NOMIRACLE workpackage 4.1
  Effects of mixtures of compounds
  
Required amount of research work:
  4 year PhD, or 3 year Postdoc (theoretical part) plus 
  4 year PhD, or 3 year Postdoc (experimental part)

Motivation:

The effect of toxic compounds on organisms can best be understood in
terms of deviations from untressed physiology. The Dynamic Energy
Budget (DEB) theory quantifies the feeding, maintenance, development,
growth, reproduction and aging processes of organisms (Kooijman,
2000), and provides a sound basis for the quantification of effects of
compounds as function of the internal concentration. This
quantification is by a change of compound-specific parameter values of
the (unstressed) physiology of resource allocation to the various
endpoints. Effects specifications in terms of external concentrations
of chemical compounds obviously involves a toxico-kinetics
module. This way of quantifying effects turns out to be effective. A
report by the ISO and OECD will soon appear (Kooyman et al, 2004) that
accepts this method for the analysis of data from standardized
ecotoxicity tests (acute mortality, (fish) body growth, (daphnia)
reproduction, algal population growth).

The method allows for the estimation of NECs of pure compounds
(Kooijman & Bedaux, 1996). This method makes use of expectations for
how effects would become visible in responses as functions of
concentration of test compound and exposure time. Although the method
might seem complex at first sight, it offers a methodology to
extrapolate e.g. from acute to chronic exposure, from effects on
individuals to that on populations. 

Research specification:

We now plan to supplement this method with methodology for the
assessment of the toxicity of mixtures. We plan to work with two
extreme cases of mixtures of compounds:

The first extreme is a mixture of two compounds, where all
  detailes of the toxicity of both compounds are still visible in the
  data.
The second extreme is a mixture of many compounds, where
  key-parameters, such as the NEC, the tocixity and the elimination
  rates of the compounds can be thought of a random trials from some
  simultaneous frequency distribution. Baysian methods will be
  developed to estimate parameter values (e.g. mean and variance
  estimates of the key-parameters) from data sets that meet specified
  quality criteria.
We will focus on profile likelihood methods to study the confidence
intervals for the NEC-estimates. We will make use of Monte Carlo
Methods toevaluate the properties of the estimation method.

Since the method is an extension of the DEB-based method for
quantifying effects, the co-limitation by food as a factor that
modifies toxic effects is automatically implied, as well as the
extrapolation form effects on individuals to those on populations.

The models will be tested against experimental data of the following
type. The effects of pH will be tested on the toxicity of a mixture of
the molecular and ionic form of a test compound to springtails and
enchytraeids. The tests will be performed in a standard soil and focus
on effects on survival, growth and reproduction. In a second set of
experiments the link will be made between toxicokinetics and effects
in two mixtures of two compounds; one mixture concerns compounds with
similar mode of action, the other with different modes of action. In a
third set of experiments the toxicity of a rich mixture of many
compounds will be evaluated, e.g. creosotes.

References:

Kooijman, S. A. L. M. (2000) Dynamic Energy and Mass Budgets in
Biological Systems. Cambridge University Press.

Kooijman, S. A. L. M., Bedaux, J. J. M., P\'{e}ry, A. R. R. and Jager,
T. (2004) Biology-based methods. In: H. Magaud (ed) Draft guidance
document on the statistical interpretation of ecotoxicity tests. ISO
and OECD, TC 147/ SC 5/ WG 10/ N0390, Paris.

Kooijman, S. A. L. M. and Bedaux, J. J. M. (1996) The analysis of
aquatic toxicity data. VU University Press, Amsterdam.