Method for Screening Human Transthyretin Interferors Using K-Nearest Neighbor Algorithm

Abstract

The invention discloses a method for screening human transthyretin interfering substances by adopting k-nearest neighbor algorithm. The method first calculates the quantitative descriptor based on the morphology correction for ionizable organic chemicals, and then uses the quantitative descriptor based on the morphology correction, functional group, molecular fragment descriptor and k-nearest neighbor algorithm to construct a binary classification model and quantitative prediction model; when screening target organic chemicals, they are firstly classified into active and inactive organic chemicals based on a binary classification model; then, quantitative models are used to predict the interference effect data of active organic chemicals; finally, judgments are made based on the predicted effect values Whether the target organic chemical is a potential human transthyretin disruptor. The descriptor mechanism of the present invention is clear and easy to calculate, the prediction method is easy to program, the prediction model has good fitting degree, robustness and prediction ability, the screening method has good scalability, and is suitable for screening potential human thyroid in the application field catalin disruptors.

Description

technical field [0001] The invention relates to a method for screening human transthyretin disruptors by using the k-nearest neighbor algorithm, and belongs to the technical field of endocrine disruptor screening strategies. Background technique [0002] The endocrine disrupting effect caused by environmental endocrine disruptors (EDCs) seriously threatens the safety of humans and wild animals, and is becoming a global environmental problem faced by human beings. For management, how to effectively identify and evaluate potential EDCs from commercial chemicals is the primary problem that chemical management departments in various countries need to solve. However, after years of practice, it has been found that only experimental methods are used to screen and evaluate potential EDCs, such as low throughput (50-100 chemicals per year), high cost (1 million US dollars per chemical), etc., making it difficult Test commercial chemicals one by one according to the existing testing...

AI Technical Summary

Problems solved by technology

However, the above method has the following limitations: (1) This method only classifies target organic chemicals based on 10 groups. If the target organic chemical does not contain these 10 groups, it cannot be classified. The interference effect of the organic chemicals of the group cannot be predicted; (2) the descriptor of this method is only the Dragon descriptor calculated based on the molecular state of organic chemicals, but Yang et al. (Yang XH, Xie HB, Chen JW, LiXH. Anionic phenolic compounds bind stronger with transthyretin than their neutral forms: nonnegligible mechanisms in virtual screening of endocrine disrupting chemicals. Chem Res Toxicol, 2013, 26(9): 1340-1347; Yang XH, Lyakurwa F, Xie XH, Chen, JW, Li Qiao XL, Cai XY. Different binding mechanisms of neutral andanionic poly- / perfluorinated chemicals to human transthyretin revealed by Insilico models. Chemosphere, 2017, 182, 574-583) with phenols, perfluoro / polyfluorocarboxylic acids, perfluorosulfonic acids organic Pollutants were used as model organic chemicals to study the interaction mechanism between ionizable organic chemicals and hTTR. It was found that the interaction between anionic organic chemicals and the protein was stronger than that of corresponding molecular forms, and the aromatic rings in phenolic organic chemicals could interact with hTTR. The residues of hTTR form cation-π interactions, that is to say, some ionizable organic chemicals will dissociate into ionic states under experimental or physiological pH conditions. During the interaction between ionizable organic chemicals and hTTR, the ionic states Like the molecular state, it has a non-negligible effect, so the method does not consider the impact of ionizable organic chemical ion state when constructing hTTR interferer prediction model

Images