iCAMP (version 1.5.12)

bNRI.bin.big: Calculate beta net relatedness index (betaNRI) for each phylogenetic bin

Description

Perform null model test based on a phylogenetic beta diversity index, beta mean pairwise distance (betaMPD), in each bin; calculate beta net relatedness index (betaNRI. Webb et al 2008), or modified Raup-Crick metric, or confidence level based on the comparison between observed and null betaMPD in each bin. The package bigmemory (Kane et al 2013) is used to handle very large phylogenetic distance matrix.

Usage

bNRI.bin.big(comm, pd.desc, pd.spname, pd.wd, pdid.bin, sp.bin,
             spname.check = FALSE, nworker = 4, memo.size.GB = 50,
             weighted = c(TRUE, FALSE), rand = 1000, output.bMPD = FALSE,
             sig.index=c("SES","Confidence","RC","bNRI"),
             unit.sum = NULL, correct.special = FALSE,
             detail.null=FALSE, special.method=c("MPD","MNTD","both"),
             ses.cut=1.96,rc.cut=0.95, conf.cut=0.975,
             dirichlet = FALSE)

Arguments

comm

matrix or data.frame, community data, each row is a sample or site, each colname is a species or OTU or gene, thus rownames should be sample IDs, colnames should be taxa IDs.

pd.desc

the name of the file to hold the backingfile description of the phylogenetic distance matrix, it is usually "pd.desc" if using default setting in pdist.big function.

pd.spname

character vector, taxa id in the same rank as the big matrix of phylogenetic distances.

pd.wd

folder path, where the bigmemmory file of the phylogenetic distance matrix are saved.

pdid.bin

list, each element is a vector of integer, indicating which rows/columns in the big phylogenetic matrix represent the taxa in a bin.

sp.bin

one-column matrix, rownames are taxa IDs (i.e. OTU IDs), the only column shows the bin ID of each taxon. Bin IDs are integers in the same order as the elements in the list of pdid.bin.

spname.check

logic, whether to check the OTU ids (species names) in community matrix and phylogenetic distance matrix are the same.

nworker

for parallel computing. Either a character vector of host names on which to run the worker copies of R, or a positive integer (in which case that number of copies is run on localhost). default is 4, means 4 threads will be run.

memo.size.GB

numeric, to set the memory size as you need, so that calculation of large tree will not be limited by physical memory. unit is Gb. default is 50Gb.

weighted

Logic, consider abundances or not (just presence/absence). default is TRUE.

rand

integer, randomization times. default is 1000.

output.bMPD

logic, if TRUE, the output will include beta mean pairwise distance (betaMPD).

sig.index

character, the index for null model significance test. SES or bNRI, standard effect size, i.e. beta net relatedness index (betaNRI); Confidence, percentage of null values less extreme than the observed value, i.e. non-parametric one-side confidence level; RC, modified Raup-Crick index (RC) based on betaMPD, i.e. count the number of null betaMPD lower than observed betaMPD plus a half of the number of null betaMPD equal to observed betaMPD, to get alpha, then calculate betaMPD-based RC as (2 x alpha - 1). default is SES. If input a vector, only the first element will be used.

unit.sum

NULL or a number or a nemeric vector. When a beta diversity index is calculated for a bin, the taxa abundances will be divided by unit.sum to calculate the relative abundances. Usually, unit.sum can be set as the sequencing depth in each sample. Default setting is NULL, means not to do this transformation.

correct.special

logic, whether to correct the special cases. Default is FALSE.

detail.null

logic, if TRUE, the output will include all the null values. Default is FALSE.

special.method

When correct.special is TRUE, which method will be used to check underestimation of deterministic pattern(s) in special cases. MPD, use null model test based on mean pairwise distance; MNTD, use null model test of mean nearest taxon distance; both, use null model test of both MPD and MNTD. Default is MPD.

ses.cut

numeric, the cutoff of significant standard effect size, default is 1.96.

rc.cut

numeric, the cutoff of significant modified Raup-Crick metric, default is 0.95.

conf.cut

numeric, the cutoff of significant one-side confidence level, default is 0.975.

dirichlet

Logic. If TRUE, the taxonomic null model for correcting special cases will use Dirichlet distribution to generate relative abundances in randomized community matrix. default is FALSE.

