AMOVA results/output interpretation

[L Tan]

Hi all,

I've been trying for a while now to find out specifically how to interpret these AMOVA results/output and was hoping someone here might be able to help clarify some things. Otherwise, if you are able to direct me to a more correct forum that would be much appreciated. I only have a basic understanding of coding in R and populations genetics, so apologies if these are common-knowledge questions.

In relation to the output below, my main questions are:

1. What, specifically, are the warning messages: In is.euclid(xdist) : Zero distance(s) and In is.euclid(distmat) : Zero distance(s)? (Highlighted aqua below)
2. Does "alter" stand for "alternative hypothesis" - and if so, what is this actually set at? (highlighted yellow below)
   
3. Is the p-value exact? Is there any way to see this at more decimal points to get a more exact p-value? (highlighted green below)

Any advice in answering the above would be much appreciated. Thanks in advance!

L

#====================================================================================

#OUTPUT

#====================================================================================

#Convert genind to dataframe

> df <- genind2df(ravenpop_DefvOth, sep="/")

> strata(ravenpop_DefvOth) <- df

> #AMOVA into object; will print automatically if you do not put into object

> amova.result <- poppr.amova(ravenpop_DefvOth, hier = ~pop, clonecorrect = FALSE, within = TRUE,

+                             dist = NULL, squared = TRUE, freq = TRUE,

+                             correction = "quasieuclid", sep = "_", filter = FALSE,

+                             threshold = 0, algorithm = "farthest_neighbor", threads = 1L,

+                             missing = "loci", cutoff = 0.10, quiet = FALSE,

+                             method = c("ade4", "pegas"), nperm = 0)

Found 1081 missing values.

3 loci contained missing values greater than 10%

Removing 3 loci: CoBr24, CoBr09, Ck1_B6G

Distance matrix is non-euclidean.

Using quasieuclid correction method. See ?quasieuclid for details.

Warning messages:

1: In is.euclid(xdist) : Zero distance(s)

2: In is.euclid(distmat) : Zero distance(s)

> #Print result

> amova.result

$call

ade4::amova(samples = xtab, distances = xdist, structures = xstruct)

$results

                           Df      Sum Sq  Mean Sq

Between pop                  1    7.057667 7.057667

Between samples Within pop 196 1216.040893 6.204290

Within samples             198 1079.245295 5.450734

Total                      395 2302.343855 5.828719

$componentsofcovariance

                                           Sigma           %

Variations  Between pop                0.01639835   0.2806057

Variations  Between samples Within pop 0.37677823   6.4473649

Variations  Within samples             5.45073382  93.2720294

Total variations                       5.84391039 100.0000000

$statphi

                         Phi

Phi-samples-total 0.067279706

Phi-samples-pop   0.064655075

Phi-pop-total     0.002806057

> #Test for p value

> amova.test <- randtest(amova.result)

>

> #Plot amova.test

> plot(amova.test)

>

> amova.test

class: krandtest lightkrandtest

Monte-Carlo tests

Call: randtest.amova(xtest = amova.result)

Number of tests:   3

Adjustment method for multiple comparisons:   none

Permutation number:   99

                        Test        Obs    Std.Obs   Alter Pvalue

1  Variations within samples 5.45073382 -3.7992722    less   0.01

2 Variations between samples 0.37677823  4.2037855 greater   0.01

3     Variations between pop 0.01639835  0.5285421 greater   0.23

[Zhian N. Kamvar]

What, specifically, are the warning messages: In is.euclid(xdist) : Zero distance(s) and In is.euclid(distmat) : Zero distance(s)? (Highlighted aqua below)

This means that some samples in your data had no differences between them. This is usually the case with clonal data, but can also happen when there is a lot of missing data.

Does "alter" stand for "alternative hypothesis" - and if so, what is this actually set at? (highlighted yellow below)

Yes. This means that the alternative hypothesis is that the observed value is less than or greater than random chance. You might want to try using method = "pegas" for an alternative presentation (which, honestly is a bit more user friendly: https://rdrr.io/cran/pegas/man/amova.html)

Is the p-value exact? Is there any way to see this at more decimal points to get a more exact p-value? (highlighted green below)

Yes. This represents the fraction of permutations + 1 that rejected the null. To get more decimal places, increase the number of repetitions.

I hope that helps.

Best,

Zhian

--
You received this message because you are subscribed to the Google Groups "poppr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to poppr+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/poppr/e42d8cf4-06f4-4495-bdb1-82bcfccfeb72o%40googlegroups.com.

[L Tan]

Thanks Zhian for your answers! Very much appreciated, that helps me immensely. I'll take a look at using the "pegas" method too.

Kind regards,

L

To unsubscribe from this group and stop receiving emails from it, send an email to po...@googlegroups.com.