Protein-Protein Interaction Prediction via Structure-Based Deep Learning.

Protein-protein interactions (PPIs) play an essential role in life activities. Many artificial intelligence algorithms based on protein sequence information have been developed to predict PPIs. However, these models have difficulty dealing with various sequence lengths and suffer from low generalization and prediction accuracy. In this study, we proposed a novel end-to-end deep learning framework, RSPPI, combining residual neural network (ResNet) and spatial pyramid pooling (SPP), to predict PPIs based on the protein sequence physicochemistry properties and spatial structural information. In the RSPPI model, ResNet was employed to extract the structural and physicochemical information from the protein three-dimensional structure and primary sequence; the SPP layer was used to transform feature maps to a single vector and avoid the fixed-length requirement. The RSPPI model possessed excellent cross-species performance and outperformed several state-of-the-art methods based either on protein sequence or gene ontology in most evaluation metrics. The RSPPI model provides a novel strategy to develop an AI PPI prediction algorithm.

Vec2image: an explainable artificial intelligence model for the feature representation and classification of high-dimensional biological data by vector-to-image conversion.

Feature representation and discriminative learning are proven models and technologies in artificial intelligence fields; however, major challenges for machine learning on large biological datasets are learning an effective model with mechanistical explanation on the model determination and prediction. To satisfy such demands, we developed Vec2image, an explainable convolutional neural network framework for characterizing the feature engineering, feature selection and classifier training that is mainly based on the collaboration of principal component coordinate conversion, deep residual neural networks and embedded k-nearest neighbor representation on pseudo images of high-dimensional biological data, where the pseudo images represent feature measurements and feature associations simultaneously. Vec2image has achieved better performance compared with other popular methods and illustrated its efficiency on feature selection in cell marker identification from tissue-specific single-cell datasets. In particular, in a case study on type 2 diabetes (T2D) by multiple human islet scRNA-seq datasets, Vec2image first displayed robust performance on T2D classification model building across different datasets, then a specific Vec2image model was trained to accurately recognize the cell state and efficiently rank feature genes relevant to T2D which uncovered potential T2D cellular pathogenesis; and next the cell activity changes, cell composition imbalances and cell-cell communication dysfunctions were associated to our finding T2D feature genes from both population-shared and individual-specific perspectives. Collectively, Vec2image is a new and efficient explainable artificial intelligence methodology that can be widely applied in human-readable classification and prediction on the basis of pseudo image representation of biological deep sequencing data.

Fine-Grained Recognition of Mixed Signals with Geometry Coordinate Attention.

With the advancement of technology, signal modulation types are becoming increasingly diverse and complex. The phenomenon of signal time-frequency overlap during transmission poses significant challenges for the classification and recognition of mixed signals, including poor recognition capabilities and low generality. This paper presents a recognition model for the fine-grained analysis of mixed signal characteristics, proposing a Geometry Coordinate Attention mechanism and introducing a low-rank bilinear pooling module to more effectively extract signal features for classification. The model employs a residual neural network as its backbone architecture and utilizes the Geometry Coordinate Attention mechanism for time-frequency weighted analysis based on information geometry theory. This analysis targets multiple-scale features within the architecture, producing time-frequency weighted features of the signal. These weighted features are further analyzed through a low-rank bilinear pooling module, combined with the backbone features, to achieve fine-grained feature fusion. This results in a fused feature vector for mixed signal classification. Experiments were conducted on a simulated dataset comprising 39,600 mixed-signal time-frequency plots. The model was benchmarked against a baseline using a residual neural network. The experimental outcomes demonstrated an improvement of 9% in the exact match ratio and 5% in the Hamming score. These results indicate that the proposed model significantly enhances the recognition capability and generalizability of mixed signal classification.

Rural Road Extraction in Xiong'an New Area of China Based on the RC-MSFNet Network Model.

