Hello, my data is from a functional DNA array and it looks like this :
phylum SCC1 SCC2 SCC3 SDC1 SDC2 SDC3 SAC1 SAC2 SAC3
Verrucomicrobia 0.525962728 0.453958735 0.50890293 0.392088402 0.545695164 0.471533122 0.910890972
1.221393907 0.786549277
Actinobacteria 0.565297666 0.31049045 0.553223779 0.462435947 0.510223144 0.392476548 1.266877979
0.327223749 0.898307454
Gammaproteobacteria 0.65658187 0.714752911 0.544786603 1.208918781 1.027978028 1.42647538 0.644828867 0.854848574 0.798517205
As showed in the first row, I have 9 samples from 3 niches. The numeric data is the expression ratio of each gene. I want to know if there is any significant difference between different sites. I have no trouble with carrying out ordination,clustering, ANOVA, and PERMANOVA analysis yet I don't know if I should go with ANOVA(univariate) or PERMANOVA(multivariate). Frankly speaking, I don't have any environmental factors data in hand.
I feel I should go with one-way PERMANOVA to compare 3 niches, as it's based on the distance matrix. However, I don't feel there's anything wrong with using ANOVA , as I can perform it on all samples and I don't really have multiple (environmental) variates.
Any idea is welcome. Thanks
Hey, i dont see any problem. In this case it is not a problem the lack of environmental data. You can definitely use PERMANOVA as univariate measure to compare 3 niches. Just remember to use Euclidean distances as recomended by Anderson et al.
Cheers,
Cesar
PERMANOVA is a better alternative when your data don't meet normality or homocedasticity. It is also the best option if you don't have many replicates, because you get more power with the permutations.
Best of luck.
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology