Towards data-driven design of visible-light photoswitches using structural features

In this manuscript we present the strategy for modeling of the photoswitch properties (maximum absorption wavelength and thermal half-life of photoisomers) of visible-light azo-photoswitches using structural data. We compile a comprehensive data set from literature sources and perform a rigorous benchmark to select the best feature type and modeling approach. The fragment counts have demonstrated excellent performance in the benchmark for both properties. We validate the models in cross-validation and on an external set. The predictions for this set are highly accurate, despite the modest size of the data set related to thermal half-life of photoisomers, especially when consensus modeling approach is applied. We also provide the interpretation of the modeling results using ColorAtom approach and the insights into the chemical space covered by the data set.