inst/examples/labrid_data_knit_.md

In cboettig/wrightscape: wrightscape

labrid data processing

ro fig.width=7.5 or

``` {r } require(phytools) require(geiger) require(wrightscape)

(This data has not been released)
``` {r }
labrid_tree <- read.nexus("../../data/labrid_tree.nex")
#fin_data <-read.csv(paste(path,"labrid.csv", sep="")) # not actually being used
diet_data <- read.csv("../../data/labriddata_parrotfish.csv")

Simple size corrections length and weight as fraction of body mass could use

``` {r eval=FALSE} for(i in c(3,4,6,7,8)){ diet_data[i] <- diet_data[i]/diet_data[5] } diet_data[5] <- log(diet_data[5])

We will take a different approach, using the phylogenetic correction, so the above code is not to be run.  

``` {r }
corrected_data <- diet_data

Use the simple trait names from @Price2010 ``` {r } traitnames <- c("Species", "group", "gape", "prot", "bodymass", "AM", "SH", "LP", "close", "open", "kt") names(corrected_data) <- traitnames

Lengths are log transformed 

``` {r }
corrected_data[["gape"]] <- log(corrected_data[["gape"]])
corrected_data[["prot"]] <- log(corrected_data[["prot"]])

masses are log(cube-root) transformed

``` {r } corrected_data[["bodymass"]] <- log(corrected_data[["bodymass"]])/3 corrected_data[["AM"]] <- log(corrected_data[["AM"]])/3 corrected_data[["SH"]] <- log(corrected_data[["SH"]])/3 corrected_data[["LP"]] <- log(corrected_data[["LP"]])/3

Traits which are ratios are fine as they are

Drop any unmatched tip-traits
``` {r }
ape <- treedata(labrid_tree, corrected_data[,3:11], corrected_data[,1])

Run Revell's phylogenetic size corrections

``` {r } ape$data["bodysize"] out <- phyl.resid(ape$phy, ape$data[,"bodymass"], ape$data[,c("gape", "prot","AM", "SH", "LP")] )

The `phyl.resid` function changes order of species listing. Merge for a set of uncorrected and corrected traits.  

``` {r } 
traits <- merge(ape$data, out$resid, by="row.names")

columns that are transformed now have gape.x for untransformed, gape.y for transformed.

format_data() gets regimes from column specified in "regimes" (e.g. this is a column id, not a # of regimes) This also converts the tree and data into ouch format (We could just hand it all traits, but these are just the tranformed and size-corrected ones)

``` {r } labrid <- format_data(labrid_tree, traits[,2:length(traits)], species_names=traits[,1])

## Painting Regimes
Select common ancestor of a Chlorurus and a Hipposcarus as the changepoint

``` {r }
intra_ancestor <- mrcaOUCH(c("Chlorurus_sordidus", "Hipposcarus_longiceps"), labrid$tree)
intramandibular <- paintBranches(intra_ancestor, labrid$tree, c("other","intramandibular"))

Select common ancestor for all parrot fish:

``` {r } pharyngeal_ancestor <- mrcaOUCH(c("Bolbometopon_muricatum", "Sparisoma_radians"), labrid$tree) pharyngeal <- paintBranches(pharyngeal_ancestor, labrid$tree, c("other","pharyngeal")) two_shifts <- paintBranches(c(pharyngeal_ancestor, intra_ancestor), labrid$tree, c("wrasses", "pharyngeal", "intramandibular") )

This leaves the branch on which the second transition occurs unspecified (fourth regime).  
We have to fix this manually

``` {r }
two_shifts[as.numeric(intra_ancestor)] <- "intramandibular"
two_shifts <- as.factor(as.character(two_shifts))
names(two_shifts) <- names(intramandibular)

rename the ouch-formated data ``` {r } dat <- labrid$data tree <- labrid$tree

rename the ape-formatted data

``` {r }
ape.phy <- ape$phy
ape.dat <- traits

Save the final output. (This data is provided with the package).

``` {r } save(list=c("intramandibular", "pharyngeal", "two_shifts", "tree", "dat", "ape.phy", "ape.dat"), file="labrids.rda")

We now have access to the following configurations:
`intramandibular`, `pharyngeal` and `two_shifts` paintings, (and `labrid$noregimes`),
and tree in `labrid$tree`

``` {r }
plot(labrid$tree, regimes=two_shifts)

cboettig/wrightscape documentation built on May 13, 2019, 2:12 p.m.