Max specific growth rate from OD data

[#1]

Hi,

I'm extremely new to JMP Pro, so I am not sure if this is a dumb question. I am thrilled that JMP can help me analyze bacterial growth data quickly and effectively using the Cell Growth and Logistic models, but sometimes it is preferable to report the more rudimentary maximum specific growth rate instead of the overall growth rate. Is there an established and high-throughput way to do that in JMP? I don't want to have to ln transform the data, find linear region, fit a linear model, etc. 

Any help would be greatly appreciated.

Thanks,

David

[#2]

Hi @ddooley2 ,

You could try saving the first-derivative formula under the curve you have applied in Fit Curve

You can then take the saved column and use Tables > Summary or Tabulate and create a table with the max first derivative (slope) values.

Hope that helps,

Thanks,

Ben

[#3]

Hi, Ben! Thanks for your quick reply. That's an interesting solution, but if I'm not mistaken, the first derivative reported is the instantaneous slope at each time point (or something that approximates to that). Unfortunately, that is a bit too localized for accurate max specific growth rate. Ideally, the slope would be calculated based on several time points (up to a few hours worth of time points) at which the growth rate is highest to prevent overestimation due to any small positive fluctuations in OD.

Thanks,

David

[#4 · Accepted solution]

I'm assuming the data can be quite easily log transformed in JMP Pro. From there, all I would need would be some kind of trimming technique that confines the data to the linear range in the log scale (i.e. the period of exponential growth). After the data is trimmed, a simple linear regression is all that is needed to find the maximum growth rate... 

[#5]

Yes, it is the instantaneous slope at a given X, but the estimate is based on the fitted curve, which used all the data.

[#6]

Is there a trimming technique in JMP Pro that automatically determines the linear region of a curve?

[#7]

There is no built-in capability to identify the 'linear region.' You could develop a script or make a formula to exclude rows. I advise against trimming the data before fitting. The best practice is to use all the data to fit the model and then use the model to obtain the desired result.