Prediction of human intestinal absorption properties based on artificial intelligence / 中国药科大学学报

@#Human intestinal absorption (HIA) is a crucial indicator for measuring the oral bioavailability of drugs.This study aims to use artificial intelligence methods to predict and evaluate the HIA of drugs in the early stages of drug discovery, thus accelerating the drug discovery process and reducing costs.This study used MOE''s 2D, 3D descriptors, and ECFP4 (extended connectivity fingerprints) to characterize the molecules and established eight models, including support vector machine (SVM), random forest (RF), and deep neural network (DNN).The results showed that the SVM model constructed using a combination of 2D, 3D descriptors and ECFP4 fingerprints was the optimal model according to comprehensive evaluation of various evaluation indicators.The area under the receiver operating characteristic curve (AUC), Matthews correlation coefficient, and Kappa coefficient of the optimal model were 0.94, 0.75, and 0.74, respectively.In conclusion, this study established a robust and generalizable machine learning model for predicting HIA properties, which can provide guidance for early molecular screening and the study of pharmacokinetic properties of drugs.

Research on the automatic classification system of pterygium based on deep learning / 国际眼科杂志(Guoji Yanke Zazhi)

@#AIM: To evaluate the application value of the automatic classification and diagnosis system of pterygium based on deep learning.<p>METHODS: A total of 750 images of normal, observational and operative anterior sections of pterygium were collected from the Affiliated Eye Hospital of Nanjing Medical University between May 2020 and April 2021. Seven triclassification models were respectively trained with original data set and enhanced data set. Totally 470 clinical images were tested, and the generalization ability of the model before and after data enhancement was compared to determine the best model for the automatic classification system of pterygium.<p>RESULTS:The average sensitivity, specificity and AUC of the best model trained on the original data set were 92.55%, 96.86% and 94.70% respectively. After data was enhanced, the sensitivity, specificity and AUC of different models were increased by 3.7%, 1.9% and 2.7% on average. The sensitivity, specificity and AUC of the EfficientNetB7 model trained on the enhanced data set were 93.63%, 97.34% and 95.47% respectively.<p>CONDLUSION: The EfficientNetB7 model, which was trained on the enhanced data set, achieves the best classification result and can be used in the automatic classification system of pterygium.This automatic classification system can diagnose diseases about pterygium better, and it is expected to be an effective screening tool for primary medical care. It also provides reference for the research on the refinement and grading of pterygium.

Construction and application of a random forest-based classification model for DNA double-strand break induced by ionizing radiation / 中华放射医学与防护杂志

Objective:To construct a random forest classification model of DNA double strand breaks (DSB) induced by ionizing radiation and investigate the genome-wide distribution of DSB.Methods:The GRCh38 reference genome was divided into 50 kilobase fragments. Then these genomic fragments were separated into low-level or high-level regions of ionizing radiation-induced DSB according to the sequencing data of MCF-7 cells. The data of eight epigenetic features were used as input. Two thirds of the data were randomly assigned to the training set, and the rest of the data was assigned to the test set. A random forest classification model with 100 decision trees was constructed. The importance of epigenetic features in the classification model was analyzed and displayed.Results:The accuracy score of the random forest classification model on the test set was 99.4%, the precision score was 98.9% and the recall score was 99.9%. The area under the receiver operating characteristic curve was 0.994. Among the eight epigenetic features, H3K36me3 and DNase markers were the most important variables. The enrichments of the two markers in DSB high-level regions were much higher than those in DSB low-level regions.Conclusions:The random forest classification model could precisely predict the genome-wide levels of DSB induced by ionizing radiation in the 50 kilobase window based on epigenetic features. Analysis revealed that these DSB might primarily distribute in the actively transcribed sites in the genome.

