grunwaldlab / poppr Goto Github PK

🌶 An R package for genetic analysis of populations with mixed (clonal/sexual) reproduction

Home Page: https://grunwaldlab.github.io/poppr

Makefile 0.13% R 79.81% C 16.22% Shell 0.09% TeX 3.72% CSS 0.03%

r multilocus-genotypes genetic-distances population-genetics multilocus-lineages genetic-analysis populations minimum-spanning-networks clonality

poppr's Issues

Fix definition of Hexp

Currently, Hexp, produced by the function poppr is cited as Nei's expected heterozygosity (Nei 1978). Equation 2 in this paper translates to the implementation here if you consider x to be a genotype instead of an allele. How this came to be is a mystery. This is actually an unbiased measure of genotypic diversity representing the probability of drawing two genotypes from a population and having them be different. It's implementation in the function locus_table is accurate.

Tasks:

Change citation
Add average expected heterozygosity (equation 3 in Nei 1978) to poppr output
Fix equation in "algo" reference manual.

Ensure that bitwise.IA works for haploids

poppr breaking with new adegenet on CRAN

CRAN's maintainer has reported the following issue caused by the new version of adegenet on CRAN:

** preparing package for lazy loading
Error in setIs(Class, class2, classDef = classDef, where = where) :
  class “genclone” cannot extend class “genind”: slot in class “genclone” must extend corresponding slot in class “genind”: fails for hierarchy
Error in setClass("genclone", contains = "genind", representation = representation(mlg = "numeric",  :
  error in contained classes ("genind") for class “genclone”; class definition removed from ‘poppr’
Error : unable to load R code in package ‘poppr’
ERROR: lazy loading failed for package ‘poppr’

Fix logical error in infinite model of Bruvo's Distance

in This line, zerocatch[0] < 0 should be changed to zerocatch[0] == 0. A test should be added to make sure bug is permanently squashed.

Add v.2.0.1 once CRAN accepts patch

Before then, all changes will still be under 2.0.0.99

amova does not work on single locus

fix diss.dist to allow for single locus (for within = FALSE)
fix pool_haplotypes to ignore missing data (for within = TRUE)

Cutoff not working for non-dissimilarity distance in imsn()

Set the distance to any other distance that "Dissimilarity" and nothing happens when you specify a cutoff.

use stored function/object name as a memory for mlg.filter

Currently, mlg.filter has a default behavior of using nei.dist to calculate the distance matrix. If the user utilizes mlg.filter as an assignment method, currently @distname slot in the object carries the name of the object or the function and the @distargs slot contains the arguments for that function.

In this way there are a couple of procedures that could happen for this.

check if @distname is a function or a distance matrix:

is.function(eval(x@mlg@distname))
is.dist(eval(x@mlg@distname))

if @distname is a function, use do.call

do.call(eval(x@mlg@distname), c(x, x@mlg@distargs))

Add support for custom MLG definitions

Create a slot that allows users to customize MLG definitions such that they would be able to increase or decrease MLG diversity based on their knowledge of the system.

Include tests to be skipped on CRAN

The package testthat has a function called skip_on_cran() that will skip tests on CRAN to save run time. This will allow the construction of a bare set of tests to make sure the package works and an expanded set to increase code coverage to test for caveats.

implement pgen and psex

The current implementation is internal and slightly broken. It would be good to implement this even if it's only in R.

This will be closed when the following tasks are complete:

function to estimate round-robin allele frequencies is in place
pgen works for genind objects
code is documented
tests are written
Stacy Krueger-Hadfield and Erik Sotka are added as contributors on the functions and DESCRIPTION
add NEWS
add functions to package man page
add functions to main vignette

add skip_on_cran() to test-filter.R

I got a lovely email from BDR today:

On Jul 18, 2015, at 03:45 , Prof Brian Ripley [email protected] wrote:

See https://cran.r-project.org/web/checks/check_results_poppr.html . There are several separate errors, but for fedora-gcc it seems you are testing floating-point equality which the manual warns you not to do.

Please fix ASAP.

Brian D. Ripley, [email protected]
Emeritus Professor of Applied Statistics, University of Oxford
1 South Parks Road, Oxford OX1 3TG, UK

Because I'm already testing on coveralls, I think I can safely add skip_on_cran() to that particular test. Since I will also be uploading the version that fixes #47, I will increase the version number hopefully the last time.

