Fangyou Yan

QSAR models for describing the toxicological effects of ILs against Staphylococcus aureus based on norm indexes

The hazardous potential of ionic liquids (ILs) is becoming an issue of great concern due to their important role in many industrial fields as green agents. The mathematical model for the toxicological effects of ILs is useful for the risk assessment and design of environmentally benign ILs. The objective of this work is to develop QSAR models to describe the minimal inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) of ILs against Staphylococcus aureus (S.xa0aureus). A total of 169 and 101 ILs with MICs and MBCs, respectively, are used to obtain multiple linear regression models based on matrix norm indexes. The norm indexes used in this work are proposed by our research group and they are first applied to estimate the antibacterial toxicity of these ILs against S.xa0aureus. These two models precisely and reliably calculated the IL toxicities with a square of correlation coefficient (R2) of 0.919 and a standard error of estimate (SE) of 0.341 (in log unit of mM) for pMIC, and an R2 of 0.913 and SE of 0.282 for pMBC.

Predicting the activity of hydroxamic acid analogues

Abstract In a novel therapeutic approach, hydroxamic acids analogues such as Histone deacetylases (HDACs) inhibitors, have been identified as attractive drugs to treat cancer and related diseases. In this work, based on the Euclidean distance matrix and property matrix of HDACs inhibitors’ structures, a new norm index was proposed, and then according which the pIC50 values (the half maximal inhibitory concentration) of 143 HDACs inhibitors were predicted using the quantitive structure–activity relationship (QSAR) method. Results suggested that this model could give satisfactory prediction effect with high relationship coefficient of leave-one-out/5-fold/10-fold cross-validation. Moreover, the application domain of this model was validated to be satisfaction by applying the leverage approach. The statistical and comparison results showed that this new model could effectively improve the accuracy and predictive ability for predicting the bioactivity of HDACs inhibitors. All the above could demonstrate that this developed QSAR model could be effectively utilised for the description of structure–activity relationship at the molecular level. Therefore, this norm index-based QSAR model could provide an effective method for predicting the bioactivity of hydroxamic acids analogues drugs, which also might be valuable for the design of potential leading drugs.

QSAR model for predicting the toxicity of organic compounds to fathead minnow

In this work, a new norm descriptor is proposed based on atomic properties. A quantitative structure-activity relationship (QSAR) model for predicting the toxicity of organic compounds to fathead minnow is further developed by norm descriptors. Results indicate that this new model based on the norm descriptors has satisfactory predictive results with the squared correlation coefficient (R2) and squared relation coefficient of the cross validation (Q2) of 0.8174 and 0.7923, respectively. Combining with Y-randomization test, applicability domain test, and comparison with other references, calculation results indicate that the QSAR model performs well both in the stability and the accuracy with wide application domain, which might be further used effectively for the safe and risk assessment of various organics.

QSAR models for describing the toxicological effects of ILs against Candida albicans based on norm indexes

The quantitative structure-activity relationship (QSAR) model is an effective alternative to traditional experimental toxicity testing, which is undoubtedly important for modern environmental risk assessment and property prediction. Based on this background, the toxicological effects of ionic liquids (ILs) against Candida albicans (C.xa0albicans) were studied via the QSAR method. A large diverse group of 141 and 85 ILs that have a minimal inhibitory concentration (MIC) and a minimum fungicidal concentration (MBC) against C.xa0albicans were used to obtain multiple linear regression models. These two models were developed based on matrix norm indexes and proposed based on the atomic character and position. Matrix norm indexes proposed in our research group were used to calculate the toxicity of these ILs towards C.xa0albicans for the first time. These two models precisely estimated the toxicity of these ILs towards C.xa0albicans with a square of correlation coefficient (R2) ofu202f=u202f0.930 and a standard error of estimate (SE) ofu202f=u202f0.254 for pMIC, and for pMBC, R2u202f=u202f0.873 and SEu202f=u202f0.243.