Estimating the degree of similarity overlap for structural data in spectral databases

A method for the chemical structure overlap of two spectral databases estimating is suggested. The technique matches all compounds from the infrared (IR) spectral database against all compounds from the mass-spectral database using their structural similarity. The structural similarity is estimated by comparing of structural fragment sets obtained for each database using preliminarily decomposing of all molecular graphs to exhaustive sets of structural fragments with 2-7 nodes. Quantitative estimates of the relationship between the threshold on the structural similarity and the overlap degree of two databases was revealed. High degree of structural similarity overlap provides new possibilities for structure elucidation using two spectral databases simultaneously.