The calculation of the number of animals required for your study involves several factors, including the desired level of statistical power, the expected difference between the treatment and control groups, and the variability in the measurements.

One approach to calculate the number of animals needed is to use a sample size calculator, which takes into account the factors mentioned above. A popular sample size calculator is the one provided by the University of California, San Francisco (UCSF). You can access it online through the following link: .

To use the UCSF sample size calculator, follow these steps:

Using the values you've described, here's what the calculator output would look like:

• Desired Power (1 - β): 0.8
• Effect Size (δ): 0.2 (a conservative estimate, representing a 20% reduction in blood glucose levels)
• Allocation Ratio (r): 1.5 (3 doses per group / 2 controls)
• Number of Observations per Group (n): 6 (3 animals x 2 time points)
• Type I Error (α): 0.05

The calculated total sample size is approximately 42 animals. However, considering that some animals might not survive or exhibit unexpected responses, it's advisable to add a few extra animals to each group. A common practice is to increase the sample size by 10-20%.

Therefore, to ensure adequate statistical power and account for potential losses, we recommend the following animal numbers for your study:

• Control groups (untreated): 8 animals (4 animals x 2 time points)
• Treatment groups (STZ-treated): 24 animals (6 animals x 2 time points x 3 doses)

In summary, a minimum of 32 animals (8 controls + 24 STZ-treated) would be required for your study, with a recommended adjustment to account for potential losses, bringing the total to around 40 animals.

Please note that this calculation assumes similar variance between the treatment and control groups. If there's considerable variation in the measurements within a particular group, you may need more animals to maintain adequate statistical power. Consultation with a statistician experienced in animal research would help refine the sample size estimation and ensure the validity of your results.

All the best

First of all, you need to use both sexes for your experiments, since there might be differences which might be substantial. Secondly, it depends on the gap result between the treated animals and controls (they should be the same number) especially for microvascular outcome. If the difference is small, you will need a large number of animals to detect it; otherwise can just be a variance of individual animals. I would not advise less than 10 animals per sex/group with standardized collection of the data and optimal histology to detect differences with proper staining, IHC, and even semithin sections for possible TEM examination. Since it is a new model, you need to validate it and the burden is higher than for an established model.

Hi, Saif Wahid

The new site for sample size calculation is . Unfortunately, there is no option for parallel group design, so I was unable to input the parameters you suggested with the corresponding figures in your answer. Can you kindly assist me?