Science in the box

P&G scientists use QSPRs and QSARs as the first step in a step-wise or tiered approach to environmental fate assessments. Through the mathematical relationships generated by the computational chemistry tools, exposure information can be estimated prior to any data from laboratory research.

We can understand a lot about how a molecule is going to behave from its structure. For example, small molecules containing from one to four carbon atoms are gases at room temperature (methane CH4, ethane C2H6, propane C3H8, butane C4H10). As more carbons are added, the substance becomes a liquid (hexane, a liquid, has six carbon atoms, C6H14) and finally a solid (octadecane, a solid, C18H38).
If we add one oxygen atom to methane (CH4), the molecule formed is a liquid we know as methanol (CH3OH). As we add chlorine atoms, nitrogen atoms or any other atom, we can predict the effect these additions will have on the molecule's behaviour. This behaviour is not limited to predicting whether the molecule is a solid, liquid or gas, but includes predicting how toxic or biodegradable the compound will be.
The tools and approaches used to generate a QSPR (Quantitative Structure Property Relationships) and a QSAR are similar. Which one the scientist uses depends on whether they are predicting a chemical property like solubility in water or a chemical activity like toxicity to algae or humans. Scientists use this information to predict where a chemical goes in the environment (see Fate pages) or the toxicity of a chemical. Since QSARs and QSPRs are equations that help to predict fate or effect values from chemical structures, they are quick and inexpensive to use.
P&G uses a step-wise or tiered approach to fate and effects work, beginning with screening level tools for an initial assessment. More sophisticated tools are used as needed. These different tiers allow P&G to eliminate compounds that will never make it to market for environmental or human health reasons, and to quickly identify less toxic, more rapidly degradable compounds. Ideally, we would like to eliminate problem compounds before much time and money are spent on research and development. QSARs and QSPRs are tier one tools and allow us to understand basic properties, such as toxicity, environmental fate and biodegradability, before any data are generated in the laboratory. From the toxicity, fate and biodegradation information, we are able to estimate exposure and risk to humans and the environment.

Environmental and human health scientists have studied the toxicity and biodegradation of molecules and have observed that these activities can be predicted from their structure via QSARs. For example, the more carbon atoms on a molecule, the more toxic it typically becomes. The more oxygen and nitrogen atoms on a molecule, the less toxic it typically is.

Toxicity tests with many chemicals have generated enough data to make QSAR equations that can be used to calculate toxicity and biodegradation based on chemical structure. We refer to toxicity and biodegradation as activities. We are not just predicting a chemical property, like the boiling point, but we are actually predicting what the molecule will do in the environment.

There are times when the equations that estimate QSAR effects work very well at predicting chemical activity at the first step in the research process. But while the data is increasing and the QSARs are progressively being refined, the thousands of potential equations for certain chemical groups means that the toxicity and biodegradation of certain compounds still cannot be accurately predicted by QSARs alone. More data and refinement is needed.

There are two basic problems with using equations.

• We cannot predict the toxicity and biodegradation of all compounds from equations. Chemicals can be divided into different families based on their structure. To develop a good QSAR, 15 to 50 (or more) effect values on chemicals from the same family are needed. That is a lot of data to generate one equation for one chemical family. With thousands of chemical families, we just do not have enough data to predict the toxicity and biodegradation of them all. Additional research is ongoing and more QSARs are being developed, but it will still be a long time before a QSAR is available for all compounds.


• Predictions are never as good as real data. Tomorrow's weather, the results of the next election, and who is going to win the big game are all examples of predictions. When the weather, the results of the election, or the outcome of the big game is important, we don't want to rely on predictions, we want to know the facts. We consider the prediction to be uncertain, even if predictions have been fairly accurate in the past.

It is the same with QSARs: The results may be pretty close to reality but we never really know until we run the test. As a result, scientists only use QSARs when they need an approximate value. For example, if a chemical is going to exist in the environment at 0.001 g/L and QSARs predict the safe concentration in the environment is 1000 g/L, the toxicologist might decide that it is not necessary to conduct a toxicity test. Even if the QSAR is close, the environment will not suffer and testing resources can be spent on compounds that pose a greater threat to the environment. For more on risk assessment and how decisions are made, see the environmental risk assessment page.
When a good QSAR relationship exists for a family of compounds and when an estimate of the toxicity value or biodegradation rate is all that is needed for the risk assessment, QSAR approaches are used.

