Identification of metal ion-binding sites in RNA structures using deep learning method.

Metal ion is an indispensable factor for the proper folding, structural stability and functioning of RNA molecules. However, it is very difficult for experimental methods to detect them in RNAs. With the increase of experimentally resolved RNA structures, it becomes possible to identify the metal ion-binding sites in RNA structures through in-silico methods. Here, we propose an approach called Metal3DRNA to identify the binding sites of the most common metal ions (Mg2+, Na+ and K+) in RNA structures by using a three-dimensional convolutional neural network model. The negative samples, screened out based on the analysis for binding surroundings of metal ions, are more like positive ones than the randomly selected ones, which are beneficial to a powerful predictor construction. The microenvironments of the spatial distributions of C, O, N and P atoms around a sample are extracted as features. Metal3DRNA shows a promising prediction power, generally surpassing the state-of-the-art methods FEATURE and MetalionRNA. Finally, utilizing the visualization method, we inspect the contributions of nucleotide atoms to the classification in several cases, which provides a visualization that helps to comprehend the model. The method will be helpful for RNA structure prediction and dynamics simulation study. Availability and implementation: The source code is available at https://github.com/ChunhuaLiLab/Metal3DRNA.

RNAfcg: RNA Flexibility Prediction Based on Topological Centrality and Global Features.

The dynamics of RNAs are related intimately to their functions. Molecular flexibility, as a starting point for understanding their dynamics, has been utilized to predict many characteristics associated with their functions. Since the experimental measurement methods are time-consuming and labor-intensive, it is urgently needed to develop reliable theoretical methods to predict RNA flexibility. In this work, we develop an effective machine learning method, RNAfcg, to predict RNA flexibility, where the Random Forest (RF) is trained by features including the topological centralities, flexibility-rigidity index, and global characteristics first introduced by us, as well as some traditional sequence and structural features. The analyses show that the three types of features introduced first have significant contributions to RNA flexibility prediction, among which the topological type contributes the most, which indicates the importance of structural topology in determining RNA flexibility. The performance comparison indicates that RNAfcg outperforms the state-of-the-art machine learning methods and the commonly used Gaussian Network Model (GNM) models, achieving a much higher Pearson correlation coefficient (PCC) of 0.6619 on the test data set. This work is helpful for understanding RNA dynamics and can be used to predict RNA function information. The source code is available at https://github.com/ChunhuaLab/RNAfcg/.

Prediction of Protein Allosteric Sites with Transfer Entropy and Spatial Neighbor-Based Evolutionary Information Learned by an Ensemble Model.

Allostery is one of the most direct and efficient ways to regulate protein functions. The diverse allosteric sites make it possible to design allosteric modulators of differential selectivity and improved safety compared with those of orthosteric drugs targeting conserved orthosteric sites. Here, we develop an ensemble machine learning method AllosES to predict protein allosteric sites in which the new and effective features are utilized, including the entropy transfer-based dynamic property, secondary structure features, and our previously proposed spatial neighbor-based evolutionary information besides the traditional physicochemical properties. To overcome the class imbalance problem, the multiple grouping strategy is proposed, which is applied to feature selection and model construction. The ensemble model is constructed where multiple submodels are trained on multiple training subsets, respectively, and their results are then integrated to be the final output. AllosES achieves a prediction performance of 0.556 MCC on the independent test set D24, and additionally, AllosES can rank the real allosteric sites in the top three for 83.3/89.3% of allosteric proteins from the test set D24/D28, outperforming the state-of-the-art peer methods. The comprehensive results demonstrate that AllosES is a promising method for protein allosteric site prediction. The source code is available at https://github.com/ChunhuaLab/AllosES.

PMSPcnn: Predicting protein stability changes upon single point mutations with convolutional neural network.