Ensure that bitwise.IA works for diploids

add backwards compatibility for hierarchy methods in version 2.0

Since the hierarchy methods were moved to adegenet 2.0, all of them have changed names slightly to account for the preferred style (camelCase in adegenet). To make the transition smoother, the hierarchy functions from poppr should be kept, but should only serve as wrappers to the methods in adegenet (note that both the assignment and functional methods need to have replacers):

poppr version	adegenet version
sethierarchy	strata
gethierarchy	strata
splithierarchy	splitStrata
addhierarchy	addStrata
namehierarchy	nameStrata

Integrate bootstrap confidence intervals for diversity stats

Introduction

The poppr command returns a table that contains diversity summary statistics and linkage summary statistics. Currently, the only summary statistics that give a level of confidence are rarefaction (eMLG) and linkage (Ia and rbarD). It only makes sense to provide confidence intervals or standard errors for the other diversity statistics present in the diversity table.

Current implementation

A working example can be found in this bit of code. It performs bootstrapping using the boot package and returns confidence intervals. Since it uses the boot package, it has inherent multicore functionality.

Tasks

integrate bootstrapping for summary statistics
allow users to select confidence intervals or standard errors
add option to keep or remove analyses from the table
document function

Explicitly document model choice options in Bruvo's distance

Bruvo's distance currently has the options to turn on and off the genome addition and genome loss models, but does not explicitly state what happens when both are on and both are off. This should be documented in all bruvo functions as well as the Algorithms and Equations vignette.

Update ramorum citation

It has been officially published and gets page numbers.

Single locus genalex files cannot be imported

This post in the poppr forum showed that a single locus data set could not be imported. Here's a reproducible example.

# Note, this is a subset of the monpop data set, which is haploid. I'm treating it as a diploid here
zz <- "1    6   1   6
7_09_BB         
Ind Pop CHMFc4  CHMFc5
A004    7_09_BB 224 85
A002    7_09_BB 224 97
A011    7_09_BB 224 97
A009    7_09_BB 224 97
A006    7_09_BB 224 97
A013    7_09_BB 224 97"

read.genalex(textConnection(z), sep = "\t")

 Error in gena2[, ((x - 1)/ploidy) + 1] <<- pop_combiner(gena, hier = x:(x +  : 
  incorrect number of subscripts on matrix

The reason for this error was because of our old friend drop = TRUE being default for subsetting matrices. I am scouring the code and setting drop = FALSE everywhere.

Add 2.0-rc tag after merge

This will allow compatibility for the link in the frontiers paper.

merge 2.0-rc into master and devel
remove 2.0-rc branch
change README and NEWS
tag as 2.0-rc so that link in frontiers paper still works and make pre-release.
manually change version and submit to CRAN
tag version once it's accepted on CRAN.

Capitalize bold fields in imsn()

The fields have inconsistent capitalization. This should be fixed, but it's not terribly urgent.

add function to recalcuate MLGs

If the user has a genclone object, the MLGs are static from when the data is read in to allow for identification downstream. There is no obvious way to reset those MLGs currently. It would be good to provide a small function that would do this.

Allow text connections for read.genalex

read.genalex currently reads in a file from disc formatted as a genalex csv. Unfortunately, it does not work well on connection objects. To alleviate this, the argument skip = 2 on this line in read.genalex should be changed to skip = ifelse("connection" %in% class(genalex), 0, 2), which would allow connections to be read in thusly:

library(poppr)
data(partial_clone)
pc <- file()
genind2genalex(partial_clone, filename = pc)
partial_clone2 <- read.genalex(pc)
close(pc)

This would make it easier to create reproducible examples.

Add function to display results of all mlg.filter algorithms

This would be similar to these functions from https://github.com/grunwaldlab/supplementary-poppr-2.0.

