Search

Acquired immune deficiency syndrome (AIDS) is a fatal disease caused by human immunodeficiency virus (HIV). Although 23 different drugs have been available, the treatment of AIDS remains challenging because the virus mutates very quickly which can lead to drug resistance. Therefore, predicting drug resistance before treatment is crucial for individual treatments. Here, based on HIV target protein sequence information, we analyzed 21-drug resistance caused by mutated residues using machine learning (ML) methods. To transform target sequences into numeric vectors, seven physicochemical properties were used, which can well represent the interacting characteristics of target proteins. Then, principal component analysis (PCA) method was adopted to reduce the feature dimensionality. Random forest (RF) and support vector machine (SVM) based on three different kernel functions, including linear, polynomial and radial basis function (RBF), were all employed. By comparisons, we found that RBF-based SVM method gives a comparative performance with RF model. Further, we added the weight information to RBF-based SVM method by four different weight evaluation methods of RF, eXtreme Gradient Boosting (XGB), CfsSubsetEval and ReliefFAttributeEval, respectively. Results show that the RF-weighted RBF-based SVM yield the superior performance and 13 out of 21 drug models provide the correlation coefficients (R 2 ) over 0.8 and 3 of them are higher than 0.9. Finally, position-specific importance analysis indicates that most of the mutation residues with high RF weight scores are proved to be closely related with drug resistance, which has been revealed in previous reports. Overall, we can expect that this method can be a supplementary tool for predicting HIV drug resistance for newly discovered mutations. Here, based on HIV target protein sequence information, we analyzed 21-drug resistance caused by mutated residues using machine learning (ML) methods by fusing the weight information of different mutation positions., (© 2021. The Author(s), under exclusive licence to Springer Nature Switzerland AG.)

Background: Computational prediction of inhibition efficiency (IE) for inhibitor molecules is a crucial supplementary way to design novel molecules that can efficiently inhibit corrosion onto metallic surfaces., Purpose: Here we are dedicated to developing a new machine learning-based predictor for the inhibition efficiency (IE) of benzimidazole derivatives., Methods: First, a comprehensively numerical representation was given on inhibitor molecules from all aspects of energy, electronic, topological, physicochemical and spatial properties based on 3-D structures and 150 valid structural descriptors were obtained. Then, a thorough investigation of these structural descriptors was implemented. The multicollinearity-based clustering analysis was performed to remove the linear correlated feature variables, so 47 feature clusters were produced. Meanwhile, Gini importance by random forest (RF) was used to further measure the contributions of the descriptors in each cluster and 47 non-linear descriptors were selected with the highest Gini importance score in the corresponding cluster. Further, considering the limited number of available inhibitors, different feature subsets were constructed according to the Gini importance score ranking list of 47 descriptors., Results: Finally, support vector machine (SVM) models based on different feature subsets were tested by leave-one-out cross validation. Through comparisons, the optimal SVM model with the top 11 descriptors was achieved based on Poly kernel. This model yields a promising performance with the correlation coefficient (R) and root-mean-square error (RMSE) of 0.9589 and 4.45, respectively, which indicates that the method proposed by us gives the best performance for the current data., Conclusion: Based on our model, 6 new benzimidazole molecules were designed and their IE values predicted by this model indicate that two of them have high potential as outstanding corrosion inhibitors.

