bomeara / phrapl

<img src="https://cloud.githubusercontent.com/assets/7992349/15196013/661bb536-1798-11e6-964f-100ae6aa1b7a.jpg" alt "phraplLogo" width= "600", height="200">

PHRAPL (Phylogeography using approximate likelihoods) is a method for phylogeographic model selection. This method can estimate the probability of observing a set of gene trees across a set of models that include population genetic parameters, e.g., population coalescence times and gene flow, in order to infer the model(s) in the set that most likely gave rise to the observed data.

Note that PHRAPL has recently moved from R-forge to Github for development. Moving to Github allows users to install pre-CRAN versions using devtools. To do this, within R, type

install.packages("devtools")
devtools::install_github("bomeara/phrapl")

phrapl's Issues

Convert ms string to phrapl model

Generate a set of random models

Include parameter values in 3d print

Error at subsampling step

Hi,

I'm trying to run PHRAPL on my dataset. I have set the popAssignments vector as follows:
popAssignments<-list(c(3,3,3))

I get following error while trying to generate subsampled trees.

Error in retainedTaxaMatrix[rep, ] <- keepTaxa :
number of items to replace is not a multiple of replacement length

Can you please explain this error?

Estimate weight of each gene favoring a gene flow vs migration model

Already have info about lik for each gene for each model; extract so we can figure out the weight. Which genes are "migration genes" (similar to "speciation genes").

Read in structure formatted file for SNPs

long vectors are not supported in .C

Hello,

I tried subsampling a set of 6.5K gene trees, each with a few tips per species. Subsampling works, but I cannot calculate the topology weights:

clyWeights.df<-GetPermutationWeightsAcrossSubsamples(popAssignments=clyAssign,observedTrees=clySubsampling2)

Error in perms(length(assignFrame[, 2][which(assignFrame[, 1] == popLabels[i])])) :
long vectors (argument 1) are not supported in .C

Is it possible to run it without subsampling? I tried GridSearch with doWeights=FALSE, but it fails...

Error in append(startingVectorList, list(collapseStarts)) :
object 'collapseStarts' not found

Talk to cipres about phrapl instance

Installation issues

Hi!

I have been trying to install PHRAPL and I couldn't do it yet.

First, I have tried in my Windows personal computer, but two packages seem to be unsuitable for my versions of R: "P2C2M" and "partitions". After changing my R version four times, I found a version that accept the installation of "partitions", but I can't find one that accept "P2C2M".
One of those versions gave me an error message saying that this package is avaiable only for Linux.

So I tried to install PHRAPL in a linux cluster that I have access with Bioconda.

But I got another type of error:

Downloading GitHub repo bomeara/phrapl@master
sh: 1: sh: 1: /bin/gtar: not found
/usr/bin/gzip: not found
sh: 1: sh: 1: /bin/gtar: not found
/usr/bin/gzip: not found
Error: Failed to install 'phrapl' from GitHub:
error in running command
In addition: Warning messages:
1: In system(paste(TAR, "--version >", tf, "2>&1")) :
error in running command
2: In system(cmd) : error in running command
3: In utils::untar(tarfile, ...) :
‘/usr/bin/gzip -dc '/tmp/Rtmp7uPNJK/file1f17521ee0c.tar.gz' | /bin/gtar -xf '-' -C '/tmp/Rtmp7uPNJK/remotes1f16ffc295b'’ returned error code 127
4: In system(paste(TAR, "--version >", tf, "2>&1")) :
error in running command

Do you know what type of error is it?

Is there any special way to install PHRAPL in Windows or linux?

Thank you in advance

Ariadne F. Sabbag
UNESP, Brazil

Add param values to PlotModel2D

Need to allow parameter estimates to affect arrow width, collapse heights, etc.