Application of immune cell infiltration in the diagnosis and prognosis of non-small cell lung cancer / 生物工程学报

Immune cell infiltration is of great significance for the diagnosis and prognosis of cancer. In this study, we collected gene expression data of non-small cell lung cancer (NSCLC) and normal tissues included in TCGA database, obtained the proportion of 22 immune cells by CIBERSORT tool, and then evaluated the infiltration of immune cells. Subsequently, based on the proportion of 22 immune cells, a classification model of NSCLC tissues and normal tissues was constructed using machine learning methods. The AUC, sensitivity and specificity of classification model built by random forest algorithm reached 0.987, 0.98 and 0.84, respectively. In addition, the AUC, sensitivity and specificity of classification model of lung adenocarcinoma and lung squamous carcinoma tissues constructed by random forest method 0.827, 0.75 and 0.77, respectively. Finally, we constructed a prognosis model of NSCLC by combining the immunocyte score composed of 8 strongly correlated features of 22 immunocyte features screened by LASSO regression with clinical features. After evaluation and verification, C-index reached 0.71 and the calibration curves of three years and five years were well fitted in the prognosis model, which could accurately predict the degree of prognostic risk. This study aims to provide a new strategy for the diagnosis and prognosis of NSCLC based on the classification model and prognosis model established by immune cell infiltration.

Un modelo predictivo de preeclampsia a partir de datos clínicos y bioquímicos / A Predictive Model of Preeclampsia from Clinical and Biochemical Data

RESUMEN Introducción: La preeclampsia es uno de los síndromes en mujeres embarazadas que afecta al menos 3 - 8 por ciento de todos los embarazos. Objetivo: Desarrollar un modelo predictivo de preeclampsia a partir del estado redox en embarazadas, que clasifique a las mujeres pertenecientes a los grupos de gestantes preeclámpticas y gestantes sanas. Métodos: Se realizó un estudio analítico transversal. Los parámetros bioquímicos y clínicos se evaluaron utilizando el análisis de componentes principales para identificar las variables más influyentes en la aparición de preeclampsia. Los seleccionados como las variables más importantes fueron evaluados por el análisis discriminante lineal de Fisher. Resultados: El análisis de componentes principales determinó la varianza del set de datos, mostrando la relación con los procesos de peroxidación lipídica, metabolismo de proteínas, daño a tejidos y microangiopático, considerados factores en la fisiopatología de la preeclampsia. Las variables más influyentes fueron usadas para modelar una función discriminante capaz de clasificar gestantes sanas y preeclámpticas. El valor de Lambda de Wilks y el alto autovalor asociado a la función discriminante muestran el poder discriminante del modelo. La ecuación obtenida fue validada con el método Leave one out y reveló un excelente poder clasificatorio del mismo. Conclusiones: El modelo predictivo puede ser considerado como apropiado para clasificar los casos de preeclampsia, y muestran a los biomarcadores como buenos candidatos para la clasificación y como potenciales indicadores predictivos de preeclampsia(AU)

ABSTRACT Introduction: Preeclampsia is one of the syndromes in pregnant women that affects at least 3 - 8 percent of all pregnancies. Objective: To develop a predictive model of preeclampsia from the redox state in pregnant women, which allows to classify them in groups of preeclamptic pregnant women and healthy pregnant women. Methods: A cross-sectional analytical study was performed. Biochemical and clinical parameters were evaluated using principal component analysis to identify the most influential variables in the occurrence of preeclampsia. Those selected as the most important variables were evaluated by Fisher's linear discriminant analysis. Results: The main component analysis determined the variance of the data set, showing the relationship with lipid peroxidation processes, protein metabolism, tissue damage and microangiopathy, considered factors in the pathophysiology of preeclampsia. The most influential variables were used to model a discriminant function capable of classifying healthy and preeclamptic pregnant women. Wilks Lambda value and the high eigenvalue associated with the discriminant function show the discriminant power of the model. The equation obtained was validated with the Leave one out method and revealed excellent classifying power. Conclusions: The predictive model can be considered as appropriate to classify pre-eclampsia cases, and to show biomarkers as good candidates for classification and as potential predictive indicators of pre-eclampsia(AU)

Study of a screening system for mild cognitive impairment based on machine learning model / 上海交通大学学报（医学版）