In pharmaceutical treatment, many non-cardiac drugs carry the risk of prolonging the QT interval, which can lead to fatal cardiac complications such as torsades de points (TdP). Although the unexpected blockade of ion channels has been widely considered to be one of the main reasons for affecting the repolarization phase of the cardiac action potential and leading to QT interval prolongation, the lack of knowledge regarding chemical structures in drugs that may induce the prolongation of the QT interval remains a barrier to further understanding the underlying mechanism and developing an effective prediction strategy. In this study, we thoroughly investigated the differences in chemical structures between QT-prolonging drugs and drugs with no drug-induced QT prolongation (DIQT) concerns, based on the Drug-Induced QT Prolongation Atlas (DIQTA) dataset. Three categories of structural alerts (SAs), namely amines, ethers, and aromatic compounds, appeared in large quantities in QT-prolonging drugs, but rarely in drugs with no DIQT concerns, indicating a close association between SAs and the risk of DIQT. Moreover, using the molecular descriptors associated with these three categories of SAs as features, the structure–activity relationship (SAR) model for predicting the high risk of inducing QT interval prolongation of marketed drugs achieved recall rates of 72.5% and 80.0% for the DIQTA dataset and the FDA Adverse Event Reporting System (FAERS) dataset, respectively. Our findings may promote a better understanding of the mechanism of DIQT and facilitate research on cardiac adverse drug reactions in drug development. [ABSTRACT FROM AUTHOR]

Predicting RNA solvent accessibility using only primary sequence data can be regarded as sequence-based prediction work. Currently, the established studies for sequence-based RNA solvent accessibility prediction are limited due to the available number of datasets and black box prediction. To improve these issues, we first expanded the available RNA structures and then developed a sequence-based model using modified attention layers with different receptive fields to conform to the stem–loop structure of RNA chains. We measured the improvement with an extended dataset and further explored the model's interpretability by analysing the model structures, attention values and hyperparameters. Finally, we found that the developed model regarded the pieces of a sequence as templates during the training process. This work will be helpful for researchers who would like to build RNA attribute prediction models using deep learning in the future. [ABSTRACT FROM AUTHOR]

Teratogenicity is one of the main concerns in clinical medications of pregnant women. Prescription of antiseizure medications (ASMs) in women with epilepsy during pregnancy may cause teratogenic effects on the fetus. Although large scale epilepsy pregnancy registries played an important role in evaluating the teratogenic risk of ASMs, for most ASMs, especially the newly approved ones, the potential teratogenic risk cannot be effectively assessed due to the lack of evidence. In this study, the analyses are performed on any medication, with a focus on ASMs. We curated a list containing the drugs with potential teratogenicity based on the US Food and Drug Administration (FDA)-approved drug labeling, and established a support vector machine (SVM) model for detecting drugs with high teratogenic risk. The model was validated by using the post-marketing surveillance data from US FDA Spontaneous Adverse Events Reporting System (FAERS) and applied to the prediction of potential teratogenic risk of ASMs. Our results showed that our proposed model outperformed the state-of-art approaches, including logistic regression (LR), random forest (RF) and extreme gradient boosting (XGBoost), when detecting the high teratogenic risk of drugs (MCC and recall rate were 0.312 and 0.851, respectively). Among 196 drugs with teratogenic potential reported by FAERS, 136 (69.4%) drugs were correctly predicted. For the eight commonly used ASMs, 4 of them were predicted as high teratogenic risk drugs, including topiramate, phenobarbital, valproate and phenytoin (predicted probabilities of teratogenic risk were 0.69, 0.60 0.59, and 0.56, respectively), which were consistent with the statement in FDA-approved drug labeling and the high reported prevalence of teratogenicity in epilepsy pregnancy registries. In addition, the structural alerts in ASMs that related to the genotoxic carcinogenicity and mutagenicity, idiosyncratic adverse reaction, potential electrophilic agents and endocrine disruption were identified and discussed. Our findings can be a good complementary for the teratogenic risk assessment in drug development and facilitate the determination of pharmacological therapies during pregnancy. [ABSTRACT FROM AUTHOR]

