Thanks a lot @Mehrbod_Estaki, but in order to generate the shannon, I use this command:
qiime diversity alpha
Is it correct? Or my input file should be something like my otu table?
Next, for the H, p and q , value, is this pic shows that the middle group is most abundance among the three with a Shannon's index of H=1.23, yet all three are not significantly different from each other?
Is it possible that we got p-value show to be not significant but q value show the opposite? Which one should we follow in this case?
The alpha diversity value is quite low for Shannon, for a “diverse” mucosa surface you would expect a Shannon Index of around 3-5. I’m not quite sure what H is, but if it is the Shannon Diversity Index it would suggest that on taxa has dominated your sample. I would generally go by the p-value, but it would appear that the p-values are not significant.
I’ve moved this question to its own post as it was unrelated to the original thread’s topic. We ask that users post new questions under a separate topic instead of tagging along other topics as to keep the discussions on point. Thanks!
Your commands to create your shannon index artifact look correct to me. If you’ve been following the tutorials your table.qza is in fact analogous to an OTU table, the difference being it was created using a denoising method and not OTU picking.
Not quite. Alpha diversity metrics estimate a community’s species diversity with regards to their richness and/or evenness. In the case of the Shannon index, it takes into consideration both richness and abundances of the species. Each metric has its own values and therefore interpretations. I therefore again recommend doing some reading on understanding the concepts of alpha diversity, specifically the metrics you are choosing to analyse so you know what their values mean.
The exact Shannon index values in your analyses would appear in the boxplots (which you haven’t included here) above your image and are different than the H value you have included.
The H values you show here are actually related to the Kruskal-Wallis test specifically, it is the test-statistic value for this test and doesn’t have any connection to “abundances”.
Based on your output here, we would look at the q-values, which are adjusted p-values, and interpret that there are no differences between the groups. The q-values should never be lower than the p-value since they are adjusted p-values to avoid false-positives inherited from multiple-testing.