Memory leak with valgrind

I haven't gotten an email yet, but there's a memtest note on CRAN. It's coming from bitwise.c, which makes me think it might be a threadding issue :(

==21510== Memcheck, a memory error detector
==21510== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
==21510== Using Valgrind-3.10.1 and LibVEX; rerun with -h for copyright info
==21510== Command: /data/blackswan/ripley/R/R-devel-vg/bin/exec/R -f test-all.R --restore --save --no-readline --vanilla
==21510== 

R Under development (unstable) (2015-07-03 r68625) -- "Unsuffered Consequences"
Copyright (C) 2015 The R Foundation for Statistical Computing
Platform: x86_64-unknown-linux-gnu (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

> library(testthat)
> test_check("poppr")
Loading required package: poppr
Loading required package: adegenet
Loading required package: ade4

   /// adegenet 2.0.0 is loaded ////////////

   > overview: '?adegenet'
   > tutorials/doc/questions: 'adegenetWeb()' 
   > bug reports/feature resquests: adegenetIssues()


This is poppr version 2.0.0. To get started, type package?poppr
==21510== Invalid read of size 4
==21510==    at 0x1BD52E0F: bitwise_distance_haploid._omp_fn.0 (packages/tests-vg/poppr/src/bitwise_distance.c:232)
==21510==    by 0x35D0809BBE: GOMP_parallel (in /usr/lib64/libgomp.so.1.0.0)
==21510==    by 0x1BD50F08: bitwise_distance_haploid (packages/tests-vg/poppr/src/bitwise_distance.c:175)
==21510==    by 0x479A92: do_dotcall (svn/R-devel/src/main/dotcode.c:1251)
==21510==    by 0x4B1C3E: Rf_eval (svn/R-devel/src/main/eval.c:655)
==21510==    by 0x4B4C35: do_set (svn/R-devel/src/main/eval.c:2106)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B36C8: do_begin (svn/R-devel/src/main/eval.c:1716)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B36C8: do_begin (svn/R-devel/src/main/eval.c:1716)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==  Address 0x1ce20844 is 772 bytes inside a block of size 1,968 alloc'd
==21510==    at 0x4A06BCF: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==21510==    by 0x4DE26D: GetNewPage (svn/R-devel/src/main/memory.c:865)
==21510==    by 0x4DFD71: Rf_allocVector3 (svn/R-devel/src/main/memory.c:2437)
==21510==    by 0x541032: Rf_allocVector (svn/R-devel/src/include/Rinlinedfuns.h:194)
==21510==    by 0x541032: HashLookup (svn/R-devel/src/main/unique.c:801)
==21510==    by 0x542C10: match5 (svn/R-devel/src/main/unique.c:902)
==21510==    by 0x4A3DFA: bcEval (svn/R-devel/src/main/eval.c:5524)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4AA831: bcEval (svn/R-devel/src/main/eval.c:5496)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4AA831: bcEval (svn/R-devel/src/main/eval.c:5496)
==21510== 
==21510== Thread 2:
==21510== Invalid read of size 4
==21510==    at 0x1BD52EE0: bitwise_distance_haploid._omp_fn.0 (packages/tests-vg/poppr/src/bitwise_distance.c:248)
==21510==    by 0x35D080DB65: ??? (in /usr/lib64/libgomp.so.1.0.0)
==21510==    by 0x35B5607529: start_thread (in /usr/lib64/libpthread-2.20.so)
==21510==    by 0x35B4F0022C: clone (in /usr/lib64/libc-2.20.so)
==21510==  Address 0x1ce20844 is 772 bytes inside a block of size 1,968 alloc'd
==21510==    at 0x4A06BCF: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==21510==    by 0x4DE26D: GetNewPage (svn/R-devel/src/main/memory.c:865)
==21510==    by 0x4DFD71: Rf_allocVector3 (svn/R-devel/src/main/memory.c:2437)
==21510==    by 0x541032: Rf_allocVector (svn/R-devel/src/include/Rinlinedfuns.h:194)
