When I studied the microbial diversity, I wanted to choose a good method for the relationships between microbial diversity and environmental factors. So I found some relatives. I thought db-RDA and CAP are good methods for my study, but I don't know the difference between them. Can you give me some advices. Thanks!
Have a look on the article from Ramette 2007 which should help you to understand more about db-RDA and CAP which both correspond overall to a RDA and CA respectively. Below are citations from this article about db-RDA:
db-RDA: “When the data set consists of a matrix of distances between objects, distance-based RDA (db-RDA; Legendre & Anderson, 1999) can be applied to determine how well additional environmental parameters can explain the variation among objects in the matrix. The technique first applies a PCoA on the distance matrix to convert it back to a rectangular table containing rows of objects by columns of PCoA coordinates. Those new, uncorrelated coordinates thus correspond to synthetic ‘species’ variables that are then related to additional environmental parameters using a
classical RDA. For instance, db-RDA was successfully used to determine how the variation in matrices of genomic distances among environmental strains could be explained by factors such as soil parameters, host plant species, and spatial scale, each factor being taken alone or in combination”
CAP: is a canonical analysis on the principal coordinates for any resemblance matrix, including a permutation test. CAP takes into consideration the structure of the data. So, it is more likely to separate your different levels if there is no strong difference and is good to show the interaction between factors.
db-RDA “can be applied to determine how well additional environmental parameters can explain the variation among objects in the matrix”. In contrast, CAP do not try to increase the explanation of the variance with other parameters, but only perform correlations. I think this is the major difference. Overall, the choice of the “best” method depends on what you want to do and show, and there is not necessarily a unique answer but better to try different options. See this RG question
https://www.researchgate.net/post/How_to_choose_ordination_method_such_as_PCA_CA_PCoA_and_NMDS
Reference:
A. Ramette (2007) Multivariate analyses in microbial ecology, FEMS Microbiology Ecology, 62, 142-160
Dear Aimeric,
Thank you so much! I really appreciate your help and direction to references.
Best wishes,
Kate
Dear Aimeric,
Thank you for your good answer. That is very helpful.
Best wishes.
Hello! I am working on validation of psychological instruments, and usually in this field we use PCE and PCA in order to know the structure and dimensions of the scale. Lately I learnt about db-RDA, and it seems really similar to PCA, do you think I could use db-RDA instead PCA?
Do you know which are the db-RDA foundations?
Thanks lot!
This is all very helpful, thank you so much! While doing CAP and PCoA on my data, I noticed that CAP gives way better separation than PCoA, however; CAP only works with factors with more than two levels (e.g., Low, Medium, High but not with Low and High alone). Has anyone come across this issue or is there anything that I should be aware of?
Altaf Hussain CAP should provide better separation as it is designed to define axes that maximize separation among predefined groups whereas PCoA is designed just to define axes that capture the most overall variation (Look up LDA vs PCA which are the euclidean equivalents of CAP and PCoA respectively). CAP should "work" with just 2 groups, but with g groups CAP can only define up to g-1 axes to separate the groups. Therefore, with only 2 groups, you can only define 1 axis for plotting the separation which can make plots look weird because typically you want to plot in 2 dimensions for visualization.
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology