ИСТИНА

The natural humic systems of non-living organic matter occur throughout the entire Earth’s environment. The complexity of molecular space of humic systems has been revealed due to advances in high resolution Fourier Transform Mass Spectrometry, which resolved in their composition millions of carbon atoms with distinct chemical environments. As a result, even simple enumeration of molecular constituents of humic systems is beyond any computational capacity: an efficient data reduction is needed. Here, we present a simple tool to quantify and map the chemical space of humic systems using cell-based partitioning approach. We have achieved this by binning the compositional space defined by a Van Krevelen diagram into cells, where the occupation density of each cell serves as a molecular descriptor. A set of these descriptors provides a unique identification for the sample. We demonstrate here applicability of this approach to classification of humic systems and to deriving QSAR for optic properties.