CN115364099B

CN115364099B - Antibacterial application of repaglinide and antibacterial activity prediction and structural novelty evaluation method

Info

Publication number: CN115364099B
Application number: CN202111293501.XA
Authority: CN
Inventors: 代绍兴; 仝新; 姜辰龙; 黄京飞; 赖仞; 李文兴; 李功华; 梁积浩; 郑阳; 杨鹏鹏
Original assignee: Kunming University of Science and Technology
Current assignee: Kunming University of Science and Technology
Priority date: 2021-11-03
Filing date: 2021-11-03
Publication date: 2023-11-03
Anticipated expiration: 2041-11-03
Also published as: CN115364099A

Abstract

The application discloses antibacterial application of repaglinide, an antibacterial activity prediction and a structural novelty evaluation method, and relates to the technical field of medicines, wherein the antibacterial activity prediction and the structural novelty evaluation of a compound are carried out, and the antibacterial activity prediction of the compound specifically comprises the following steps: step 1, collecting high-flux data for finishing antibacterial activity to form antibacterial activity reference data; and 2, generating Dayleight molecular fingerprint characteristics of the reference compound by using Pybel and PyDPI and constructing an activity prediction model. And 3, predicting and evaluating antibacterial activity and structural novelty of the compound to be tested by using the model and fmcsR. And step 4, performing experimental verification on the repaglinide with high potential. The application can provide important thought and guidance for the development of novel antibacterial drugs, and more importantly, provides a novel low-toxicity antibacterial active compound Repaglinide (Repaglinide) to cope with the increasingly serious bacterial drug resistance crisis.

Description

Antibacterial application of repaglinide and antibacterial activity prediction and structural novelty evaluation method

Technical Field

The application relates to the technical field of medicines, in particular to antibacterial application of repaglinide and an antibacterial activity prediction and structural novelty evaluation method.

Background

Antibiotics are the basic stone of the current medicine, and the discovery and use of antibiotics save countless lives. Alexander fleming found penicillin in 1929 and used for the treatment of various infectious diseases, marking the advent of the antibiotic era. During 1960-1970, scientists developed numerous antibacterial drugs and have been widely used in medicine. Due to various factors such as economic benefits, inherent difficulty in developing antibacterial drugs, etc., the development of antibacterial drugs has been seriously delayed in recent decades, and the development speed of new antibacterial drugs has far not kept up with the development speed of bacterial resistance.

The spread of bacterial drug resistance is controlled, and the development of novel antibacterial drugs becomes the consensus of the whole society. However, the development of new antibacterial drugs faces a number of difficulties. First, antibiotic classes have been largely explored in the early stages, and it is now difficult to find new antibacterial agents. The industry is severely starved of antimicrobial agents in research pipelines. How to discover novel antibacterial active compounds from billions of compounds comprising huge chemical structural space is of great importance for antibacterial drug development. However, there is currently no ready-made method and framework to achieve this, which is still a challenging task.

Repaglinide (Repaglinide) is a short acting insulin secretagogue used in type 2 diabetic (non-insulin dependent) patients whose hyperglycemia is not effectively controlled by diet control, weight loss and exercise. At present, repaglinide (Repaglinide) has not been reported to have antibacterial activity. Secondly, there is no method in the prior art which allows the discovery of novel antibacterial active compounds and the experimental verification of these compounds. To this end, we propose a method for predicting antibacterial activity and evaluating structural novelty by which the antibacterial potential of all marketed drugs is systematically explored and evaluated.

Disclosure of Invention

Object of the application

In view of the above, the application aims to provide antibacterial application of repaglinide, antibacterial activity prediction and structural novelty evaluation method, so as to predict antibacterial activity of a compound and evaluate structural novelty of the compound.

