Application Of Quantitative Structure Activity Relationship (QSAR) In The Research Of Medical Drugs

Drugs are special chemical substances, which are used for regulating physiological functions, overcoming diseases, maintaining healthy and improving life quality. On account of ceaseless appearance of new diseases and intense drug-resistant effects, it is becoming increasingly difficult to create new drugs and the traditional ways for drug design have already not satisfied the needs of new drugs nowadays. With the interaction and rapid development of subjects, such as modern biology, chemistry and computer science in recent 20 years, the availability of active compounds have already been shifted from the blindly random synthesis to the rational drug design. Computer-assisted drug design （CADD） makes full use of the interaction principle between ligands and their receptors, greatly improving the efficiency of drug development, shortening period and reducing research cost. Thus, it has become a modern efficient technique to create new drugs and the recent hotspot in the fields of medicinal chemistry and life science.In this thesis, PCA, HCA and Var.Jarvis-Patrick cluster methods in Cerius² software were used to reasonably classify compounds and then GFA, Stepwise-MLR and MFA-G/PLS were employed to establish some significant and credible 2D/3D-QSAR models for a series of anti-HIV drugs, anti-depression drugs and anti-HCV drugs. These QSAR models can indicate the quantitative change rules between bioactivity and microcosmic structural parameters in drugs, which can be used to guide the design and the synthesis of new drugs. The main results obtained are as follows:（1） Thiourea anti-HIV drugs. QSAR model shows that Hbond Acceptor, LUMO, Shadow_XYfrac and S_aaCH are main microcosmic structural parameters in controlling the activity and Shadow_XYfrac index should be crucially considered while designing new drugs of this kind. Increase of electrostatic substituents of molecule connecting with receptor can improve the activity.（2） Aryl alkanol piperazine anti-depressant drugs. The best 2D-QSAR models show that Atype_C6, Dipole-mag, Jurs-PNSA-3 and S_sssCH mainly control the reuptake inhibition activity of 5-HT and HOMO, PMI-mag, Shadow-XZ and S_sssN mainly control the reuptake inhibition activity of NA. The best 3D-QSAR models show that steric interaction and electrostatic interaction control the reuptake inhibition activity of 5-HT and NA, respectively. Finally, the activity of newly design molecules was predicted by the best QSAR models and molecule 38,39,44 and 46 were found having higher anti-depressant activity. （3） 1-Aryl-tetrahydroisoquinoline anti-HIV drugs. QSAR models shows that Sr, Density, Jurs-PPSA-3, AlogP98, AlogP, Fh2o, Foct and S_aaaC are main microcosmic structural parameters in controlling the activity and IC index should be crucially considered while designing new drugs of this kind. Appropriate increment of hydrophobic substituents near R group is favorable for the activity. Tiny change of substituents at Ar region would greatly influence the activity and withdrawing and hydrophobic substituents are favorable at this position. Hydrogen bonding interactions should be considered around the C4 and N of tetrahydroisoquinoline ring.（4） Benzo [b] thiophene-1,1-dioxide anti-HCV drugs. The best 2D-QSAR models show that Atype_C11, Atype_C6, Sr, S_dssC and Jurs-PNSA-3 are main microcosmic structural parameters in controlling the activity and Atype_Cll should be crucially considered while designing new drugs of this kind. The best 3D-QSAR models show that steric interaction and electrostatic interaction have great influence on the activity. The hydrophobic substituents are needed in the region between R₂ and R₃ groups, electrostatic substituents are favored around R₃ group. Increase in electrostatic and hydrophobic substituents around R₄ group are unfavorable.