R install failure

Hi Brian,

I'm using OSX 10.11.16, R v3.2.1, and I am having an issue with the PHRAPL install. When I follow your instructions, I get an error apparently due to failure to unpack a file with a name ending in "-1":

> devtools::install_github("bomeara/phrapl")
Downloading GitHub repo bomeara/phrapl@master
from URL https://api.github.com/repos/bomeara/phrapl/zipball/master
Installing phrapl
trying URL 'http://cran.cnr.Berkeley.edu/bin/macosx/mavericks/contrib/3.2/binom_1.1-1'
Content type 'application/x-gzip' length 331405 bytes (323 KB)
==================================================
downloaded 323 KB

Error: Don't know how to decompress files with extension 1-1
> 
> 

In response, I downloaded the master dir and tried to use devtools to install locally, but this also failed:

> install("/Users/justinbagley/Downloads/bomeara-phrapl-688acb1")
Installing phrapl
trying URL 'http://cran.cnr.Berkeley.edu/bin/macosx/mavericks/contrib/3.2/binom_1.1-1'
Content type 'application/x-gzip' length 331405 bytes (323 KB)
==================================================
downloaded 323 KB

Error: Don't know how to decompress files with extension 1-1
> 
>

Please help me find a solution to this so I can start working with PHRAPL.
Thanks.

Best,

Justin

Function to get confidence from AIC grid

Parse the AIC grid to get the params say delta two log likelihood from the optimum to get a confidence estimate.

Error in PrepSubsampling

##Hi,
I'm getting the below error when running the PrepSubsampling step:

> observedTrees<-PrepSubsampling(
+     assignmentsGlobal=PopsA,      # the population assignments table
+     observedTrees=TreesA,           # the original trees
+     popAssignments=list(c(3,3,3)),        # the number of tips subsampled per population
+     subsamplesPerGene=subsamplesPerGene,  # the number of replicate subsamples to take per locus
+     outgroup=TRUE,                       # whether an outgroup is present in the dataset (TRUE or FALSE)
+     outgroupPrune=TRUE) 
Error in retainedTaxaMatrix[rep, ] <- keepTaxa : 
  number of items to replace is not a multiple of replacement length

I have 482 trees from UCE/sequence-capture data. I did midpoint root my trees to double check that they were all rooted using TreesA <-midpoint(TreesA) after I initially received the error, but that didn't appear to change anything.

One of my populations has 9 individuals in it, but I still received an error even when moving popAssignments for that population to 1, just in case there were not enough individuals to subsample.
Thanks for any suggestions!

Image for README

Empty migrationArray

Dear Brian,

I am trying to generate a set of models from a fixed topology, but my migrationArray returns an empty list. Here is the code:

library(ape)
popAssignments<-list(c(3,3,3,3,3))
library(phrapl)
library(rPython)
library(P2C2M)

#Fix the tree from stacey

pop1 <- c(0,0,0,4)
pop2 <- c(0,0,3,NA)
pop3 <- c(0,2,NA,NA)
pop4 <- c(1,NA,NA,NA)
pop5 <- c(1,2,3,4)
collapseList <- list(pop1,pop2,pop3,pop4,pop5)


popVector<-popAssignments[[1]]  
maxK<-5  
maxMigrationK=0 
maxN0K=1
maxGrowthK=1
forceTree=T  
forceSymmetricalMigration=T  
migrationArray<-GenerateMigrationIndividuals(popVector=popVector,maxK=maxK, collapseList = collapseList,
                                             maxMigrationK=maxMigrationK,maxN0K=maxN0K,maxGrowthK=maxGrowthK,
                                             forceTree=forceTree,forceSymmetricalMigration=forceSymmetricalMigration,
                                             verbose = T)

Am I doing anything wrong?

Thanks,

Pedro

Deal with dangly stem plotting properly in models with merge at base

Allow specification of tip label font and color in plot2d

Make clickable R GUI for building models

Click on rootward end of edges to add a collapse, midpoint of edges to add a migration event, etc. Prototype code was in an email, need to merge with the PlotModel2D

Docker or virtual box instance of phrapl

Get rid of library calls within main code

All packages should be in DESCRIPTION & NAMESPACE files. Note this isn't a simple matter of just deleting library lines; need to verify that the code works correctly after this change (unit testing would come to the rescue in this case...).