(II) technical scheme

In order to achieve the technical aim, the application provides antibacterial application of repaglinide, antibacterial activity prediction and structural novelty evaluation methods:

the method comprises the steps of predicting the antibacterial activity of a compound and evaluating the structural novelty of the compound, wherein the step of predicting the antibacterial activity of the compound specifically comprises the following steps of;

step 1, collecting high-flux data related to antibacterial activity and cytotoxicity, and performing finishing filtration and analysis on the data to form an antibacterial activity benchmark database;

step 2, generating a composition descriptor, a topology descriptor, a molecular connection, a molecular charge descriptor and a Dayleight molecular fingerprint feature of a compound to be predicted by using Pybel and PyDPI;

step 3, selecting the features generated in the step 2 through a feature selection module in the scikit-learn to obtain corresponding features to be detected;

step 4, constructing a support vector machine prediction model and a support random forest prediction model;

step 5, evaluating classification performance of all models constructed in the step 4 by adopting a 5-fold cross validation method for 10 times;

step 6, combining the support vector machine with good prediction performance and the random forest prediction model to form an antibacterial prediction model after the evaluation in the step 5;

and 7, predicting the characteristics to be detected in the step 3 through an antibacterial prediction model, and when both models predict that the compound has antibacterial activity, using the compound as a candidate antibacterial compound.

Preferably, the data collected in step 1 specifically include baseline data sets of antibacterial activity and all marketed small molecule drug data sets;

wherein the baseline data set of antibacterial activity is all antibacterial activity data downloaded from the ChEMBL database;

all the marketed small molecule drug data sets are all marketed small molecule drugs downloaded from the drug bank drug database and their corresponding information.

Preferably, the evaluation in the step 5 is performed by using five indexes of ROC curve, accuracy, precision, recall and F1 Score, and the calculation formula is as follows:

wherein TP is true positive, TN is true negative, FP is false positive, and FN is false negative.

Preferably, in the step 4, the construction of the support vector machine prediction model is specifically implemented by using libsvm27 encapsulated in a Python-based machine learning module library Scikit-learn.

Preferably, in the step 4, the construction of the model for supporting random forest prediction is specifically to train and predict the sample by using a random forest classifier in a machine learning module library Scikit-learn based on Python, so as to construct the model for supporting random forest prediction.

Preferably, the evaluating the structural novelty of the compound specifically includes the following steps:

step 1, calculating the overall structural similarity of a candidate compound and all known antibacterial drugs through Pybel, and measuring through a valley coefficient TC, wherein the calculation formula of the TC value is as follows:

tc=c (i, j)/U (i, j), where C (i, j) represents the number of features in common in the molecular fingerprints of two small molecules i and j and U (i, j) represents the number of features in common in the molecular fingerprints of two small molecules i and j;

step 2, using Pybel to generate FP2 molecular fingerprint and calculating TC value;

and 3, judging whether the calculated TC value is lower than 0.5, if so, the similarity of two small molecules is very low, and the structure of the selected compound is novel.

Preferably, the evaluating the structural novelty of the compound specifically further includes the following steps:

step 1, constructing a substructure library of the effective groups of all known antibacterial drugs on the market, and then utilizing fmcsR to perform substructure search on newly discovered candidate antibacterial compounds;

step 2, if the candidate compound does not contain the active substructure of the known antibacterial agent and the overall similarity is less than 0.5, the compound has structural novelty.

In addition, we provide the use of repaglinide for the manufacture of an antibacterial medicament.

From the above technical scheme, the application has the following beneficial effects:

the application develops a novel method capable of accurately predicting the antibacterial activity of the compound and evaluating the structural novelty through a machine learning method and integrating antibacterial data, which is used for exploring novel antibacterial active compounds. The accuracy of the method exceeds 91%, and hundreds of millions of compound libraries can be rapidly screened. Screening against a drug bank drug database can relocate from all drugs on the market to obtain novel antibacterial drugs. Since marketed drugs have generally passed safety tests, they have low toxicity.

The Repaglinide (Repaglinide) which is predicted and experimentally verified by the application has the following characteristics different from the existing antibacterial drugs besides the inhibition activity, safety and low toxicity characteristics of various bacteria and fungi. Repaglinide (Repaglinide) is chemically distinct from existing marketed antibacterial agents, does not contain the active substructure of known antibacterial agents and has an overall similarity of less than 0.3. Repaglinide (Repaglinide) is a short acting insulin secretagogue for use in type 2 diabetic (non-insulin dependent) patients whose hyperglycemia is not effectively controlled by diet control, weight loss and exercise. At present, repaglinide (Repaglinide) has not been reported to have antibacterial activity. Is expected to be applied to the preparation of medicines for resisting drug-resistant bacteria.