High-resolution remote sensing imagery, reaching meter or sub-meter levels, provides essential data for extracting and identifying road information. However, rural roads are often narrow, elongated, and have blurred boundaries, with textures that resemble surrounding environments such as construction sites, vegetation, and farmland. These features often lead to incomplete extraction and low extraction accuracy of rural roads. To address these challenges, this study introduces the RC-MSFNet model, based on the U-Net architecture, to enhance rural road extraction performance. The RC-MSFNet model mitigates the vanishing gradient problem in deep networks by incorporating residual neural networks in the downsampling stage. In the upsampling stage, a connectivity attention mechanism is added after dual convolution layers to improve the model's ability to capture road completeness and connectivity. Additionally, the bottleneck section replaces the traditional dual convolution layers with a multi-scale fusion atrous convolution module to capture features at various scales. The study focuses on rural roads in the Xiong'an New Area, China, using high-resolution imagery from China's Gaofen-2 satellite to construct the XARoads rural road dataset. Roads were extracted from the XARoads dataset and DeepGlobe public dataset using the RC-MSFNet model and compared with some models such as U-Net, FCN, SegNet, DeeplabV3+, R-Net, and RC-Net. Experimental results showed that: (1) The proposed method achieved precision (P), intersection over union (IOU), and completeness (COM) scores of 0.8350, 0.6523, and 0.7489, respectively, for rural road extraction in Xiong'an New Area, representing precision improvements of 3.8%, 6.78%, 7.85%, 2.14%, 0.58%, and 2.53% over U-Net, FCN, SegNet, DeeplabV3+, R-Net, and RC-Net. (2) The method excelled at extracting narrow roads and muddy roads with unclear boundaries, with fewer instances of omission or false extraction, demonstrating advantages in complex rural terrain and areas with indistinct road boundaries. Accurate rural road extraction can provide valuable reference data for urban development and planning in the Xiong'an New Area.

IMU-Based Fitness Activity Recognition Using CNNs for Time Series Classification.

Mobile fitness applications provide the opportunity to show users real-time feedback on their current fitness activity. For such applications, it is essential to accurately track the user's current fitness activity using available mobile sensors, such as inertial measurement units (IMUs). Convolutional neural networks (CNNs) have been shown to produce strong results in different time series classification tasks, including the recognition of daily living activities. However, fitness activities can present unique challenges to the human activity recognition task (HAR), including greater similarity between individual activities and fewer available data for model training. In this paper, we evaluate the applicability of CNNs to the fitness activity recognition task (FAR) using IMU data and determine the impact of input data size and sensor count on performance. For this purpose, we adapted three existing CNN architectures to the FAR task and designed a fourth CNN variant, which we call the scaling fully convolutional network (Scaling-FCN). We designed a preprocessing pipeline and recorded a running exercise data set with 20 participants, in which we evaluated the respective recognition performances of the four networks, comparing them with three traditional machine learning (ML) methods commonly used in HAR. Although CNN architectures achieve at least 94% test accuracy in all scenarios, two traditional ML architectures surpass them in the default scenario, with support vector machines (SVMs) achieving 99.00 ± 0.34% test accuracy. The removal of all sensors except one foot sensor reduced the performance of traditional ML architectures but improved the performance of CNN architectures on our data set, with our Scaling-FCN reaching the highest accuracy of 99.86 ± 0.11% on the test set. Our results suggest that CNNs are generally well suited for fitness activity recognition, and noticeable performance improvements can be achieved if sensors are dropped selectively, although traditional ML architectures can still compete with or even surpass CNNs when favorable input data are utilized.

Enhancing soil texture classification with multivariate scattering correction and residual neural networks using visible near-infrared spectra.

Soil texture is one of the most important indicators of soil physical properties, which has traditionally been measured through laborious procedures. Approaches utilizing visible near-infrared spectroscopy, with their advantages in efficiency, eco-friendliness and non-destruction, are emerging as potent alternatives. Nevertheless, these approaches often suffer from limitations in classification accuracy, and the substantial impact of spectral preprocessing, model integration, and sample matrix effect is commonly disregarded. Here a novel 11-class soil texture classification strategy that address this challenge by combining Multiplicative Scatter Correction (MSC) with Residual Network (ResNet) models was presented, resulting in exceptional classification accuracy. Utilizing the LUCAS dataset, collected by the Land Use and Cover Area frame Statistical Survey project, we thoroughly evaluated eight spectral preprocessing methods. Our findings underscored the superior performance of MSC in reducing spatial complexity within spectral data, showcasing its crucial role in enhancing model precision. Through comparisons of three 1D CNN models and two ResNet models integrated with MSC, we established the superior performance of the MSC-incorporated ResNet model, achieving an overall accuracy of 98.97 % and five soil textures even reached 100.00 %. The ResNet model demonstrated a marked superiority in classifying datasets with similar features, as observed by the confusion matrix analysis. Moreover, we investigated the potential benefit of pre-categorization based on land cover type of the soil samples in enhancing the accuracy of soil texture classification models, achieving overall classification accuracies exceeding 99.39 % for woodland, grassland, and farmland with the 2-layer ResNet model. The proposed work provides a pioneering and efficient strategy for rapid and precise soil texture identification via visible near-infrared spectroscopy, demonstrating unparalleled accuracy compared to existing methods, thus significantly enhancing the practical application prospects in soil, agricultural and environmental science.