Protein missense mutations and resulting protein stability changes are important causes for many human genetic diseases. However, the accurate prediction of stability changes due to mutations remains a challenging problem. To address this problem, we have developed an unbiased effective model: PMSPcnn that is based on a convolutional neural network. We have included an anti-symmetry property to build a balanced training dataset, which improves the prediction, in particular for stabilizing mutations. Persistent homology, which is an effective approach for characterizing protein structures, is used to obtain topological features. Additionally, a regression stratification cross-validation scheme has been proposed to improve the prediction for mutations with extreme ΔΔG. For three test datasets: Ssym, p53, and myoglobin, PMSPcnn achieves a better performance than currently existing predictors. PMSPcnn also outperforms currently available methods for membrane proteins. Overall, PMSPcnn is a promising method for the prediction of protein stability changes caused by single point mutations.

Insights into Activation Dynamics and Functional Sites of Inwardly Rectifying Potassium Channel Kir3.2 by an Elastic Network Model Combined with Perturbation Methods.

The inwardly rectifying potassium channel Kir3.2, a member of the inward rectifier potassium (Kir) channel family, exerts important biological functions through transporting potassium ions outside of the cell, during which a large-scale synergistic movement occurs among its different domains. Currently, it is not fully understood how the binding of the ligand to the Kir3.2 channel leads to the structural changes and which key residues are responsible for the channel gating and allosteric dynamics. Here, we construct the Gaussian network model (GNM) of the Kir3.2 channel with the secondary structure and covalent interaction information considered (sscGNM), which shows a better performance in reproducing the channel's flexibility compared with the traditional GNM. In addition, the sscANM-based perturbation method is used to simulate the channel's conformational transition caused by the activator PIP2's binding. By applying certain forces to the PIP2 binding pocket, the coarse-grained calculations generate the similar conformational changes to the experimental observation, suggesting that the topology structure as well as PIP2 binding are crucial to the allosteric activation of the Kir3.2 channel. We also utilize the sscGNM-based thermodynamic cycle method developed by us to identify the key residues whose mutations significantly alter the channel's binding free energy with PIP2. We identify not only the residues important for the specific binding but also the ones critical for the allosteric transition coupled with PIP2 binding. This study is helpful for understanding the working mechanism of Kir3.2 channels and can provide important information for related drug design.

Endothelial Progenitor Cells Derived From Wharton's Jelly of Human Umbilical Cord Attenuate Ischemic Acute Kidney Injury by Increasing Vascularization and Decreasing Apoptosis, Inflammation, and Fibrosis.

Ischemia-reperfusion (I/R) injury to the kidney, a major cause of acute renal failure in humans, is associated with a high mortality, and the development of a new therapeutic strategy is therefore highly desirable. In this study, we examined the therapeutic potential of implantation of endothelial progenitor cells (EPCs) isolated from Wharton's jelly of human umbilical cords in the treatment of renal I/R injury in mice. To visualize the localization of the transplanted EPCs, the cells were labeled with Q-tracker before injection into the renal capsule. Mice with renal I/R injury showed a significant increase in blood urea nitrogen and creatinine levels, and these effects were decreased by EPC transplantation. The kidney injury score in the mice with I/R injury was also significantly decreased by EPC transplantation. EPC transplantation increased the microvascular density, and some of the EPCs surrounded and were incorporated into microvessels. In addition, EPC transplantation inhibited the I/R-induced cell apoptosis of endothelial, glomerular, and renal tubular cells, as demonstrated by TUNEL staining, and significantly reduced reactive oxygen species production and the expression of the inflammatory chemokines macrophage inflammatory protein-2 and keratinocyte-derived cytokine, as shown by immunostaining and ELISA. Moreover, EPC transplantation reduced I/R-induced fibrosis, as demonstrated by immunostaining for S100A4, a fibroblast marker, and by Jones silver staining. To our knowledge, this is the first report that transplantation of EPCs from Wharton's jelly of human umbilical cords might provide a novel therapy for ischemic acute kidney injury by promoting angiogenesis and inhibiting apoptosis, inflammation, and fibrosis.