inra / rgs3 Goto Github PK
View Code? Open in Web Editor NEWR package wrapping the GS3 program for genomic selection (http://genoweb.toulouse.inra.fr/~alegarra/).
License: GNU General Public License v3.0
R package wrapping the GS3 program for genomic selection (http://genoweb.toulouse.inra.fr/~alegarra/).
License: GNU General Public License v3.0
GS3 returned the following error:
At line 1199 of file gs3.f90 (unit = 10, file = 'config.txt')
Fortran runtime error: Bad integer for item 1 in list input
It seems to be due to the fact that NITER
is written as 1e+05
instead of 100000
.
To eventually prevent continuation of the MCMC when execGs3()
is run several times in the same directory.
Dear Timothée,
I was trying to download rgs3 from github but it failed, so I cloned the git repo and tried to build it myself on debian, Unfortunaltely, I am having problems. Do you think you could help?
Best Gregor
R CMD build rgs3
--- re-building ‘intro-rgs3.Rmd’ using rmarkdown
Quitting from lines 308-311 (intro-rgs3.Rmd)
Error: processing vignette 'intro-rgs3.Rmd' failed with diagnostics:
isValidData(x, inds, col.id, col.traits, binary.traits) ist nicht TRUE
--- failed re-building ‘intro-rgs3.Rmd’
SUMMARY: processing the following files failed:
‘crossval-rgs3.Rmd’ ‘intro-rgs3.Rmd’
Fehler: Vignette re-building failed.
Ausführung angehalten
In writeDataForGs3, the default value for binary.traits as only one value, but if there are missing data, binary.traits is expected to be a vector of length the number of columns of x because there is a test on binary.traits[c] with c in 1:ncol(x).
However, at the beginning there is a test on length(binary.traits) == length(col.traits) so length(binary.traits) cannot be ncol(x)
I think replacing
if (binary.traits[c]) {
by
if (binary.traits[which(c == col.traits)]) {
should work.
Hélène
It's seems that the GS3 manual at http://genoweb.toulouse.inra.fr/~alegarra/gs3_folder/ is not up-to-date (May 2014). The github repository contains a more recent version (Nov. 2016).
The link in the rgs3 vignette should probably therefore be modified to:
https://github.com/alegarra/gs3/blob/master/manual/manualgs3_6.pdf
or
https://github.com/alegarra/gs3/raw/master/manual/manualgs3_6.pdf
for a direct download.
I found a problem in the way missing data are handled in writeDataForGs3: if the data.frame with the phenotypic data has several columns before the trait of interest and these columns contain missing data, for the moment writeDataForGs3 will not transform them into -9999 and this will prevent reading data further by gs3. I had a phenotypic dataset with a missing data in the second trait at the 6th line, and when I asked for analysis of the third trait, only 5 records were read.
Concerning writeGenosForGs3, as gs3 can work with missing data in the genotypic data, and these have to be coded as "5", writeGenosForGs3 should accept values of 5 in the matrix of genotypic data.
So the stopping rules at the beginning should be
stopifnot(is.matrix(x), all(x %in% c(0, 1, 2, 5)), !is.null(rownames(x)),
is.vector(inds), is.numeric(inds), !is.null(names(inds)),
all(inds == 1:length(inds)), all(rownames(x) %in% names(inds)))
Hope this helps !
Hélène
Seems to be due to the usage of invisible
in function execGs3
.
Currently, the output files of the i-th fold are sol.txt_foldi
, config_train_gs3_foldi.txt_finalEstimates
, ..., which can be overwritten if several jobs run in the same directory.
In some data sets, genotypes have been phenotyped several times, but some of these data could be missing, which currently results in a NA
in cross-validation results such as the RMSPE.
This should be doable by executing GS3 without any configuration file, and parsing the stdout or stderr.
2varapqpi
-> X2varapqpi
Look at this task view, and notably at the parallel and doParallel packages.
install and library loading is OK but I couldn't the vignette on Windows:
browseVignettes("rgs3")
No vignettes found by browseVignettes("rgs3")
The histogram of gEBVs in vignette "intro-gs3" is not centered on 0, why?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.