In clinical cancer research, it is a hot topic on how to accurately stratify patients based on genomic data. With the development of next-generation sequencing technology, more and more types of genomic features, such as mRNA expression level, can be used to distinguish cancer patients. Previous studies commonly stratified patients by using a single type of genomic features, which can only reflect one aspect of the cancer. In fact, multiscale genomic features will provide more information and may be helpful for clinical prediction. In addition, most of the conventional machine learning algorithms use a handcrafted gene set as features to construct models, which is generally selected by a statistical method with an arbitrary cut-off, e.g., p value < 0.05. The genes in the gene set are not necessarily related to the cancer and will make the model unreliable. Therefore, in our study, we thoroughly investigated the performance of different machine learning methods on stratifying breast cancer patients with a single type of genomic features. Then, we proposed a strategy, which can take into account the degree of correlation between genes and cancer patients, to identify the features from mRNAs and microRNAs, and evaluated the performance of the models with the new combined features of the multiscale genomic features. The results showed that, compared with the models constructed with a single type of features, the models with the multiscale genomic features generated by our proposed method achieved better performance on stratifying the ER status of breast cancer patients. Moreover, we found that the identified multiscale genomic features were closely related to the cancer by gene set enrichment analysis, indicating that our proposed strategy can well reflect the biological relevance of the genes to breast cancer. In conclusion, modelling with multiscale genomic features closely related to the cancer not only can guarantee the prediction performance of the models but also can effectively provide candidate genes for interpreting the mechanisms of cancer. [ABSTRACT FROM AUTHOR]

Background: The aim of gene expression-based clinical modelling in tumorigenesis is not only to accurately predict the clinical endpoints, but also to reveal the genome characteristics for downstream analysis for the purpose of understanding the mechanisms of cancers. Most of the conventional machine learning methods involved a gene filtering step, in which tens of thousands of genes were firstly filtered based on the gene expression levels by a statistical method with an arbitrary cutoff. Although gene filtering procedure helps to reduce the feature dimension and avoid overfitting, there is a risk that some pathogenic genes important to the disease will be ignored. Results: In this study, we proposed a novel deep learning approach by combining a convolutional neural network with stationary wavelet transform (SWT-CNN) for stratifying cancer patients and predicting their clinical outcomes without gene filtering based on tumor genomic profiles. The proposed SWT-CNN overperformed the state-of-art algorithms, including support vector machine (SVM) and logistic regression (LR), and produced comparable prediction performance to random forest (RF). Furthermore, for all the cancer types, we firstly proposed a method to weight the genes with the scores, which took advantage of the representative features in the hidden layer of convolutional neural network, and then selected the prognostic genes for the Cox proportional-hazards regression. The results showed that risk stratifications can be effectively improved by using the identified prognostic genes as feature, indicating that the representative features generated by SWT-CNN can well correlate the genes with prognostic risk in cancers and be helpful for selecting the prognostic gene signatures. Conclusions: Our results indicated that gene expression-based SWT-CNN model can be an excellent tool for stratifying the prognostic risk for cancer patients. In addition, the representative features of SWT-CNN were validated to be useful for evaluating the importance of the genes in the risk stratification and can be further used to identify the prognostic gene signatures. [ABSTRACT FROM AUTHOR]

Identifying protein-RNA binding residues is essential for understanding the mechanism of protein-RNA interactions. So far, rigid distance thresholds are commonly used to define protein-RNA binding residues. However, after investigating 182 non-redundant protein-RNA complexes, we find that it would be unsuitable for a certain amount of complexes since the distances between proteins and RNAs vary widely. In this work, a novel definition method was proposed based on a flexible distance cutoff. This method can fully consider the individual differences among complexes by setting a variable tolerance limit of protein-RNA interactions, i.e. the double minimum-distance by which different distance thresholds are achieved for different complexes. In order to validate our method, a comprehensive comparison between our flexible method and traditional rigid methods was implemented in terms of interface structure, amino acid composition, interface area and interaction force, etc. The results indicate that this method is more reasonable because it incorporates the specificity of different complexes by extracting the important residues lost by rigid distance methods and discarding some redundant residues. Finally, to further test our double minimum-distance definition strategy, we developed a classifier to predict those binding sites derived from our new method by using structural features and a random forest machine learning algorithm. The model achieved a satisfactory prediction performance and the accuracy on independent data sets reaches to 85.0%. To the best of our knowledge, it is the first prediction model to define positive and negative samples using a flexible cutoff. So the comparison analysis and modeling results have demonstrated that our method would be a very promising strategy for more precisely defining protein-RNA binding sites. [ABSTRACT FROM AUTHOR]

