Accounting for phylogenetic relation between samples in the formula?

[Jigyasa3]

Hey!

I have two questions-

1. I have a biom file of microbial KEGG pathways for 100 host samples which are phylogenetically related. I was wondering if I can account for phylogenetic relation between host samples in the co-variates via patristic distance?

For example,
--formula "Depth+Temperature+patristic(phy)"

2. Are the other flags (as given below) default in songbird command?
   --epochs 10000
   --differential-prior 0.5
   --summary-interval 1 \

[mortonjt]

Question 1 : See our paper on phylogenetic methods. It sounds like you want something like phylogenetic GLS - but it's not clear to me. I suggest looking at phylofactor : https://github.com/reptalex/phylofactor Songbird does not support phylogenetic methods, and won't for the foreseeable future. Question 2 : the defaults should be in the command. Does it show when you run it on the command line?

…

On Tue, Nov 12, 2019, 1:31 AM Jigyasa3 ***@***.***> wrote: Hey! I have two questions- 1. I have a biom file of microbial KEGG pathways for 100 host samples which are phylogenetically related. I was wondering if I can account for phylogenetic relation between host samples in the co-variates via patristic distance? For example, --formula "Depth+Temperature+patristic(phy)" 1. Are the other flags (as given below) default in songbird command? --epochs 10000 --differential-prior 0.5 --summary-interval 1 \ — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#97?email_source=notifications&email_token=AA75VXP3ZBY4OVZATAZJI4TQTJETRA5CNFSM4JL64RE2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HYTII2Q>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA75VXK6I6OG7ZJBUST4PP3QTJETRANCNFSM4JL64REQ> .