Objective:To evaluate the reliability and validity of a computerized cognitive assessment system designed for screening mild cognitive impairment (MCI), and compare the screening accuracy among constructed different machine learning classification models. Methods:A group of random stratified samples of over 55 years old residents in the communities, nursing homes and memory-clinics from Shanghai and Henan were selected to assess their cognitive status using Montreal Cognitive Assessment (MoCA) by well-trained investigators. The reliability and validity were assessed by intrinsic consistency analysis and factor analysis, respectively. Taking the results of MoCA as standards, four machine learning classification algorithms, i.e., naïve Bayesian classification model, random forest classifier, Logistic regression classifier, and K-nearest neighbor classifier, were compared in accuracy and area under curve (AUC). Results:A total of 359 participants were included, the median age of whom was 63 years old. And 82.80% of them were secondary school graduates or below. According to the results of MoCA, 147 of them might be MCI. The Cronbach's α and KMO of this system were 0.84 and 0.78, respectively; Bartlett's sphericity test was significant (P<0.05); thirteen common factors could explain 75.10% of the system. The best classification model was naïve Bayesian classification model, and its accuracy and AUC were 88.05% and 0.941, respectively. Conclusion:The new designed computerized cognitive assessment system has been proved to be reliable and valid. The naïve Bayesian classification model has good classification accuracy.

Predicting the Number of People for Meals of an Institutional Foodservice by Applying Machine Learning Methods: S City Hall Case

Predicting the number of meals in a foodservice organization is an important decision-making process that is essential for successful food production, such as reducing the amount of residue, preventing menu quality deterioration, and preventing rising costs. Compared to other demand forecasts, the menu of dietary personnel includes diverse menus, and various dietary supplements include a range of side dishes. In addition to the menus, diverse subjects for prediction are very difficult problems. Therefore, the purpose of this study was to establish a method for predicting the number of meals including predictive modeling and considering various factors in addition to menus which are actually used in the field. For this purpose, 63 variables in eight categories such as the daily available number of people for the meals, the number of people in the time series, daily menu details, weekdays or seasons, days before or after holidays, weather and temperature, holidays or year-end, and events were identified as decision variables. An ensemble model using six prediction models was then constructed to predict the number of meals. As a result, the prediction error rate was reduced from 10%~11% to approximately 6~7%, which was expected to reduce the residual amount by approximately 40%.

Network mechanism of effective constituents from the compound Yizhihao against influenza / 中国药理学与毒理学杂志

Influenza caused by influenza virus,seriously threaten human life and health.Drug treatment is one of the effective measurement. However, there are only two classes of drugs, one class is M2 blockers and another is neuraminidase (NA)inhibitors. The recent antiviral surveillance studies reported a global significant increase in M2 blocker resistance among influenza viruses, and the resistant virus strains against NA inhibitor are also reported in clinical treatment.Therefore thediscovery of new medicines with low resistance has become very urgent.As all known,traditional medicines with multi-target features and network mechanism often possess low resistance. Compound Yizhihao, which consists of radix isatidis,folium isatidis,Artemisia rupestris,is one of the famous traditional medicine for influenza treatment in China, however its mechanism of action against influenza is unclear. In this study, the multiple targets related with influenza disease and the known chemical constituents from Compound Yizhihao were collected, and multi-target QSAR (mt-QSAR) classification models were developed by Na?ve Bayesian algorithm and verified by various datasets. Then the classification models were applied to predict the effective constituents and their drug targets.Finally,the constituent-target-pathway network was constructed,which revealed the effective constituents and their network mechanism in Compound Yizhihao. This study will lay important basis for the clinical uses for influenza treatment and for the further research and development of the effective constituents.

Network pharmacology study of the effective constituents in the Compound Yizhihao against influenza disease / 药学学报

Compound Yizhihao, consists of Radix isatidis, Folium isatidis, Artemisia rupestris, has a significant therapeutic effect on the treatment of influenza and fever. However, the mechanism of its action is still unclear. In this investigation, we collected the key target molecule of influenza disease and the chemical constituents of Compound Yizhihao, and developed Naïve Bayesian classification models based on the input molecular fingerprints and molecule descriptors. The built models were further applied to construct classifiers for predicting the effective constituents. We used the professional network-building software to build the constituent-target network and target-pathway network, which revealed the network pharmacology of the effective constituents in Compound Yizhihao. It will contribute to the further research of mechanism of Compound Yizhihao.