Hepatocellular carcinoma (HCC) is currently still a major factor leading to death, lacking of reliable biomarkers. Therefore, deep understanding the pathogenesis for HCC is of great importance. The emergence of circular RNA (circRNA) provides a new way to study the pathogenesis of human disease. Here, we employed the prediction tool to identify circRNAs based on RNA-seq data. Then, to investigate the biological function of the circRNA, the candidate circRNAs were associated with the protein-coding genes (PCGs) by GREAT. We found significant candidate circRNAs expression alterations between normal and tumor samples. Additionally, the PCGs associated with these candidate circRNAs were also found have discriminative expression patterns between normal and tumor samples. The enrichment analysis illustrated that these PCGs were predominantly enriched for liver/cardiovascular-related diseases such as atherosclerosis, myocardial ischemia and coronary heart disease, and participated in various metabolic processes. Together, a further network analysis indicated that these PCGs play important roles in the regulatory and the PPI network. Finally, we built a classification model to distinguish normal and tumor samples by using candidate circRNAs and their associated genes, respectively. Both of them obtained satisfactory results (~ 0.99 of AUC for circRNA and PCG). Our findings suggested that the circRNA could be a critical factor in HCC, providing a useful resource to explore the pathogenesis of HCC. [ABSTRACT FROM AUTHOR]

Motivated by timesaving when dealing with the large-scale calculation for data modeling in parallel with multiple CPU cores or machines and result comparison, PML was designed in this study. PML has the ability to do dimension reduction and grid search in parallel, support both classification and regression, and could generate HTML pages as output for results comparison. Written in PERL, PML is compatible with Windows and Linux. This open source software is free available together with the code, manual, examples and license from http://cic.scu.edu.cn/pml or https://github.com/limlcic/PML . [ABSTRACT FROM AUTHOR]

Protein–protein interactions play fundamental roles in almost all biological processes. Determining the protein–protein binding affinity has been recognized not only as an important step but also as a challenging task for further understanding of the molecular mechanism and the modeling of the biological systems. Unlike the traditional methods like empirical scoring algorithms and molecular dynamic which are time consuming, we developed a fast and reliable machine learning method for the prediction of protein–protein binding affinity. Based on diverse protein–protein interface features calculated using commonly used available tools, 432 features were obtained to represent hydrogen bond, Van der Waals force, hydrophobic interaction, electrostatic force, interface shape and configuration and allosteric effect. Considering the limited number of the available structures and affinity-known protein complexes, in order to avoid overfitting and remove noises in the feature set, feature importance evaluation was implemented and 154 optimal features were selected, then the prediction model based on random forest (RF) was constructed. We demonstrate that the RF model yields promising results and the predictive power of our method is better than other existing methods. Using leave-one-out cross-validation, our model gives a correlation coefficient ( r ) of 0.708 on the whole benchmark dataset of 133 complexes and a high r of 0.806 on the validated set of 53 samples. When performing the same two independent datasets, our method outperforms other two methods and achieves a high r of 0.793 and 0.907 respectively. All results indicate that our method can be a useful implement in determining protein–protein binding affinity. [ABSTRACT FROM AUTHOR]

