Currently in Toolbox, UVCBs could be recognized only by their names. Their representation based on a selected single representative chemical is inaccurate. In this respect associated data points does not correspond to structure representation of UVCB. In other words, it is very difficult to select UVCB analogues and use their data points for read-across prediction of a target UVCB.
On the other hand Impurities and Additives to the target chemical (Reach dissemination data in ECHA CHEM) are indicated and could be removed among analogues of the target discrete chemical by “Substance type” metadata field. For details in this respect see comment to a similar question here:
https://qsartoolbox.freshdesk.com/solution/articles/16000043016-how-can-i-remove-the-data-for-uvcbs-from-the-data-points-used-for-qsar-prediction-
In collaboration with ECHA, LMC has developed a system for structure elucidation of UVCBs which is in the private domain, see attached ppt in this respect: Methodology-UVCBs.ppt .
We should mention that the structural elucidation in UVCBs requires additional information for the substances, methodology and the associated software is not implemented in the Toolbox.