==21510==    by 0x541032: HashLookup (svn/R-devel/src/main/unique.c:801)
==21510==    by 0x542C10: match5 (svn/R-devel/src/main/unique.c:902)
==21510==    by 0x4A3DFA: bcEval (svn/R-devel/src/main/eval.c:5524)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4AA831: bcEval (svn/R-devel/src/main/eval.c:5496)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4AA831: bcEval (svn/R-devel/src/main/eval.c:5496)
==21510== 
==21510== Thread 1:
==21510== Invalid read of size 4
==21510==    at 0x1BD5331F: bitwise_distance_diploid._omp_fn.1 (packages/tests-vg/poppr/src/bitwise_distance.c:439)
==21510==    by 0x35D0809BBE: GOMP_parallel (in /usr/lib64/libgomp.so.1.0.0)
==21510==    by 0x1BD5125F: bitwise_distance_diploid (packages/tests-vg/poppr/src/bitwise_distance.c:374)
==21510==    by 0x479A92: do_dotcall (svn/R-devel/src/main/dotcode.c:1251)
==21510==    by 0x4B1C3E: Rf_eval (svn/R-devel/src/main/eval.c:655)
==21510==    by 0x4B4C35: do_set (svn/R-devel/src/main/eval.c:2106)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B36C8: do_begin (svn/R-devel/src/main/eval.c:1716)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B36C8: do_begin (svn/R-devel/src/main/eval.c:1716)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==  Address 0x1d27d254 is 724 bytes inside a block of size 1,968 alloc'd
==21510==    at 0x4A06BCF: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==21510==    by 0x4DE26D: GetNewPage (svn/R-devel/src/main/memory.c:865)
==21510==    by 0x4DFD71: Rf_allocVector3 (svn/R-devel/src/main/memory.c:2437)
==21510==    by 0x446671: Rf_allocVector (svn/R-devel/src/include/Rinlinedfuns.h:194)
==21510==    by 0x446671: do_nzchar (svn/R-devel/src/main/character.c:123)
==21510==    by 0x4A3DFA: bcEval (svn/R-devel/src/main/eval.c:5524)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4AA831: bcEval (svn/R-devel/src/main/eval.c:5496)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4AA831: bcEval (svn/R-devel/src/main/eval.c:5496)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510== 
==21510== Thread 2:
==21510== Invalid read of size 4
==21510==    at 0x1BD533F0: bitwise_distance_diploid._omp_fn.1 (packages/tests-vg/poppr/src/bitwise_distance.c:455)
==21510==    by 0x35D080DB65: ??? (in /usr/lib64/libgomp.so.1.0.0)
==21510==    by 0x35B5607529: start_thread (in /usr/lib64/libpthread-2.20.so)
==21510==    by 0x35B4F0022C: clone (in /usr/lib64/libc-2.20.so)
==21510==  Address 0x1d27d254 is 724 bytes inside a block of size 1,968 alloc'd
==21510==    at 0x4A06BCF: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==21510==    by 0x4DE26D: GetNewPage (svn/R-devel/src/main/memory.c:865)
==21510==    by 0x4DFD71: Rf_allocVector3 (svn/R-devel/src/main/memory.c:2437)
==21510==    by 0x446671: Rf_allocVector (svn/R-devel/src/include/Rinlinedfuns.h:194)
==21510==    by 0x446671: do_nzchar (svn/R-devel/src/main/character.c:123)
==21510==    by 0x4A3DFA: bcEval (svn/R-devel/src/main/eval.c:5524)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4AA831: bcEval (svn/R-devel/src/main/eval.c:5496)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4AA831: bcEval (svn/R-devel/src/main/eval.c:5496)
==21510==    by 0x4B178F: Rf_eval (svn/R-devel/src/main/eval.c:558)
==21510== 
Loading required package: parallel
testthat results ================================================================
OK: 221 SKIPPED: 52 FAILED: 0
> 
> proc.time()
   user  system elapsed 
385.387  78.932 391.750 
==21510== 
==21510== HEAP SUMMARY:
==21510==     in use at exit: 237,692,259 bytes in 121,847 blocks