[Bowel Sounds Detection Method Based on ResNet-BiLSTM and Attention Mechanism].

Bowel sounds can reflect the movement and health status of the gastrointestinal tract. However, the traditional manual auscultation method has subjective deviation and is time-consuming and labor-intensive. In order to better assist doctors in diagnosing bowel sounds and improve the reliability and efficiency of bowel sound detection, this study proposed a deep neural network model that combines a residual neural network (ResNet), a bidirectional long short-term memory network (BiLSTM), and an attention mechanism. Firstly, a large number of labeled clinical data was collected using the self-developed multi-channel bowel sound acquisition system, and the multi-scale wavelet decomposition and reconstruction method was used to preprocess the bowel sounds. Then, log Mel spectrogram features were extracted and sent to the network for training. Finally, the performance and effectiveness of the model were evaluated and verified by 10-fold cross-validation and an ablation experiment. The experimental results showed that the precision, recall, and F 1 score of the model reached 83%, 76%, and 79%, respectively, and it could effectively detect bowel sound segments and locate their start and end times, performing better than previous algorithms. This algorithm can not only provide auxiliary information for doctors in clinical practice but also offer technical support for further analysis and research of bowel sounds.

Deep-learning-based ring artifact correction for tomographic reconstruction.

X-ray tomography has been widely used in various research fields thanks to its capability of observing 3D structures with high resolution non-destructively. However, due to the nonlinearity and inconsistency of detector pixels, ring artifacts usually appear in tomographic reconstruction, which may compromise image quality and cause nonuniform bias. This study proposes a new ring artifact correction method based on the residual neural network (ResNet) for X-ray tomography. The artifact correction network uses complementary information of each wavelet coefficient and a residual mechanism of the residual block to obtain high-precision artifacts through low operation costs. In addition, a regularization term is used to accurately extract stripe artifacts in sinograms, so that the network can better preserve image details while accurately separating artifacts. When applied to simulation and experimental data, the proposed method shows a good suppression of ring artifacts. To solve the problem of insufficient training data, ResNet is trained through the transfer learning strategy, which brings advantages of robustness, versatility and low computing cost.

A comparative study of deep learning-based knowledge-based planning methods for 3D dose distribution prediction of head and neck.

PURPOSE: In this paper, we compare four novel knowledge-based planning (KBP) algorithms using deep learning to predict three-dimensional (3D) dose distributions of head and neck plans using the same patients' dataset and quantitative assessment metrics. METHODS: A dataset of 340 oropharyngeal cancer patients treated with intensity-modulated radiation therapy was used in this study, which represents the AAPM OpenKBP - 2020 Grand Challenge dataset. Four 3D convolutional neural network architectures were built. The models were trained on 64% of the data set and validated on 16% for voxel-wise dose predictions: U-Net, attention U-Net, residual U-Net (Res U-Net), and attention Res U-Net. The trained models were then evaluated for their performance on a test data set (20% of the data) by comparing the predicted dose distributions against the ground-truth using dose statistics and dose-volume indices. RESULTS: The four KBP dose prediction models exhibited promising performance with an averaged mean absolute dose error within the body contour <3 Gy on 68 plans in the test set. The average difference in predicting the D99 index for all targets was 0.92 Gy (p = 0.51) for attention Res U-Net, 0.94 Gy (p = 0.40) for Res U-Net, 2.94 Gy (p = 0.09) for attention U-Net, and 3.51 Gy (p = 0.08) for U-Net. For the OARs, the values for the D m a x ${D_{max}}$ and D m e a n ${D_{mean}}$ indices were 2.72 Gy (p < 0.01) for attention Res U-Net, 2.94 Gy (p < 0.01) for Res U-Net, 1.10 Gy (p < 0.01) for attention U-Net, 0.84 Gy (p < 0.29) for U-Net. CONCLUSION: All models demonstrated almost comparable performance for voxel-wise dose prediction. KBP models that employ 3D U-Net architecture as a base could be deployed for clinical use to improve cancer patient treatment by creating plans with consistent quality and making the radiotherapy workflow more efficient.

An Automated Skill Assessment Framework Based on Visual Motion Signals and a Deep Neural Network in Robot-Assisted Minimally Invasive Surgery.