[Display omitted] • FTIR spectra combined with chemometrics methods for rapid quality control of C.sinensis. • Multiple samples over 3500 were collected which could be a great supplement to fingerprint mapping studies of C.sinensis. • Comprehensive cross-sectional comparison between conventional machine learning methods and CNN combined with different data pretreatment methods. • Robust multi-class models constructed for prediction of C.sinensis's part and region. Cordyceps sinensis(C.sinensis) is a precious traditional Chinese herbal medicine with a wide variety of beneficial pharmacological effects. In recent years, due to the limitation of natural conditions and the over-harvesting, natural C.sinensis resources have been increasingly scarce and far from meeting the market demand. The phenomenon of counterfeit and substandard in the market is also becoming more and more serious. Thus the establishment of an identification method with good accuracy, high sensitivity and specificity is imminent. For this purpose, 3559 samples of C.sinensis were collected from different sources (part, region and phase) and we proposed the use of FTIR spectroscopy combined with chemometrics methods to construct robust and effective multi-class discriminant models. Cross-sectional comparison was performed between conventional machine learning methods (SVM, PLS-DA, LDA) and convolution neural network (CNN) under different classification tasks. Matthews correlation coefficient (MCC) was chosen as the optimized metrics for imbalanced classification problem. The best model for prediction of C.sinensis's part was SVM combined with SNV as the pretreatment method (MCC was 0.8918 on cross-validation dataset and 0.9017 on independent testing dataset) and LDA with no pretreatment for prediction of C.sinensis's region (MCC was 0.7721 on cross-validation dataset and 0.7332 on independent testing dataset). Classification of C.sinensis's phase was less predictive (MCC was around 0.25) but was significantly superior to random prediction. The model performance of CNN is comparable to that of traditional machine learning methods and increasing training sample size might have the potential to further improve the model performance of CNN. Our results demonstrated that FTIR spectra combined with an optimized multivariable discriminant model could be used as an effective and sensitive method for rapid quality control of C.sinensis. [ABSTRACT FROM AUTHOR]

Abstract: MicroRNA (miRNA) is the negative regulator of gene expression, also known as guide strand of transient miRNA:miRNA* duplex. It is critical in maintaining the normal physiological processes such as development, differentiation, and apoptosis in many organisms. With increasing miRNA data, it is desirable to design methods to identify guide strand based on machine learning algorithms. In this study, the random forest models based on local sequence–structure features were proposed to identify miRNA in four species. The accuracies achieved were 86.51% for Homo sapiens, 81.66% for Ornithorhynchus anatinus, 82.33% for Mus musculus and 85.71% for Schmidtea mediterranea, respectively. Furthermore, the important analysis of feature elements was carried out by using the conditional feature importance strategy. The analysis results revealed that most of the significant elements were related to guanine–cytosine (GC) base pair. We believed that our method could be beneficial to annotate the function of miRNA and help the further understanding of the RNA interference mechanism. [Copyright &y& Elsevier]

Recently, multiple researches of solvent accessibility on protein are well established. However, research on solvent accessibility of RNA faces a few challenges such as instability and diversity of RNA tertiary structure. Nowadays, no study has examined the predicting performance from different datasets built from multi-models and multi-attributes, but it is an important part of measuring the overall performance of modeling. Therefore, we performed a comprehensive comparison for predicting RNA solvent accessibility based on two datasets. 15,923 (12,229 ​+ ​3694) samples and 512(336 ​+ ​176) attributes were generated for attribute selection, finally 336 models were built for predicting. 12 modeling methods and 2 attribute selection methods were used for modeling and evaluating. This work provided a strategy for getting stable expectation when predicting RNA solvent accessibility. These results would be useful in further experimental or computational design for RNA solvent accessibility predicting and we hope this work could help the related researches which need a workflow for stable prediction. • A strategy was provided for getting stable expectation when predicting RNA solvent accessibility. • Two datasets were used and 6 groups of attributes were generated for attribute selecting. • 336 models consisted by 2 attribute selection methods and 12 machine learning methods were used for modeling and evaluating. • Models using Random Forest attribute selection exhibited the contribution from none-sequential attributes. [ABSTRACT FROM AUTHOR]