mecodev gamma and beta diversity calculations

[jamorillo]

Hi Chi,

I have a question regarding the mecodev package. I noticed that in one of the steps it calculates both beta and gamma diversity using a microeco object, which is very useful. However, I would like to clarify a few points about these numbers:

1. I understand that the gamma diversity calculation refers to an estimation of the total number of ASVs detected in a group of samples (Site). Is that correct? Is there any kind of estimation involved, or are these the total detected ASVs/group?

2. Regarding beta diversity, do these numbers represent the average distance (Bray in this case) within each group? Is it something similar/equivalent to betadisper?

3. Can I apply apply sqrt to calculate betadiversity? then giving more weight to less frequent features. Might be interesting. How?

4. Which paper should I cite if I use mecodev?

Many thanks!
reagrds
Jose

Example with one of my datasets:

microecoITS_norm
microtable-class object:
sample_table have 156 rows and 18 columns
otu_table have 3128 rows and 156 columns
tax_table have 3128 rows and 8 columns
phylo_tree have 3128 tips
rep_fasta have 3128 sequences

diversity_object <- trans_gamma$new(dataset = microecoITS_norm,
group = "Site",
method = "bray")

diversity_object$cal_observed()
The result is stored in object$res_observed ...

diversity_object$res_observed
Sample Group beta gamma
1 observed FR 0.6718374 1637
2 observed DP 0.4799911 172
3 observed MC 0.6504501 705
4 observed BG 0.6054494 763
5 observed NR 0.6180772 514

[ChiLiubio]

Hi Jose,

1. The gamma diversity here is actually the number of total species observed, no estimation.
2. Yes. It is the mean dispersion from the centroid of distance matrix based on the betadisper as you say. Thanks. I will explain these in the package documentation.
3. Sure. You can.
4. No need to cite the paper. Just provide a link to the package.

Best,
Chi

[jamorillo]

Hi Chi,
Thanks for the confirmation of betadisper. One thing more, respect to tansformations of ASV table (sqrt, log etc), is that integrated in microeco, perhaps before (I guess) than cal_betadiv?
thanks!
PD: I think that found a way:

microecoITS_norm$cal_abund()

transformation here:

microecoITS_norm$taxa_abund$ASVs <- sqrt(microecoITS_norm$taxa_abund$ASVs)

microecoITS_norm$cal_betadiv(unifrac = FALSE)

beta_ITS <- trans_beta$new(dataset = microecoITS_norm,
group = "Site",
measure = "bray")

[ChiLiubio]

Hi. It should be :

microecoITS_norm$otu_table <- sqrt(microecoITS_norm$otu_table)