==21510==   total heap usage: 814,885 allocs, 693,038 frees, 1,242,042,234 bytes allocated
==21510== 
==21510== Thread 1:
==21510== 18,352 bytes in 31 blocks are possibly lost in loss record 4,689 of 5,786
==21510==    at 0x4A08946: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==21510==    by 0x35B46129E4: _dl_allocate_tls (in /usr/lib64/ld-2.20.so)
==21510==    by 0x35B5607EC7: pthread_create@@GLIBC_2.2.5 (in /usr/lib64/libpthread-2.20.so)
==21510==    by 0x35D080E0B6: ??? (in /usr/lib64/libgomp.so.1.0.0)
==21510==    by 0x35D0809BB9: GOMP_parallel (in /usr/lib64/libgomp.so.1.0.0)
==21510==    by 0x1BD50F08: bitwise_distance_haploid (packages/tests-vg/poppr/src/bitwise_distance.c:175)
==21510==    by 0x479A92: do_dotcall (svn/R-devel/src/main/dotcode.c:1251)
==21510==    by 0x4B1C3E: Rf_eval (svn/R-devel/src/main/eval.c:655)
==21510==    by 0x4B4C35: do_set (svn/R-devel/src/main/eval.c:2106)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B36C8: do_begin (svn/R-devel/src/main/eval.c:1716)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510== 
==21510== 60,656 bytes in 3,791 blocks are definitely lost in loss record 5,303 of 5,786
==21510==    at 0x4A08946: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==21510==    by 0x4D9140: R_chk_calloc (svn/R-devel/src/main/memory.c:3186)
==21510==    by 0x1BD55344: bruvo_dist (packages/tests-vg/poppr/src/poppr_distance.c:425)
==21510==    by 0x1BD56050: bruvo_distance (packages/tests-vg/poppr/src/poppr_distance.c:296)
==21510==    by 0x479A92: do_dotcall (svn/R-devel/src/main/dotcode.c:1251)
==21510==    by 0x4B1C3E: Rf_eval (svn/R-devel/src/main/eval.c:655)
==21510==    by 0x4B4C35: do_set (svn/R-devel/src/main/eval.c:2106)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B36C8: do_begin (svn/R-devel/src/main/eval.c:1716)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4B18E8: Rf_eval (svn/R-devel/src/main/eval.c:674)
==21510== 
==21510== 68,128 (34,064 direct, 34,064 indirect) bytes in 2,129 blocks are definitely lost in loss record 5,347 of 5,786
==21510==    at 0x4A08946: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==21510==    by 0x4D9140: R_chk_calloc (svn/R-devel/src/main/memory.c:3186)
==21510==    by 0x1BD55356: bruvo_dist (packages/tests-vg/poppr/src/poppr_distance.c:432)
==21510==    by 0x1BD56050: bruvo_distance (packages/tests-vg/poppr/src/poppr_distance.c:296)
==21510==    by 0x479A92: do_dotcall (svn/R-devel/src/main/dotcode.c:1251)
==21510==    by 0x4B1C3E: Rf_eval (svn/R-devel/src/main/eval.c:655)
==21510==    by 0x4B4C35: do_set (svn/R-devel/src/main/eval.c:2106)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B36C8: do_begin (svn/R-devel/src/main/eval.c:1716)
==21510==    by 0x4B1AB0: Rf_eval (svn/R-devel/src/main/eval.c:627)
==21510==    by 0x4B6DBC: Rf_applyClosure (svn/R-devel/src/main/eval.c:1039)
==21510==    by 0x4B18E8: Rf_eval (svn/R-devel/src/main/eval.c:674)
==21510== 
==21510== LEAK SUMMARY:
==21510==    definitely lost: 94,720 bytes in 5,920 blocks
==21510==    indirectly lost: 34,064 bytes in 4,258 blocks
==21510==      possibly lost: 18,352 bytes in 31 blocks
==21510==    still reachable: 237,545,123 bytes in 111,638 blocks
==21510==         suppressed: 0 bytes in 0 blocks
==21510== Reachable blocks (those to which a pointer was found) are not shown.
==21510== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==21510== 
==21510== For counts of detected and suppressed errors, rerun with: -v
==21510== ERROR SUMMARY: 11 errors from 7 contexts (suppressed: 0 from 0)

