Hi
I have alpha group significance results for three different groups of 2 samples in each. I am wondering how qiime2 calculates the median based on two samples.
How correct these results are? How to analyze these results based on the similarities or differences among groups?
Thanks and regrads
Rashmi Ira
Hello!
Let's go through it step-by-step:
Sampling size is too low and I would be very careful with reporting such results. At least 5 samples per group would be much better.
In that case, the median is equal to the mean since there are only two samples per group. You can see the boxplots look like they were calculated for the mean value, not for the median (perfect symmetry).
You can use this data as preliminary results (indicating that there is potential and that more studies with a bigger sampling size are required), but you can't draw any conclusions from it as it is since the sampling size is too low.
Comments to the legend:
Boxplots represent alpha diversity based on Shannon entropy. Significance can be added with asteriscs or p-values for each pair. Treatment is a factor, not the group, since you have several groups within the Treatment as a factor. Kruskal-Wallis test was used for comparing groups within Treatment, not for obtaining Shannon entropy. Shannon entropy was calculated before you run alpha group-significance plugin.
Best,
Thank you so much @timanix for such a good explanation. I will work on it and do correction as suggested.
Best Regards,
Rashmi
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.