base taxon in complex study design #143
Replies: 1 comment
-
Also A follow up to my question above. when comparing diversity across groups is it possible to have multiple fixed effects and possibly interaction effects. for example in the code below I am getting an estimate of the effect of "treatment" as defined by X. But you can see from my plot that I am also interested in "filtersize". I would like to specify a model that has the format ( = treatment * filtersize). But when I try to add multiple variables to X <- breakaway::make_design_matrix(ps_divnet, "treatment" + "filtersize) the code doesn't run and I am not sure how to specify if the model should be additive or include interaction effects even if I were to figure out how to add multiple fixed effects. I have also tried coma separated and even as C(list) Thank you. divnet_result$shannon %>% estimates <- divnet_result$shannon %>% summary %$% estimate |
Beta Was this translation helpful? Give feedback.
-
Hi, so I am trying to calculate some beta diversity estimates for 16S data. And I run into the problem that I do not have an ASV that is shared across all samples. I have read the supporting materials and I understand the suggestion to just choose a few abundant taxon and show that the results a robust to taxon choice, but I still have some questions that are specific to my data structure.
My data is made up of distinct communities that do not overlap in ASVs i.e. communities: A, B, C and I also have mixed composite communities AB, ABC. The number of samples I have per community is quite limited because this is from laboratory flask experiments so lets say 4 biological replicates for every community type both exclusive and composite.
My question is: Is it best that I subset my data by community types and calculate the indexes of interest ie. Shannon for that community type and then repeat for the next community and so on using (base_taxon = "NULL")? If I do this can the results be compared across community types? Or should I choose a highly abundant ASV for a base_taxon even if its completely absent from some community types? I have run the code both ways and the results are similar but not identical and I want to know which approach is more statistically sound for the use of your tool. Thank you very much for your time and if my question is not clear let me know.
divnet_result <- ps_divnet %>%
divnet(base_taxon = "NULL")
Beta Was this translation helpful? Give feedback.
All reactions