Support sites and documents
On the software. Manuals
and documentation can be downloaded from the download page. There is no
helpdesk for the use of the software. However, the DEBtox discussion board
has an area for discussions on GUTS. We hope that in the
future this becomes a place where users can help each other.
For reporting bugs, errors or wishes for future releases,
please use the discussion board or send an email to Tjalling
Jager (see email at bottom of this page).
For the Matlab version, feel free to
contact Tjalling Jager directly. He will be able to assist
with small problems. Larger problems or additional
functionality can be discussed as well, although they will
generally require funding.
On the statistical/optimisation
framework. The general statistical framework
and the automated optimisation algorithm as developed for
openGUTS have been described in detail in a publication,
including an extensive supporting information: Jager, T.
(2021). Robust likelihood-based approach for automated
optimization and uncertainty analysis of
toxicokinetic-toxicodynamic models. Integr. Environ.
Assess. Manag. 17(2):388-397 https://doi.org/10.1002/IEAM.4333.
On GUTS. The e-book on GUTS is
the definitive guide to the framework. It also provides the
conceptual background, as well as the technical details on
the model equations and the statistical framework. This book
is freely-downloadable from Leanpub (payment is optional).
On risk assessment with GUTS.
The EFSA
opinion on TKTD models is the current go-to document
when it comes to risk assessment for pesticides. Also
relevant in this context is the EFSA
opinion on good modelling practice. Note that openGUTS
can also be used to calculate a no-effect concentration
(threshold), or the traditional LCx,t values,
for a range of effect levels x and time points t
(using the model as calibrated to the complete data set with
all observations over time).
See also the following paper: Brock et al.
(2021). Application of GUTS models for regulatory aquatic
pesticide risk assessment illustrated with an example for
the insecticide chlorpyrifos. Integr Environ Assess Manag
17(1):243-258. https://doi.org/10.1002/ieam.4327
Publications on GUTS.
The DEBtox information website maintains a list of
publications that apply GUTS, or predecessors now
viewed as special cases (the list only includes papers that
use stochastic death). Currently, this list contains more
than 100 entries.
Alternative software for GUTS analyses
If openGUTS cannot do the analysis that you want, or if
you want a 'second opinion', feel free to try one of the
alternative software platforms. We have not tested these
versions in detail, so we cannot comment on their
correctness or user-friendliness. However, these platforms
(often in an older version) were included in the ring test,
which is reported in the e-book. A
potential source of differences between these platforms lies
in how they constrain parameter space; for some data sets,
parameters will not be constrained by the data, and can go
to zero or infinity (consult the interpretation guide from
the download page for typical
examples). 'Frequentist' applications generally use hard
min-max boundaries, whereas Bayesians use (weakly)
informative prior distributions.
MOSAIC. A web-based and
user-friendly way to perform GUTS calculations (reduced SD
and IT models only) is available at https://mosaic.univ-lyon1.fr/guts.
See also the papers of Baudrot et al
(2018) and Charles
et al (2022). Under the hood, it applies Bayesian
inference with the MORSE R-package (see below). Predictions
for exposure profiles can be made with the associated Shiny app. The only concern
with this platform is that the rules for designing priors,
as used in this version, block 'slow kinetics'
(specifically: the prior for the threshold does not extend
to low-enough values). An example is provided in the "test
results for openGUTS" document from the download page. as well as in the
supplement of Jager (2021).
Since 'slow kinetics' would often be a 'worst case'
interpretation of the data in risk assessment, use of this
software requires due care (note any warnings that may be
shown in the results).
GUTS-3S. Developed at
the Fraunhofer Institute in VB.net as a standalone
executable. More information. This
implementation (v.1.0) applies Euler integration with a
variable step size, rather than the (partial) analytical
solutions of openGUTS.
EasyGUTS. Under
development by RIFCON. A test version is available on
request. More
information (scroll down to EasyGUTS). It is a shell
around the GUTS R-package of Albert and Vogel (see below).
R-packages. There are
two R-packages available for GUTS calculations in a Bayesian
framework. The first package was developed by Carlo Albert
& Sören Vogel, is currently maintained by RIFCON, and
can be found at https://cran.r-project.org/package=GUTS.
The second is the MORSE R-package, as developed and
maintained by the University of Lyon. It contains the
reduced SD and IT models. This package can be found at https://cran.r-project.org/package=morse
and forms the basis for the web-based interface MOSAIC.
Requires installation of R (free). It is good to note that
the rules for designing priors, as used in this version,
block 'slow kinetics'. An example (using MOSAIC, which
applies the same rules) is provided in the "test results for
openGUTS" document from the download
page.
Python package. A
Python toolbox for GUTS calculations was developed by
Raymond Nepstad (SINTEF, Trondheim), and can be downloaded
from GitHub: https://github.com/nepstad/epytox.
Requires installation of Python (free).
Matlab packages. Apart
from the Matlab version of openGUTS, there is also an
extensive Matlab toolbox as part of the BYOM platform
(developed and maintained by Tjalling Jager, DEBtox
Research). This toolbox includes all GUTS cases (also the
full model), can do 'frequentist' and Bayesian inference,
but has limited user-friendliness. Requires a license for
Matlab.
Mathematica package.
Developed by Andreas Focks (WUR, Wageningen). This package
is part of the huge file with supporting information of the
EFSA
opinion (if you are able to find it), but the
Mahematica part can also be downloaded here
(ZIP file, 17 MB). Requires a license for Mathematica. This
software will remain a research version,
and is not being further developed for public release.
Links with more information or useful tools
- Roman Ashauer's website about models in ecotoxicology
at: www.ecotoxmodels.org.
- The DEBtox information site at www.debtox.info.
This site contains information on GUTS and DEBtox, and
also includes lists of publications that feature these
models.
- SETAC-Europe interest group on effect modelling. More
information.
|
|