CATABOL
Biodegradation is an important environmental process because it results in the destruction of the ingredient. However, some compounds do not biodegrade all the way to CO2. Metabolism (the breakdown of the molecule by bacteria) stops before the compound is fully degraded, leaving a partially degraded compound called a metabolite. Due to the importance of biodegradation for P&G products and the need to understand the formation of metabolites, we developed a QSAR program called CATABOL. CATABOL predicts the biodegradation pathway and predicts the possibility that degradation may stop at one or more metabolites.
ECOSAR
ECOSAR (Ecological Structure Activity Relationships) is a computer program used to estimate the toxicity of chemicals to aquatic organisms. To use this program, some chemistry background is useful since for many compounds you will need to enter the compound's structure using SMILES notation. Fortunately, the HELP section of the program contains all the information needed to learn about SMILES. Although the program predicts both acute and chronic toxicity for many compounds, we prefer to use the acute toxicity relationships because they are usually based on more compounds and more data and thus we have greater confidence in the results. In using any QSAR prediction, it is important to understand how many compounds the QSAR is based on and whether they are similar to the compound you are interested in.
DEREK
To better understand how humans respond to our ingredients, P&G uses an expert system called DEREK to assess skin sensitization, irritation, carcinogenicity or reproductive affects.
TIMES
To predict mutagenicity , P&G developed TIMES. TIMES makes predictions on the parent molecule and then looks for the potential for metabolic activation or deactivation.

• Cronin, M., Jaworska, J., Walker, J., Comber, M., and Watts. C. Use of QSARs in International Decision-Making Frameworks to Predict Ecological Effects and Environmental Fate of Chemical Substances. Environmental Health Perspectives, in Press.
• Dyer, S.D., Lauth, J.R., Morrall, S.W., Herzog, R.R., and Cherry, D.S., 1997. Development of a Chronic Toxicity Structure-Activity Relationship for Alkyl Sulfates. Environmental Toxicology and Water Quality, 12, pp. 295-303.
• Dyer, S.D., Stanton, D.T., Lauth, J.R., and Cherry, D.S., 2000. Acute and Chronic Structure Activity Relationships for Alcohol Ethersulfates. Environmental Toxicology and Chemistry, 19, pp. 608-616.
• Jaworska, J., Dimitrov, S., Nikolova, N., and Mekenyan, O., 2002. Chemical Biodegradability. Probabilistic Prediction Based on a Metabolic Pathway. SAR and QSAR in Environmental. Research, 13, pp. 307-323.
• Jaworska, J., Dimitrov, S., Nikolova, N., and Mekenyan, O., 2002. Chemical Biodegradability. Probabilistic Prediction Based on a Metabolic Pathway. SAR and QSAR in Environmental. Research, 13, pp. 307-323.
• Jaworska, J., Howard, P., and Boethling, R.S., 2003. Quantitative Structure Biodegradadation Relationships - A Review. Environmental Toxicology and Chemistry, in Press.
• Morrall, D.D., Belanger, S.E., and Dunphy, J.C. Acute and Chronic Aquatic Toxicity Structure-Activity Relationships for Alcohol Ethoxylates. Ecotoxicology and Environmental Safety, in Press.
• Morrall, S.M., Rosen, M.J., Zhu, Y., Versteeg, D.J., and Dyer, S.D., 1997. Physicochemical Descriptors for Development of Aquatic Toxicity QSARs for Surfactants. In Chen, F. and Schuurmann, G. (Eds.), QSAR '96, 7th International Workshop on QSARs in Environmental Sciences. SETAC Press, Pensacola, FL, pp. 299-313.
• Rosen, M.J., Li, F., Morrall, S.M., and Versteeg, D.J., 2001. The Relationship between the Interfacial Properties of Surfactants and Their Toxicity to Aquatic Organisms. Environmental Science and Technology, 35, pp. 954-959.
• Versteeg, D.J., Stanton, D.T., Pence, M.A., and Cowan, C.E., 1997. Effects of Surfactants on the Rotifer, Brachionus Calyciflorus, in a Chronic Toxicity Test and the Development of QSARs. Environmental Toxicology and Chemistry, 16, pp. 1051-1058.