order rate fit on one of the product concs ('DCP Output [M] in the data sheet, see script 'DCP - Fit Curve') but can't get any other rate equation fits on the remaining products ('Product [M]', 'Isomer [M]', 'Bis [M]' in data sheet). The paper applies first and second order rate equations with Dynochem Scale-Up systems and MODDE for their DoE, but I appear to get the following error:

The data set is relatively small, with only 5 residence time values. Does anyone know the cause for the failed convergence and/or a way to still apply first/second order rate equations?

Thanks,

Ben

Definitive screening designs for kinetic models in flow Updated.jmp

Accepted answers

Byron_JMP

Woah! this is a fantastic data set!! I'm really glad you posted a question with it.

Ok, so the problem is with the "shape" of the data. The JMP model is expecting something different. There might be an unstated assumption in the name of the model. It looks like these are first and second order rate "decay" models.

I tried fitting 1 / :"Product (M)"n the first and second order rate models converge.

The data for Product, Isomer and Bis are appropriate for a Michaelis-Menten kinetics curve.

All comments

Phil_Kay

Hi Ben,

Just to clarify: I think that in the paper they used Dynochem/Scale Up Systems for the kinetic modelling. MODDE was used for the DOE modelling - just in case anyone is coming up against MODDE in a competitive situation.

Cheers,

Phil

Bingh94

Apologies - yes thanks for clarifying! I've edited the text.

Byron_JMP

Woah! this is a fantastic data set!! I'm really glad you posted a question with it.

I tried fitting 1 / :"Product (M)"n the first and second order rate models converge.

The data for Product, Isomer and Bis are appropriate for a Michaelis-Menten kinetics curve.

Phil_Kay

This page of the documentation is useful, I think. It shows you what shapes of curves you can fit with each model. I guess it might be helpful if you could see these visuals when selecting the model in Fit Curve.

Mark_Bailey

Another approach is to treat this example as a functional DOE. That is, fit functional models for each response. Launch FDE with the factors as supplementary variables. I attached your data table with a table script for this idea.

Definitive screening designs for kinetic models in flow Updated.jmp

Bingh94

This is great! Thank you