I have two treatment groups of sample plus a group control. Each treatment and control have own three replicates. I am using Qiime2 for data analysis. I am in dilemma to the following ways below. My question is here:
For normalization prior to statistical analysis, do I need to get an arrange value of three replicates then normalize my data or without calculation the average?
Hello Mehrdad,
I think it will be better to normalize each replicate separately, and treat each replicate as a different sample, for preliminary analysis to check that the replicates are close.
Next, to reduce noise, you can take the average of the normalized values for the replicates.
I hope this helps.
Do you mean the normalization that is done to approximate the data to a normal (Gaussian) distribution? If that is, it is done directly on the original values and the more values (replicates or repetitions per treatment) you have, the better. If you don't have limitations of resources (money and time) to increase the number of replicates, I recommend it.
Regards !
I would definitely leave each replicate as a separate data point and normalize them individually. More discrete data points will generally give a more statistically robust result. I would suggest testing for normality before you go ahead and normalize the dataset, though. You'll want to be able to explain why you chose to normalize. While most biological data needs some form of normalization, it's not a bad idea to take a look first.
- Badges
- Science topic