Microbiome Discovery 11: Statistical testing part 2

[@RobertKwapich]

Hi Dan Knights, what other learning sources (essential books, lectures, papers) would you suggest for techniques/pipelines in studying microbiome? Especially for statistical testing. 18:00 we use absolute abundances with "-a", not relative :)

[@RobertKwapich]

Another question: 18:21. If we are operating on counts, then "pruning" or "subsetting" out genus table with ">= .1 " will not give us genuses with abundance greater than 10%. We're operating on counts right now, and coun't could be unrarefied - i.e. different number of total count for each sample.

[@desichase79]

Hi, thanks for providing these videos. I have a question. I used spearman correlations to test associations between levels of stress and RELATIVE abundances of certain microbes. Many were significantly associated, however, the microbiome data were not normally distributed so I conducted Negative Binomial regressions as a follow-up. Now, nothing is significant and I think it's because I switched from using relative abundances in the spearman correlations to using absolute abundances for the negative binomial regressions. In which types of regressions can I continue to use RELATIVE abundances instead of absolute? Thanks!! Any tips would be EXTREMELY appreciated! p.s. I did not use the same package as you use here to run the Negative Binomial regressions. Could that be the problem?

[@yikeshen2971]

18:10 it's absolute abundance, not relative abundance.

[@AD-uw3xs]

Hello! Thank you for your tutorials. I have been working through them to understand statistical analyses for microbiomes better. I am now at the section of GLM. Whn doing analyses with edgeR package, do I have to run the wrapper script every single time? I thought that could be the problem, since removing rare taxa with 'genus.a[,colMeans(genus.a > 0) >= .1]' (Error in colMeans(genus.a > 0) : 'x' must be an array of at least two dimensions) and running 'colnames(genus.a)

[@VSerman13GR]

Thanks for another excellent lecture. One question, what is a confounding variable?

[@patitadelrosario4910]

Very good videos, Thank you so much. I am looking for the analysis PERMANOVA which has been used in last papers of microbiota (animal). I need to understand concepts, when and how to use it, then how to run it. Could you please explain about that? Best regards from Chile.

[@fulaibaowang]

very useful, straightforward course, found it helpful even though I used mothur for already 2 year