How to avoid testing of all mixture`s components?
In order representative structures to be selected we need to be sure that all constrituents have consistent properties. To this end, several new functionalities have been implemented in the OASIS software.
- Cluster analysis (clustering) is the task of grouping a set of objects in such a way that objects in the same group (called cluster) are more similar (in some sense or another) to each other than to those in other groups (clusters).
- Cluster analysis is based on similarities between each two chemicals of an initial set. Similarity values are compared to a selected threshold in a repeated procedure.
- Resulting clusters include only substances which are similar to each other above the required threshold.
Clustering is most appropriate as a first step for selection of representatives. This will ensure the obtaining of manageable groups (clusters) of chemicals which are similar to each other based on a certain threshold or that have the same structural features of interest for the evaluation. Initial knowledge on the chemicals that will be clustered and the specific purpose of grouping should be available. This information is used to define the clustering criteria and the sequence of their application. In general, when more than one criterion for clustering is applied in a hierarchy, the least specific criterion should be applied first. For example, if structural and metabolic similarity criteria are used for clustering of a chemical set, then the first level in the clustering hierarchy should be based on the structural similarity (i.e., less specific similarity criteria). Once clusters with structurally similar members are formed, each of the clusters could be additionally subclustered based on the metabolic similarity between the chemicals (i.e., more specific similarity criteria). As a result of this two‐tier hierarchical clustering, groups of structurally and metabolically similar chemicals will be produced.
- Selection of representatives
Once groups consisting of similar (according to the selected criteria) chemicals are obtained, a representative could be selected out of each group (cluster).
The representatives could be selected in two ways:
- based on the worst-case scenario - in this approach we select the chemical with the worst-case properties, e.g. is the selection is based on bioavailability, we could select the chemical with the highest logKow value
- representing adequately the entire group where they are selected from - e.g. in this approach we could select the chemical being the most similar (according to the selected criteria) to all other chemicals in the same group
Example for an hierarchical scheme for selection of representatives for PBT screening/testing purposes:
Download the related article at Environmental Toxicology and Chemistry