RDA: Redundancy Analysis
In walterxie/ComMA: Community Matrix Analysis

Description Usage Arguments Details Note Examples

Redundancy Analysis using vegan capscale.

doRDA(tcm.or.dist, env, verbose = TRUE)

matchCMEnv(cm, env, is.transposed = FALSE, verbose = TRUE)

plotCorrelations(df.numeric, corr.gram = FALSE, cex.axis = 0.75,
  cex.cor = 0.9, col = "#333333")

plotRDA(rda, env, colour.id = "Elevation",
  title = "Backward RDA, Jaccard distance", x.lab = "", y.lab = "",
  palette = c("blue", "orange"), scale.limits.min = NULL,
  no.legend = FALSE, legend.title = "Elevation (m)", verbose = TRUE, ...)

printXTable.RDA(rda, matrix.name = "", taxa.group = "", table.file = NULL,
  invalid.char = FALSE)

`tcm.or.dist`	A transposed community matrix or dist object of distances between samples. Rows are samples.
`env`	The enviornmental meta-data, where rows are samples, and they must be same as rownames(tcm.or.dist) inlcuding order. In addition, make sure rownames (enviornmental variables) are valid to `formula`.
`verbose`	More details. Default to TRUE.
`cm`	A community matrix.
`is.transposed`	If TRUE, then the community matrix is already transposed to be the valid input of `vegdist`. Default to FASLE.
`df.numeric`	The data frame or matrix containing numeric variables (columns) to plot.
`corr.gram`	Logical, if use `corrgram` instead of `plot`.
`rda`	The ordination result from `doRDA`.
`colour.id`	The column name in `env` to colour points and texts.
`title, x.lab, y.lab`	Title, x, y label.
`palette`	Refer to `ggOptPalette` in gg1L. Default to c("blue", "orange").
`scale.limits.min`	Manually set the minimum data range of the colour scale, for example, in legend. The code set `limits` in `discrete_scale` to `c(scale.limits.min, max(df[,colour.id]))`.
`no.legend, legend.title`	Configure legend.
`matrix.name`	The string to locate the matrix from its file name. Only used for table name and label here.
`taxa.group`	The taxonomic group. Only used for table name and label here.
`table.file`	If NULL, then print the results to console, otherwise print them to the file. Default to NULL.
`rda`	The list of results from `doRDA`.

doRDA makes Constrained Analysis of Principal Coordinates for eDNA data sets given environmental variables.

matchCMEnv matches the sample names between community matrix and the enviornmental meta-data including the order, in order to provide the valid input to RDA analysis doRDA.

Preprecessing can be applied by preprocessCM and preprocessEnv.

plotCorrelations plots numeric variables (columns).

Tip: use "%<a-%" in pryr to save plots.

plotRDA plots a ordination result from doRDA. However, there are still two columns "Plot" and "shortIDs" in the intermediate data "sites" hard coded to be able to make shorter texts. sites$Plot <- gsub("-[A-Z]", "", rownames(sites)) and sites$shortIDs <- gsub("(CM30|CM31|Plot)", "", rownames(sites)) This is expecting to be improved in future.

printXTable.RDA prints xtable given rda results.

Make sure you are using the correct data type for both the community matrix and enviornmental meta-data. Run sapply(env, class) to check if the value should be discrete(character) or numeric. convertType will convert data frame columns to different type easily.

# 1. get the community matrix and enviornmental meta-data
cm <- getCommunityMatrix("16S", min2=T, by.plot=F)
env <- getEnvData(by.plot=F)

# 2. preprocessing
cm.prep <- preprocessCM(cm, rm.samples=c("CM30b51","CM30b58"), min.abund=5, mean.abund.thr=0.025) 
env.prep <- preprocessEnv(env, rm.samples=c("CM30b51","CM30b58"), log.var=c(14:20), sel.env.var=c(4,5,8,9,14:22))
sapply(env, class)
 
# 3. match samples and transpose cm
tcm.env <- matchCMEnv(cm.prep, env.prep)

# 4. RDA
rda <- doRDA(tcm.env$tcm, tcm.env$env)

# 5. result
rda.pl <- plotRDA(rda[["backward"]], tcm.env$env)
rda.pl$plot
printXTable.RDA(rda, matrix.name="16S", taxa.group="all")



# already transposed
tcm.env <- matchCMEnv(tcm, env, is.transposed=T)
# Note colSums(cm) are based on samples
tcm.env <- matchCMEnv(cm, env)

# before RDA
require(pryr)
p %<a-% plotCorrelations(tcm.env$env)

rda.pl <- plotRDA(rda.list[[1]][["backward"]], env.prep, scale.limits.min=0)

printXTable.RDA(rda, matrix.name="16S", taxa.group="BACTERIA")