Due to the widespread prevalence, deoxyuridine triphosphatase (UTPase) is considered by modern biochemists and physicians as a promising target for the development of drugs with a wide range of activities. The therapeutic effect of these drugs will be due to suppression of DNA biosynthesis in various viruses, bacteria and protozoa. In order to rationalize the search for new dUTPase inhibitors, domestic and foreign researchers are actively using the QSAR methodology at the selection stage of hit compounds. However, the practical application of this methodology is impossible without existence of valid QSAR models. With the use of the GUSAR 2013 program, a quantitative analysis of the relationship between the structure and efficacy of 135 dUTPase inhibitors based on uracil derivatives was performed in the IC50 range of 30¸185000 nmol/L. Six statistically significant valid consensus models, characterized by high descriptive ability and moderate prognostic ability on the structures of training and test samples, are constructed. To build valid QSAR models for dUTPase inhibitors can use QNA or MNA descriptors and their combinations in a consensus approach.
S ispol'zovaniem programmy GUSAR vypolnen kolichestvennyĭ analiz vzaimosviazi mezhdu strukturoĭ i aktivnost'iu 135 ingibitorov dezoksiuridintrifosfatazy cheloveka na osnove proizvodnykh uratsila, v intervale znacheniĭ IC50=30¸185000 nmol'/l. Postroeno shest' konsensusnykh modeleĭ QSAR, obladaiushchikh dostatochno vysokoĭ opisatel'noĭ tochnost'iu i predskazatel'noĭ sposobnost'iu. V rezul'tate virtual'nogo skrininga bazy dannykh ChEMBL s ispol'zovaniem postroennykh konsensusnykh modeleĭ vyiavleno 10 potentsial'nykh ingibitorov dezoksiuridintrifosfatazy, kotorye rekomendovany dlia éksperimental'nogo testirovaniia.
Keywords: GUSAR 2013 program; MNA and QNA descriptors; QSAR modeling; deoxyuridine triphosphatase inhibitors.