From listmaster animalgenome.org Fri Feb 22 13:21:11 2019
Return-Path: <listmasteranimalgenome.org>
From: Andres Legarra-Albizu <andres.legarrainra.fr>
Subject: Re: software for exact p-values for association
analysis on single
Postmaster: submission approved by list moderator
To: Members of AnGenMap <angenmapanimalgenome.org>
Date: Fri, 22 Feb 2019 13:21:11 -0600
Patrik
the marker effect is shrunken, but its s.error. is shrunken as well. As a
result, the test (effect estimate / standard error) is the same as fixed
regression. This was proven (mathematically) years ago in bernal-rubio
et al. You may try to amuse yourself to verify with an example data set;
we did it because we could not believe.
Andres
On 21 Feb 2019, at 20:08, Patrik Waldmann <patrik.waldmannslu.se> wrote:
Max,
you're right, these test aren't exact. GBLUP is a form of ridge regression,
and as such provides shrinked, biased estimates of the regression
coefficients, and the variance (and standard deviation) of the estimates
will always be smaller compared to OLS estimates. The p-values cannot be
used in any form of hypothesis testing. It would be better to perform
single SNP regressions, and do some form of multiple comparison correction,
for example the local false discovery rate proposed by Bradley Efron.
Best regards,
Patrik Waldmann SLU, Uppsala Sweden
________________________________________
Frn: Rothschild, Max F [AN S] [mfrothsciastate.edu>]
Skickat: den 21 februari 2019 18:08
Till: Members of AnGenMap
mne: Re: software for exact p-values for association analysis on single step
gblup
Dear colleagues
I admit I maybe out of my depth but it is hard for me to understand how the
tests can be exact.
Please explain.
Best Max
________________________________
.From: Steibel, Juan <steibeljmsu.edu>
.Sent: Wednesday, February 20, 2019 8:14:10 PM
.To: Members of AnGenMap <angenmapanimalgenome.org>
.Subject: Re: software for exact p-values for association analysis on single
step
gblup
Dear Yuri
It is in fact a fixed SNP effect test derived from an animal centric GBLUP.
I recommend you take a look on the following paper (especially the
appendix): https://onlinelibrary.wiley.com/...oi/10.1111/age.12378
https://www.ncbi.nlm.nih.gov/pubmed/26607299
Meta-analysis of genome-wide association from genomic prediction models. -
PubMed - NCBI https://www.ncbi.nlm.nih.gov/pubmed/26607299 Anim Genet. 2016
Feb;47(1):36-48. doi: 10.1111/age.12378. Epub 2015 Nov 26. Meta-Analysis;
Research Support, Non-U.S. Gov't; Research Support, U.S. Gov't, Non-P.H.S.
http://www.ncbi.nlm.nih.gov
Where we show analytically the equivalence between a "snp-test" in a GBLUP
model and the EMMAX approach used for GWA (which is the answer to your
question, I guess).
Also, here are some simulation results showing the properties of the test
when we first proposed it: https://www.ncbi.nlm.nih.gov/pubmed/25038782
Rapid screening for phenotype-genotype associations by linear
transformations of genomic evaluations. - PubMed - NCBI BMC Bioinformatics.
2014 Jul 19;15:246. doi: 10.1186/1471-2105-15-246. Research Support, Non-
U.S. Gov't; Research Support, U.S. Gov't, Non-P.H.S. http://www.ncbi.nlm.nih.gov
These two paper (and others) are cited in Andres' manuscript as the basis
for their implementation in BLUPF90.
Cheers JP
________________________________
.From: Yuri Tani Utsunomiya <ytutsunomiyagmail.com>
.Sent: Wednesday, February 20, 2019 9:43 AM
.To: Members of AnGenMap <angenmapanimalgenome.org>
.Subject: Re: software for exact p-values for association analysis on single
step
gblup
Dear Andres,
I am a little bit confused with the proposed test.
As far as my limited knowledge goes, classic t-distributed statistics are
ratios between estimates of parameters of interest and their respective
standard errors. In the manuscript, the denominator of the ratio seems to
be the square root of the variance due to the tested marker instead (i.e.,
marker standard deviation), which is not the same as the standard error of
the estimate. The latest information I had was that the standard errors of
marker effects in GBLUP and ssGBLUP do not have a simple or even known
expression for fast computation, and therefore they should be approximated
with resampling procedures such as bootstraping, subsampling or
jackknifing. Please do not take this observation as a critic, but rather as
a question that - I believe - many of us users of blupf90 have. Last but
not least, I apologize if my read of the proposal is incorrect.
Cheers, Yuri Utsunomiya
On Wed, Feb 20, 2019 at 10:36 AM Andres Legarra-Albizu
<andres.legarrainra.fr> wrote:
Dear all,
thanks to collaborative work and extensive programming, I, on behalf of all
the blupf90 team, am happy to announce that the blupf90 suite now computes
exact p-values for gwas analysis - same as one at a time classical
gwas by regression. The authoritative reference is this biorxiv deposit:
MS ID#: BIORXIV/2019/555243
MS TITLE: Exact p-values for large-scale single step genome-wide
association, with an application for birth weight in American Angus
These requires the use of OPTION snp_p_value in programs blupf90 (or
blupf90test) and postGSf90. More details in the corresponding sections in
the wiki: http://nce.ads.uga.edu/...n_programs
regards, Andres
--
Andres Legarra
+33 561285182
INRA, UMR1388 GenPhySE
CS 52627
31326 Castanet Tolosan, France
--
*Yuri T. Utsunomiya*
DVM/MSc/PhD
Sao Paulo State University (UNESP - Brazil)
School of Veterinary Medicine - Department of Support, Production and
Animal Health (Araatuba/SP)
International Atomic Energy Agency (IAEA) Collaborating Centre on Animal
Genomics and Bioinformatics
Laboratory of Animal Biochemistry and Molecular Biology
--
Andres Legarra
+33 561285182
INRA, UMR1388 GenPhySE
CS 52627
31326 Castanet Tolosan, France
http://genoweb.toulouse.inra.fr/~alegarra
|