make sure win.ia and samp.ia have bitwise.IA internally

win.ia already has this, but samp.ia does not. This should be remedied.

Pairwise AMOVA

GenAlex generates pairwise PhiPT values for any AMOVA analysis run with more than one between group comparison. We have many between group comparisons to make. Is there an easy way using Poppr to generate something similar to what GenAlex produces? Please add pairwise_amova

Thanks!
S

Add tests for haploid index of association for genlight objects

These need not be anything fancy. They can be similar to the current tests. I will write them.

fix_negative_branch internal function gives erroneous branch lengths

The internal poppr function fix_negative_branch returns erroneous branch lengths. This is due to using the %in% operator as opposed to the match function.

This bug has few consequences normally as it mainly affects clones whose branch lengths are effectively zero. It becomes a problem when a long branch is accidentally swapped with a short branch.

Text encoding errors

Several files have UTF-8 em-dashes and directional quotes.

From Dr. Ripley:

There are also issues when checking in a Latin-1 locale on Linux:

checking whether package 'poppr' can be installed ... [10s/11s] WARNING
Found the following significant warnings:
Warning: unable to re-encode 'amova.r' line 175
Warning: unable to re-encode 'distances.r' lines 158, 161, 167, 171, 178, 181, 187, 191, 197, 200
Warning: unable to re-encode 'internal.r' lines 96, 111, 112
Warning: unable to re-encode 'poppr.R' lines 65, 67, 68, 69, 70, 71, 76, 96

Implement index selection (filtering) in bitwise.IA

bitwise.IA accepts as input a vector of locus indices to be used in its calculations. Currently this vector is ignored. The final version should be able to compute the index of association over this limited set of loci.

misspelling in package startup

unvailable is not a word.

add explicit package for non-base functions

Based on http://developer.r-project.org/blosxom.cgi/R-devel/NEWS/2015/06/29#n2015-06-29

Basically, import packages like stats and utils and then use the double colon to indicate what non-base functions exist (e.g. stats::optimize).

[.table method in R-devel breaks poppr.amova

This is an issue that will not affect most of the user base at the moment, but will do so when R releases a new version:

From Kurt Hornik:

Dear maintainer,

Current r-devel has added a [.table method which by default preserves
the table class in "most" cases.

This breaks your package with

amova.result <- poppr.amova(agc, ~Pop/Subpop)

No missing values detected.
Warning in Ops.factor(left, right) : ‘<’ not meaningful for factors
Warning in Ops.factor(left, right) : ‘<’ not meaningful for factors
Error in if (any(samples < 0)) stop("Negative value in samples") : 
 missing value where TRUE/FALSE needed
Calls: poppr.amova -> <Anonymous>

I think the problem is that in mlg.table() you have

 mlgtab <- mlgtab[, which(colSums(mlgtab) > 0)]
 return(mlgtab)

where subscripting used to drop the table class, but no longer does, and
so

 xtab    <- t(mlg.table(x, bar = FALSE, quiet = TRUE, mlgsub = unique(mlg.vector(x))))
 xtab    <- as.data.frame(xtab)

ends up calling as.data.frame.table().

To fix, use e.g.

 xtab    <- as.data.frame(unclass(xtab))

I think.

Can you pls fix this as necessary, and upload a new version as quickly
as possible?

Alongside, pls also fix

The Title field should be in title case, current version then in title case:
‘Genetic Analysis of Populations With Mixed Reproduction’
‘Genetic Analysis of Populations with Mixed Reproduction’

and single quote 'adegenet' in the Description.

Best
-k

Allow poppr.amova to take subsetted genclone objects

Currently poppr.amova partitions the mlg table by numerical index at this line. In the genclone objects, unfortunately, this index doesn't work as some MLGs might be excluded. The error produced is this:

Error in xtab[unique(mlg.vector(x)), ] : subscript out of bounds

To get around this, mlg.table should be called with the flag mlgsub = unique(mlg.vector(x, quiet = TRUE)), where x is a genclone object. This will automatically sort the MLG table to prevent this line from being needed.

Allow for named color vector input for msn functions

Currently the msn functions only allow palettes to be input. This becomes burdensome when the user wants specific colors for the nodes and has to input the palette such as function(x) myPalette

Give the option to return per-locus estimates of Bruvo's distance

This would likely be a list of distance matrices.

UBSAN shows a memory-access error

After submission of poppr version 1.1.0 to CRAN, a notification was sent from Dr. Ripley notifying me of an error in UBSAN while running tests:

Analytical value tests : poppr_distance.c:664:5: runtime error: index 1 out of bounds for type 'int [zerocatch[miss_ind]]'

This error stems from illegal memory allocation. I will sanitize the code and post a patch.

Hexp calcuation is erroneous

The problem:

After re-reading Nei (1978), I realized that my implementation of Hexp is:

N/(N - 1) * 1 - sum(p^2)

Where N is the number of allelic states and p is the vector of allele frequencies. Nei's definition is:

kn/(kn - 1) * 1 - sum(p^2)

Where n is the number of observed samples at a locus and k is the ploidy (to account for dosage).

User-facing impacts:

poppr()
locus_table()

What needs to be fixed:

internal function locus_table_pegas()

Impacts after fix:

Polyploids - Because the calculation is dependent on ploidy, this means that it would inappropriate to calculate this for polyploids due to ambiguous dosage.
locus_table(type = "genotype") - Hexp will change to unbiased simpson's index.

Unfortunately, I might have to wait until just before August to submit this patch, lest I anger our CRAN overlords.

