Choose Language Hide Translation Bar
Level I

Custom Design for CFU/mL



I am trying to set up my DOE but cannot seem to generate the right design. Here is my experiment: Three trails (performed on three separate days), 9 treatments (GH, GM, GL, SH, SM, SL, MH, MM, ML) that will be tested with 4 rehydration solutions (NaCl, Peptone, PBS, Maltose). There will be two dilution factors (1:1, 1:10), 3 plates per dilution factor (1, 2, 3), and 0.1ml volume plated. I will record the colonies and use the formula : Colonies x Dilution Factor/ volume. Right now, i have the custom design set up based on the attached file. As you can see, i have to add in the dilution factor, total volume, and colonies columns into the generated design. Not only that, but when I re-organized/grouped the rehydration/treatment, i noticed that there are not enough rows generated by my design for two dilution factors (example Maltose GH). Any help would be great thanks!

Staff (Retired)

Re: Custom Design for CFU/mL

Based on your description, it seems as if you would need to include additional factors in your design. If your intention is to 3 plates for each dilution factor, reydration solution, treatment combination for each day, perhaps a blocking factor of day and the dilution factor need to be included in the model. Please see the screen shot attached. When setting up the Blocking factor, I specified 216 runs per block. Note the 72 replicate runs that should allow for each combination three times each day.






Re: Custom Design for CFU/mL

The first problem is the definition "Number of Replicate Runs" means in the Custom Design tool.

When you set that option to 3, you get a total of three replicated runs. If the minimum number of runs for your experiment was 12 and you put a "3" in the "Number of Replicate Runs" box, then you would get 15 runs as the new minimum, not 36.  (which is what would happen in some of the classical designs.) Clearly a place where confusion can happen


A second issue has to do with the replicate plates and the experimental unit. Each combination of treatment, and solution are one experimental units. The experimental unit is diluted or not (1:10 or 1:1), and then the unit is plated (measured) three times.  The three plates aren't replicates in the design; however, the mean of the three plates does represent the experimental unit.  I would also assume that in the analysis one might only use low dilution counts unless it is TNTC, ortherwise, the higher dilution would be used. 


Because you want to perform all the factor combinations the custome designer builds you a 36 run full factorial design. The interactions can be estimated, in this design, so adding interactions in the custom designer just adds another 36 runs.  


On each day you can set up the 36 treatment solution conbinations. Then make your two dilutions, and the plate each dilution 3 times. Doing 216 plates in a day isn't unreasonable, sounds like fun. I'd join you but... I think I have something going on that day.




JMP Systems Engineer, Pharm and BioPharm Sciences
Article Labels

    There are no labels assigned to this post.