R R G I B B S
A program for simple random regression analyses via Gibbs Sampling
Random regression (RR) models
have become a popular choice for the analysis of longitudinal data or
'repeated' records. Typically, analyses require numerous parameters,
i.e. (co)variances between RR coefficients and measurement error
variances, to be estimated, especially if the model of analysis
includes additional random effects such as maternal effects. Programs
for RR model analysis using restricted maximum likelihood (REML) are
available. However, the high computational demands of REML analyses
for RR models severely limit the feasibility of RR analyses for data
sets sufficiently large to support estimation of the pertaining
(co)variance components, in particular for models fitting many RR
Bayesian analyses using Gibbs sampling provide an alternative
which is markedly simpler to implement than REML. Whilst the range of
models which can be accommodated via Gibbs sampling may be more
restrictive and the total computing time required may be longer than
for corresponding REML analyses, memory requirements are substantially
less. Hence Bayesian methodology readily facilitates large scale
analyses. Apart from these practical advantages, of course, it
provides estimates of complete sampling distributions rather than just
simple point estimates.
RRGIBBS performs a single task : the analysis of a simple class of
RR models using Bayesian methodology. Models may involve :
- multiple fixed effects, including cross-classified effects and
'standard' covariables, as well as fixed regression(s) on Legendre
polynomials of the meta-meter;
- sets of random regression coefficients, regressing
on orthogonal polynomials or user-defined functions of a single,
continuous covariable, the so-called 'meta-meter';
- multiple random
effects, distributed proportionally to an identity matrix or the
numerator relationship matrix between animals,
- different orders of polynomial
fit for each random effect,
- homogeneous or heterogeneous measurement error
variances modelled as a step function of the covariable,
- a single trait
Model specification is via a parameter file. The run time behaviour
of RRGIBBS can be modified by a number of command line
options. RRGIBBS is a 'no-frills' program. It basically
offers little more than a reasonably efficient Gibbs sampler for a
range of random regression analyses. The main output are files with
the successive samples of (co)variance components drawn, ready for
your favourite post-Gibbs analysis. In addition, limited summary
information is produced, including estimates of covariance matrices
among RR coefficients and measurement error variances obtained as
means over samples (after 'burn-in'), and approximate 95% highest
posterior density regions.
is written in standard Fortran 95 and self-contained,
except for routines to generate random numbers. The source code is
available for downloading. Compiled versions are available for :
Linux, Compaq Alpha stations and SUN work stations.
is available free of charge to the scientific community under the conditions
that it remains my copyright, that it is not modified other than to adapt
it to the local computing environment or for personal research, and that
its use is credited
in any publications.
- While every effort
has been made to ensure that RRGIBBS
does what it claims to do, there is absolutely no guarantee for
- You are using RRGIBBS
entirely at your own risk, and there is no user-support service.
Meyer, K. (2002). RRGIBBS - A program for simple random regression analyses via Gibbs sampling. Seventh World Congress on Genetics Applied to Livestock Production, Montpellier, France, August 19-23, 2002, Paper No. 28-27, 2pp.
(2 pages, pdf file, 0.1Mb)
The FORTRAN source code for RRGIBBS
(together with a Unix Makefile) has been packaged into
a Unix 'tape archive' and compressed using 'gzip'. It can be downloaded as
1859 downloads since 18/6/2007).
As distributed, RRGIBBS is set
up to use routines from the RANDLIB90 package from the Department of Biomathematics
of the M.D. Anderson Cancer Centre at the University of Texas to the obtain
random samples required by the Gibbs sampler. If you choose to use this
software, download source code, documentation and installation instructions
- LINUX : RRGIBBS for Linux has been compiled
under Red Hat Fedora 2, using the Lahey/Fujitsu FORTRAN 95
compiler (version L6.20c).
1255 downloads since 19/6/2007).
Updated to accommodate sparse basis
functions such as B-splines.
- LINUX 64-bit version: compiled under Red Hat Fedora 5, using the Pathscale F95 compiler (version 2.4, 2006).
0 downloads since 19/11/2007).
COMPAQ Alpha station : RRGIBBS
has been compiled for a Compaq True64 work station, using the Compaq FORTRAN
compiler (V5.5-2602; UNIX V5.1A) : Download :
675 downloads since 19/6/2007).
- Updated to accommodate sparse basis functions such as B-splines.
A corresponding file for
a 32 bit machine, compiled using the Compaq FORTRAN compiler (V5.3-915)
is available as :
972 downloads since 19/6/2007).
SUN work station :
A Solaris version has been compiled using Sun Workshop 6 Fortran 95 6.0 on a SPARC
Ultra work station. Download :
1235 downloads since 19/6/2007).
The manual for RRGIBBS
comprises more than 30 A4 pages. It is available as a
PDF file. Download :
3527 downloads since 19/6/2007).
The worked example provided is
the same as for program DXMRR (in DFREML). Download
2137 downloads since 19/6/2007).
A small FORTRAN 90 program to
calculate the values of covariables when fitting B-splines basis
functions (run with command line option -u for an outline of its use,
option -h for valid command line options).
2927 downloads since 19/6/2007).
or a pre-compiled Linux executable
1184 downloads since 19/6/2007).
A tiny FORTRAN 90 program to
calculate translate the binary file of Gibbs samples (samples.rr)
into a formatted file (Samples.out).
2806 downloads since 19/6/2007).
or a pre-compiled Linux executable
2770 downloads since 19/6/2007).