I am doing optimization of fungi in liquid fermentation. There are 36 total number of experiments in my case. First I will take Dry weight of fungi and after that I want to do HPLC analysis to check the concentration of different compounds. For D.W I am going with 3 replications of each experiment, So how many replication will be used for HPLC? If 3 then 36*3=108 sample?
108 will be the sample analyzed in HPLC.
Replications are the numbers of sample which come from the same experimental (and sampling) conditions.
If you are intending to perform some multivariate analysis (PCA, MDS, NMDS, etc...) later on, I will advise running at least five (05) replicates per treatment (5*36=180).
Roberto Molteni Thank you very much i got my answer. I will go with 108 samples for HPLC.
Sir I have an another question, Normally the incubation period for my fungus is 7 days.One of my friend told me that for HPLC you need to incubate fungus for at least 18 days. Will I go with 7 days or 18 days??? can you guide me?
Dear Hamza Noushahi ,
if your friend gave you explanation for 18 days, e.g. for your strain, experimental conditions and compounds that you want to check, you can easly guide yourself because you know better of anyone your experiment.
Dear Steve B. S. Baleba ,
it is good to have more data for multivariate procedure, but why at least 5 replicates?
Hamza Noushahi wants to check the concentration of different compounds and only based on the number of these compounds we can suggest a minimum number of replicates.
3 replicates per sample is fine which means you can calculate a %RSD value and a mean value. 108 HPLC injections is fine but it may mean an overnight sampling/injection sequence.
Roberto Molteni, more you have replicates, better are your conclusions (closer to what is happening in the reality). That is why I have suggested 5 replicates.
3 replicates is common in the pharmaceutical industry, which has the highest rick to the consumer.
Thank you, dear Bruce Neagle for that precision. In the discipline of chemical ecology 5 replicates is common.
Best
Steve Baleba
Even though 3 replicates are sufficient, nowadays, pharmaceutical methods demand 6 replicates for RSD calculation.
You could try designing a DOE to look the most relevants variables in ur objective. It could help you for a further analysis in the future. Nevertheless, you can try doing 3 injections per sample to make some stadistics. I have read some articles where autors do from 3 to 6 injections per sample. Some of then do that every 2 days per example. So if you wanna see the consumption during the fermentation time you can variate the days of sampling and measurement.
If you want to do a good RSD in ur method to identify those compounds by HPLC, I suggest you inject the standard solutions (not the samples) at least 6 times at different concentrations.
The number of replicates needed depends on two aspects:
1) the imprecision of your method (often concentration dependant)
2) the alloable error for the result
For example is your method has an imprecision of 5% and the allowable error is 10%, you could do with two replicates. However if your allowable error is 2% you need 3 or more replicates.
A statistician can give you a more precise answer.
You have first of all assume that the results of your method has a normal distribution, by a validation study. Assuming that, you guarantee that if you inject 3 or more replicates the results must have the RSD obtained by your validation.
- Badges
- Science method