HSMC usage to infer microbial communities? [discussion]

[LafontRapnouilTristan]

Hello,

Kind of a weird 'issue' as it's more a 'forum like' topic. I am completely new to JSDMs and have some questions about some thoughts I had about how to use them.

For instance :

• How insane would it sound to use HSMC to predict (try to) microbial communities (e.g. soil), given the complexity of microbiome and the very nature of metabarcoding data (and associated biases)? Would you guys be confident in the output to use it for diversity maps?
• To do so, would you recommend some communities simplification at some point (removing rare species?)
• Is it 'ok' (I know its not meaning much but I would be glad to have your thoughts on this) to aggregate communities at a given taxonomical rank in HSMC (I mean we could make predictions about bacterial order rather than finer taxonomical ranks?). Making the assumption that close phylogenetic entities are responding similarly to environment and accepting that our output would be a composite the responses of different OTUs/ASVs.

I see in the literature that JSDMs are progressively being used in microbial ecology but have not found yet any tentative of communities predictions.

I know github isn't an especially fitted place for such open questions and discussion and I wonder if there is an FAQ or dedicated forum where it could be discussed.

Additionally I acknowledge that my questions are quite broad but it's more about the philosophy behind this kind of usage rather than technical implementations (but if you have thoughts about that, they are more than welcome!)

Thanks to HSMC team and to anyone that will take some time to discuss this. Cheers

[LafontRapnouilTristan]

Re- Thinking about it, would high rank variables (e.g., relative abundances of dominant taxa or alpha-diversity) be reliable even if we aren't confident about the exact predicted OTUs/ASVs presence and counts?