Development of a Stress Classification Model Using Deep Belief Networks for Stress Monitoring / 대한의료정보학회지

OBJECTIVES: Stress management is related to public healthcare and quality of life; an accurate stress classification method is necessary for the design of stress monitoring systems. Therefore, the goal of this study was to design a novel stress classification model using a deep learning method. METHODS: In this paper, we present a stress classification model using the dataset from the sixth Korea National Health and Nutrition Examination Survey conducted from 2013 to 2015 (KNHANES VI) to analyze stress-related health data. Statistical analysis was performed to identify the nine features of stress detection, and we evaluated the performance of the proposed stress classification by comparison with several stress detection models. The proposed model was also evaluated using Deep Belief Networks (DBN). RESULTS: We designed profiles depending on the number of hidden layers, nodes, and hyper-parameters according to the loss function results. The experimental results showed that the proposed model achieved an accuracy and a specificity of 66.23% and 75.32%, respectively. The proposed DBN model performed better than other classification models, such as support vector machine, naive Bayesian classifier, and random forest. CONCLUSIONS: The proposed model in this study was demonstrated to be effective in classifying stress detection, and in particular, it is expected to be applicable for stress prediction in stress monitoring systems.

A study on Bayesian classification model for common respiratory infectious diseases / 预防医学

Objective To provide diagnostic clue for the investigation and laboratory examination in outbreak of common respiratory infectious diseases using a computer -aided classification model.Methods The variables were extracted from medical literature,case data of infectious diseases,reports of outbreaks such as symptoms and signs,abnormal lab test results,epidemiologic features,the incidence rates of the infectious diseases.Then a classification model was constructed using Naive Bayesian classifier and SAS 9.1 .3 Data from eight historical outbreaks of respiratory infectious diseases were used to test the model.Results Among eight outbreaks,the discriminate probability of diagnosing a disease correctly by ranking it first on the output lists of the model was 53.85%.The sensitivity was 53.85%,and specificity was 1 00.00%, and +LR was from 5.73 to ∞.The discriminant probability of diagnosing a disease correctly by ranking it within the three most probable diseases on these lists was 98.34%.The sensitivity was 98.34% and the specificity was 82.1 4%,and +LR was from 1 .26 to ∞.Conclusion A Bayesian classification model could be applied to classification and discriminant of common respiratory infectious diseases,and could improve the ability for early diagnosis of the outbreak caused by respiratory infectious diseases.

A maximum margin-based kernel width estimator and its application to the response to neoadjuvant chemotherapy

INTRODUCTION: Function induction problems are frequently represented by affinity measures between the elements of the inductive sample set, and kernel matrices are a well-known example of affinity measures. METHODS: The objective of the present work is to obtain information about the relations between data from a calculated kernel matrix by initially assuming that those geometric relations are consistent with known labels. To assess the relation between the data structure and the labels, a classifier based on kernel density estimation (KDE) was used. The performance of the selected width using the method presented in this paper was compared to the performance of a method described in the literature; the literature method was based on minimizing error minimization and balancing bias and variance. The main case study, which was to predict the response to neoadjuvant chemotherapy treatment, consists of evaluating whether a set of training data from genomic expression data from breast tumors and the genomic expression from the tumor of one patient can be used to determine whether there will be a pathological complete response. RESULTS: For the tested databases, the proposed method showed statistically equivalent results with the literature method; however, in some cases, the proposed method had a better overall performance when considering both large and small classes. CONCLUSION: The results demonstrate the feasibility of selecting models by directly calculating densities and the geometry from the class separation.

Research on Identification Model of Chinese Herbal Medicine by Texture Feature Parameter of Transverse Section Image / 世界科学技术-中医药现代化

This study was aimed to establish the classification method of Chinese herbal medicine based on feature parameters extracted from images of herbal transverse section, in order to explore the feasibility of automatic identi-fication method of herbal medicine. The extracted 26 parameters of 18 herbal medicine images by gray-level co-oc-currence matrix and grayscale gradient matrix were used as the basic data set. And the minimum covariance determi-nant (MCD) was used to delete the outliers. A total of 18 identification models were established using the native Bayes method and BP neural network methods. The results showed that the average correct rates of models were 90%. It was concluded the feasibility of using these models in the establishment of the automatic identification method of herbal medicines. It provided new technologies for the quantitative, scientific and objective identification of Chinese herbal medicine.