Fix poppr for upcoming release of adegenet/pegas/hierfstat

The adegenet package is going through major changes on github and many functions in poppr will break due to these changes.

Three main things will change:

the tab slot of genind objects will store data as integers as opposed to floats
the ploidy slot will contain the ploidy for all samples
genind and genlight objects will gain genclone's hierarchy slot

Changes that need to be implemented in poppr:

remove hierarchy slots from genclone objects as they will be inherited
remove hierarchy methods
change all references to $tab and @tab to tab().
change na.replace(object) in missingno to tab(object) and deprecate.
ensure that distance methods are compatible.
change SEXP pairdiffs to take integer input.
change SEXP permute_shuff to take integer input.
add tab method to bootgen objects.

This is not an exhaustive list as I know bugs will crop up as I encounter them, so I'm putting this here:

WRITE A TEST FOR EVERY BUG ENCOUNTERED

Allow named vector of repeat lengths for Bruvo's distance

Bruvo's distance requires knowledge of repeat lengths for calculations. The current method has the user supply a numeric vector with repeat lengths, but they must be in the exact order as the loci. A nice feature to have would be to have the option to submit named repeat lengths and match them to the loci.

Diploid missing data not imported correctly with read.genalex

It imports missing data as "NA/NA"

Git Reports Issue

Submitter: test
Email: [email protected]
This is a test. close this issue.

Add tabular output for number of alleles/private alleles

In Kepler et al. 2015, Table 1 is a very nice representation of showing the number of alleles vs. private alleles per population. Implementing this output would be quite welcome in poppr.

Bug in permutation scheme with triploids

There is a bug in the default shuffling algorithm for shufflepop and ia that comes down to numerical precision. The general process is:

Take the sum of the columns (alleles) at a locus and multiply them by the ploidy to get the counts of each allele.
Create a vector of indices of alleles where each index is replicated the number of times that allele is represented.
Use that to shuffle the locus.

The problem occurs when there are triploids and 1/3 by 3 doesn't always equal 1.

mat <- structure(c(0.333333333333333, 0.333333333333333, 0.333333333333333, 
0.333333333333333, 0, 0, 0.333333333333333, 0, 0.666666666666667, 
0.666666666666667, 0.333333333333333, 0.666666666666667), .Dim = c(4L, 
3L), .Dimnames = list(structure(c("1", "2", "3", "5"), .Names = c("1", 
"2", "3", "4")), c("D13.000", "D13.144", "D13.190")))

bucket <- colSums(mat) * 3
bucket
## D13.000 D13.144 D13.190
##       4       1       7

rep(1:length(bucket), bucket)
## [1]11112333333

print(bucket, dig = 20)
## D13.000 D13.144 D13.190
## 4.0000000000000000000 1.0000000000000000000 6.9999999999999991118

A control scheme to check for the correct number of alleles is needed.

Allow for more tree functions in aboot

Currently, the only tree constructors are nj and upgma. Given that there are more available in ape, they should all be available.

ensure bitwise.IA can handle missing data.

User model choice being ignored in bruvo.msn

Current workaround solution: calculate bruvo's distance with `bruvo.dist()` and then feed it into `poppr.msn()`

Problem: all calls to bruvo.dist() from bruvo.msn() and poppr:::singlepop_msn() do not give the add or loss parameters, thus defaulting to average over genome loss and genome addition models, despite the user choice.

I will fix these in poppr's current devel branch and then make a pull request for @JonahBrooks so that he can incorporate it into his code.

Add accessors for loc.nall and loc.fac

Keeping up with the Jombarts: See This change.

Parallelize bitwise.IA

bitwise.IA was designed with multi-threading in mind, but the functionality hasn't been incorporated into the function yet. It needs to be added before this function can be considered complete.

Add names to colors in *.msn functions

Currently the colors of the pies will be represented by unnamed character vectors of hexadecimal colors. This vector of colors is used as a lookup table when the color palette needs to be changed, but the assumption is that all colors are different. In the case that not all colors are different, the names of the populations are important at differentiating the colors. If the vector resulting from the color palette is named with the population names, then the colors in the pie.colors will be named as well resulting in an easier lookup.

example here can add the additional line:

names(color) <- [email protected]

This should be changed for the functions bruvo.msn and poppr.msn and the functions below update_poppr_graph should also be updated.