Value

Output is a list with following elements:

index

list, each element is a square matrix of betaNRI (or RC or Confidence based on betaMPD) values of a bin. The elements (bins) are in the same order as in the input pdid.bin.

sig.index

character, indicates the index for null model significance test, SES (i.e. betaNRI), RC, or Confidence.

betaMPD.obs

Output only if output.bMPD is TRUE. A list, each element is a square matrix of observed beta MPD values of a bin. The elements (bins) are in the same order as in the input pdid.bin.

rand

Output only if detail.null is TRUE. A list, each element is a matrix with null values of beta MPD for each turnover of a bin. The elements (bins) are in the same order as in the input pdid.bin.

special.crct

Output only if detail.null is TRUE. NULL if correct.special is FALSE. A list with three elements, corresponding to three different null model significance testing indexes, i.e. SES, RC, and Confidence. Each element is a matrix, where the value is zero if the result for a turnover of a bin does not need to correct, otherwise there will be a corrected value.

Details

The beta net relatedness index (betaNRI; Webb et al. 2008, Stegen et al 2012) is calculated for each phylogenetic bin. betaNRI is a standardized measure of the mean pairwise distance between samples/communities (betaMPD). Parallel computing is used to improve the speed.

The null model algorithm is "taxa shuffle" (Kembel 2009), i.e. shuffling taxa labels across the tips of the phylogenetic tree to randomize phylogenetic relationships among species. In this function, taxa will be randomized across all bins.

In the betaNRI of each bin, the diagonal are set as zero. If the randomized results are all the same, the standard deviation will be zero and betaNRI will be NAN. In this case, betaNRI will be set as zero, since the observed result is not differentiable from randomized results.

Modified RC (Chase et al 2011) and Confidence (Ning et al 2020) are alternative significance test indexes to evaluate how the observed beta diversity index deviates from null expectation, which could be a better metric than standardized effect size (betaNRI) in some cases, e.g. null values do not follow normal distribution.

References

Webb, C.O., Ackerly, D.D. & Kembel, S.W. (2008). Phylocom: software for the analysis of phylogenetic community structure and trait evolution. Bioinformatics, 24, 2098-2100.

Kembel, S.W. (2009). Disentangling niche and neutral influences on community assembly: assessing the performance of community phylogenetic structure tests. Ecol Lett, 12, 949-960.

Stegen, J.C., Lin, X., Konopka, A.E. & Fredrickson, J.K. (2012). Stochastic and deterministic assembly processes in subsurface microbial communities. Isme Journal, 6, 1653-1664.

Chase, J.M., Kraft, N.J.B., Smith, K.G., Vellend, M. & Inouye, B.D. (2011). Using null models to disentangle variation in community dissimilarity from variation in alpha-diversity. Ecosphere, 2, 1-11.

Ning, D., Yuan, M., Wu, L., Zhang, Y., Guo, X., Zhou, X. et al. (2020). A quantitative framework reveals ecological drivers of grassland microbial community assembly in response to warming. Nature Communications, 11, 4717.

Kane, M.J., Emerson, J., Weston, S. (2013). Scalable Strategies for Computing with Massive Data. Journal of Statistical Software, 55(14), 1-19. URL http://www.jstatsoft.org/v55/i14/.

See Also

bNRIn.p,bmpd

Examples

# NOT RUN {
# this function is usually used in icamp.big when setting phylo.rand.scale="across",
# means randomization across all bins in phylogenetic null model.
data("example.data")
comm=example.data$comm
tree=example.data$tree
pdid.bin=example.data$pdid.bin
sp.bin=example.data$sp.bin

# since pdist.big need to save output to a certain folder,
# the following code is set as 'not test'.
# but you may test the example on your computer after change the path for 'save.wd'.

# }
# NOT RUN {
wd0=getwd()
save.wd=paste0(tempdir(),"/pdbig.bNRI.bin.big")
# you may change save.wd to the folder you want to save the pd.big output.
nworker=2 # parallel computing thread number
pd.big=pdist.big(tree = tree, wd=save.wd, nworker = nworker)
rand.time=20 # usually use 1000 for real data.

bNRIbins=bNRI.bin.big(comm=comm, pd.desc=pd.big$pd.file, pd.spname=pd.big$tip.label,
                      pd.wd=pd.big$pd.wd, pdid.bin=pdid.bin, sp.bin=sp.bin,
                      spname.check = FALSE, nworker = nworker, memo.size.GB = 50,
                      weighted = TRUE, rand = rand.time, output.bMPD = FALSE,
                      sig.index="SES",unit.sum = NULL, correct.special = TRUE,
                      detail.null=FALSE, special.method="MPD")
setwd(wd0)
# }