PERMANOVA/ADONIS result

[Wossata (Wossen Mengesha)]

Hi Guys, my beta diversity significant is highly significant when i run via galaxy (https://cancer.usegalaxy.org/) but the same data when I import to r and do it by vegan, none of the factors had significant contribution to beta diversity. Besides, when I view the result of galaxy via qiime2view, it is devoid of any information such as number of groups, test statistics or group significance box plots (that were illustrated in one of the free online qiime2 workshops that I am referring on YouTube. So, Shall I go ahead and report significance from Adonis via galaxy or just r (insignificance table)?? Please help. Besides, the PCoA generated using core-metrics-weighted-unifrac was very different in separation of clusters than the pcoa I generated via r (qiime2 via galaxy gave me better separation). I must admit sample size is smaller (18) and is 16S pacBio (if that meant anything). I know cleaver people on here to rescue me cheers

[colinbrislawn (Colin J Brislawn)]

Hello Wossen,

Welcome back to the forums!

If you are able, can you post the output artifacts you get from Qiime2 running on Galaxy so we can take a look!

@Wossata

Besides, when I view the result of galaxy via qiime2view, it is devoid of any information such as number of groups, test statistics or group significance box plots

This makes me there there's something else going wrong.

[Wossata (Wossen Mengesha)]

thanks @colinbrislawn. I am putting here the screenshots first from galaxy and then r outputs:

r outputs
adonis2(formula = dist_matrix ~ treatment + bay, data = final_metadata, permutations = 999)
Df SumOfSqs R2 F Pr(>F)
Model 3 0.036644 0.16246 0.9052 0.561
Residual 14 0.188907 0.83754
Total 17 0.225550 1.00000
(for all factors tested individually in r, (e.g. "treatment" and "bay" seen in galaxy) significance is the same (p> 0.5)
PcoA
1)QIIME2

P.S (Legend: Blue-unfumigated soil, Red-fumigated soil
PcoA from r

Much appreciated

[colinbrislawn (Colin J Brislawn)]

Thanks!

How about this?

@colinbrislawn

If you are able, can you post the output artifacts you get from Qiime2 running on Galaxy

[Mehrbod_Estaki]

One common issue I've noticed with folks unfamiliar with vegan's adonis is that it assumes your distance matrix is sorted in the same order as your grouping vector in your metadata file. This is something that QIIME2 handles nicely under hood but it is not controlled in R. Worth a double check there.