In summary, the application not only can provide important ideas and guidance for the development of novel antibacterial drugs, but also provides a novel low-toxicity antibacterial active compound Repaglinide (Repaglinide) to cope with increasingly serious bacterial drug resistance crisis.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present application, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of a method for predicting antibacterial activity of a compound and evaluating structural novelty.

Fig. 2 is a chemical structure diagram of Repaglinide (Repaglinide) provided by the present application.

Fig. 3 is a diagram showing the performance of the support vector machine model and the random forest model in antibacterial activity prediction.

Fig. 4 is a graph showing the results of structural similarity calculation and activity prediction of Repaglinide (Repaglinide) provided by the present application.

Fig. 5 is a diagram of structural novelty evaluation results of Repaglinide (Repaglinide) provided by the present application.

Fig. 6 is the antibacterial activity experimental data of Repaglinide (Repaglinide) provided by the present application.

Detailed Description

The following description is merely exemplary in nature and is not intended to limit the present disclosure, application, or uses. It should be understood that throughout the drawings, the same or similar reference numerals indicate the same or similar parts and features. The drawings merely schematically illustrate the concepts and principles of embodiments of the disclosure and do not necessarily illustrate the specific dimensions and proportions of the various embodiments of the disclosure. Specific details or structures may be shown in exaggerated form in particular figures to illustrate related details or structures of embodiments of the present disclosure.

Referring to fig. 1-6:

example 1

A method for predicting the antibacterial activity of a compound and evaluating the structural novelty of the compound comprises the steps of predicting the antibacterial activity of the compound and evaluating the structural novelty of the compound, wherein the method for predicting the antibacterial activity of the compound specifically comprises the following steps of;

In addition, the data collected in step 1 specifically include baseline data sets for antibacterial activity and all marketed small molecule drug data sets;

wherein the baseline data set for antibacterial activity is all antibacterial activity data downloaded from the ChEMBL database. Active and inactive compounds are specifically filtered according to the semi-inhibitory concentration (IC 50). An antibacterial compound with an IC50 of less than 1000 nM (10. Mu.M) and an inactive compound with an IC50 of greater than 10000nM (100. Mu.M). The baseline dataset contained 1097 antimicrobial active compounds and 578 inactive compounds in total. Since the number of negative data samples in the data set is smaller than that of positive data, a mode of randomly sampling the positive data set is adopted to acquire balanced data sets with the same number as that of the negative data samples. This process is repeated multiple times, ensuring that the predictive model does not deviate significantly between each repetition.

All the marketed small molecule drug data sets are all marketed small molecule drugs downloaded from the drug bank drug database and their corresponding information. A total of 4196 small molecule drugs on the market were included, of which there were 427 antibacterial drugs.

Specifically, in the evaluation in the step 5, five indexes including ROC curve, accuracy, precision, recall and F1 Score are adopted for evaluation, and the calculation formula is as follows:

In the step 4, the construction of the support vector machine prediction model is specifically performed by using libsvm27 encapsulated in a Python-based machine learning module library Scikit-learn.

Specifically, the area under line (AUC) of the ROC curve is used to select the best model and parameters. And finally, determining the kernel function adopted by the optimal model as 'rbf', wherein the penalty parameter C is 50, and other parameters adopt default settings.

The construction of the support random forest prediction model is specifically to train and predict samples by using a random forest classifier in a machine learning module library Scikit-learn based on Python, so as to construct the support random forest prediction model.

Specifically, the parameters were set as follows: (1) The number of decision trees is 950, and the parameter is selected by using the area under line (AUC) of the ROC curve; (2) other parameters employ default settings.

Example two

The method for predicting antibacterial activity of a compound and evaluating structural novelty comprises the following steps of:

and 3, judging whether the calculated TC value is lower than 0.5, if so, judging that the similarity of two small molecules is very low, and selecting a compound structure to be novel.

