I have gone through the tutorials for both differential abundance testing methods (ANCOM and gneiss). It has come to my attention that the add-pseudocount step is necessary first step in both methods, but it was done in different ways (composition add-pseudocount and gneiss add-pseudocount). When I read the description of each script, they are 100% similar:
So I don’t get why there have to be 2 scripts in 2 plugins (composition and gneiss) if they do the same thing? Or if they are somewhat different and I haven’t been able to recognize the difference? If they are actually similar, would it be more suitable to have one script only (composition add-pseudocount) for less confusion?
Yes, these do the same thing as far as I'm aware. Long story, but to keep it short:
- Different plugins are designed by different people and sometimes these are 3rd parties, so some overlap is almost inevitable. Often plugin developers design these packages to be standalone packages and other times they just aren't aware of functionality in other plugins...
- q2-composition was actually briefly removed from a year or two ago, so q2-gneiss really did need its own
So I agree with you — one makes more sense, but two does no harm, and in this particular case can be important for the purposes of making these plugins more independent.
Thanks for asking!
Thanks Nick. Does this mean that the add-pseudocount step should be performed with gneiss add-pseudocount from now (even for ANCOM analysis)?
No, I mean it just does not matter — presumably the two are interchangeable.
It is useful to keep both, though. In the future, users will only download the plugins they need from a “plugin library” instead of automatically downloading all available plugins. So someone might have q2-gneiss but not q2-composition, or vice versa. So having this simple utility method replicated in each of these may be useful, if redundant.
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.