In addition, as a preferable mode, the method for evaluating the structural novelty of the compound specifically comprises the following steps:

Specifically, the substructure library comprises the substructures of the effective groups of all known antibacterial agents on the market, such as sulfonamides, penicillins, cephalosporins, carbapenems, chloramphenicol, tetracyclines, aminoglycosides, macrolides, glycopeptides, quinolones, linezolid, lipopeptides and the like.

The Repaglinide was predicted and evaluated using the above method, resulting in a calculated maximum similarity of 0.38 for Repaglinide (Repaglinide) and the existing 427 antibacterial agents, with an overall average similarity of only 0.20 (see fig. 4). While Repaglinide (Repaglinide) does not contain the active substructure of the common 10 broad classes of antibacterial drugs (see figure 5). Repaglinide is thus structurally novel. It should be noted that an overlay of less than 1 in the table indicates that the substructure is not included.

Example III

The repaglinide is predicted and evaluated by the method for predicting the antibacterial activity of the compound and evaluating the structural novelty, so that the repaglinide has the antibacterial activity and the structural novelty, and the antibacterial activity is verified by experiments, wherein the method comprises the following specific steps:

step 1, 20 mu l of each of escherichia coli, candida albicans, bacillus subtilis and staphylococcus aureus are respectively added into an LB liquid culture medium to be shake-cultured at 37 ℃ until the escherichia coli, candida albicans, bacillus subtilis and staphylococcus aureus are in a cloud form (OD 600 is approximately equal to 0.6). The cloudy fungus solution was centrifuged (5000 rpm,5 min), 100. Mu.l of liquid medium was left, and the fungus was blown and mixed uniformly and spread on the surface of solid LB medium. A sterilized 5mm diameter piece of filter paper was placed on the surface of the medium, and 10. Mu.l of sample Repaglinide (Repaglinide) was added dropwise to the center of the filter paper. After the sample is completely absorbed, the sample is poured into 37 ℃ to be cultured for 10-12 hours, and the formation of the inhibition zone is observed, wherein the formation of the inhibition zone indicates that the antibacterial activity is achieved.

Step 2, according to the same procedure as described above, using Ampicillin and Fluconazole as positive controls, 4. Mu.l of positive control solution was added dropwise to the center of a 5mm diameter filter paper sheet.

Step 3, calculating the bacteriostatic activity value (relative to a positive control) of the sample to be tested according to the following formula:

the activity value (%) = (diameter of the bacteriostasis ring of the sample to be detected/diameter of the positive control bacteriostasis ring) of the sample to be detected) is 100

Step 4, the antibacterial activity of Repaglinide (Repaglinide) was determined as shown in fig. 6. Fig. 6 shows the inhibitory activity of Repaglinide (Repaglinide) against various bacteria and fungi, showing a pronounced inhibitory effect of Repaglinide (Repaglinide) against all of escherichia coli, staphylococcus aureus, bacillus subtilis, candida albicans. Repaglinide (Repaglinide) has an inhibitory activity against escherichia coli of 81.53% of ampicillin and 134.4% of fluconazole against candida albicans.

The above experimental results show that Repaglinide (Repaglinide) has a chemical structure different from that of the antibacterial drugs on the market and has inhibitory activity against various bacteria. Therefore, repaglinide (Repaglinide) is an antibacterial compound with a novel structure and can be applied to the preparation of antibacterial drugs.

For this purpose we provide the use of repaglinide, i.e. for the preparation of an antibacterial drug.

The exemplary implementation of the solution proposed by the present disclosure has been described in detail hereinabove with reference to the preferred embodiments, however, it will be understood by those skilled in the art that various modifications and adaptations can be made to the specific embodiments described above and that various combinations of the technical features, structures proposed by the present disclosure can be made without departing from the scope of the present disclosure, which is defined by the appended claims.

Claims

1. The application of repaglinide is characterized in that the repaglinide is used for preparing medicines for resisting escherichia coli, candida albicans, bacillus subtilis and staphylococcus aureus.