Surgical skill assessment can quantify the quality of the surgical operation via the motion state of the surgical instrument tip (SIT), which is considered one of the effective primary means by which to improve the accuracy of surgical operation. Traditional methods have displayed promising results in skill assessment. However, this success is predicated on the SIT sensors, making these approaches impractical when employing the minimally invasive surgical robot with such a tiny end size. To address the assessment issue regarding the operation quality of robot-assisted minimally invasive surgery (RAMIS), this paper proposes a new automatic framework for assessing surgical skills based on visual motion tracking and deep learning. The new method innovatively combines vision and kinematics. The kernel correlation filter (KCF) is introduced in order to obtain the key motion signals of the SIT and classify them by using the residual neural network (ResNet), realizing automated skill assessment in RAMIS. To verify its effectiveness and accuracy, the proposed method is applied to the public minimally invasive surgical robot dataset, the JIGSAWS. The results show that the method based on visual motion tracking technology and a deep neural network model can effectively and accurately assess the skill of robot-assisted surgery in near real-time. In a fairly short computational processing time of 3 to 5 s, the average accuracy of the assessment method is 92.04% and 84.80% in distinguishing two and three skill levels. This study makes an important contribution to the safe and high-quality development of RAMIS.

Bearing Fault Reconstruction Diagnosis Method Based on ResNet-152 with Multi-Scale Stacked Receptive Field.

The axle box in the bogie system of subway trains is a key component connecting primary damper and the axle. In order to extract deep features and large-scale fault features for rapid diagnosis, a novel fault reconstruction characteristics classification method based on deep residual network with a multi-scale stacked receptive field for rolling bearings of a subway train axle box is proposed. Firstly, multi-layer stacked convolutional kernels and methods to insert them into ultra-deep residual networks are developed. Then, the original vibration signals of four fault characteristics acquired are reconstructed with a Gramian angular summation field and trainable large-scale 2D time-series images are obtained. In the end, the experimental results show that ResNet-152-MSRF has a low complexity of network structure, less trainable parameters than general convolutional neural networks, and no significant increase in network parameters and calculation time after embedding multi-layer stacked convolutional kernels. Moreover, there is a significant improvement in accuracy compared to lower depths, and a slight improvement in accuracy compared to networks than unembedded multi-layer stacked convolutional kernels.

A novel end-to-end method to predict RNA secondary structure profile based on bidirectional LSTM and residual neural network.

BACKGROUND: Studies have shown that RNA secondary structure, a planar structure formed by paired bases, plays diverse vital roles in fundamental life activities and complex diseases. RNA secondary structure profile can record whether each base is paired with others. Hence, accurate prediction of secondary structure profile can help to deduce the secondary structure and binding site of RNA. RNA secondary structure profile can be obtained through biological experiment and calculation methods. Of them, the biological experiment method involves two ways: chemical reagent and biological crystallization. The chemical reagent method can obtain a large number of prediction data, but its cost is high and always associated with high noise, making it difficult to get results of all bases on RNA due to the limited of sequencing coverage. By contrast, the biological crystallization method can lead to accurate results, yet heavy experimental work and high costs are required. On the other hand, the calculation method is CROSS, which comprises a three-layer fully connected neural network. However, CROSS can not completely learn the features of RNA secondary structure profile since its poor network structure, leading to its low performance. RESULTS: In this paper, a novel end-to-end method, named as "RPRes, was proposed to predict RNA secondary structure profile based on Bidirectional LSTM and Residual Neural Network. CONCLUSIONS: RPRes utilizes data sets generated by multiple biological experiment methods as the training, validation, and test sets to predict profile, which can compatible with numerous prediction requirements. Compared with the biological experiment method, RPRes has reduced the costs and improved the prediction efficiency. Compared with the state-of-the-art calculation method CROSS, RPRes has significantly improved performance.

Artificial intelligence classification model for macular degeneration images: a robust optimization framework for residual neural networks.

BACKGROUND: The prevalence of chronic disease is growing in aging societies, and artificial-intelligence-assisted interpretation of macular degeneration images is a topic that merits research. This study proposes a residual neural network (ResNet) model constructed using uniform design. The ResNet model is an artificial intelligence model that classifies macular degeneration images and can assist medical professionals in related tests and classification tasks, enhance confidence in making diagnoses, and reassure patients. However, the various hyperparameters in a ResNet lead to the problem of hyperparameter optimization in the model. This study employed uniform design-a systematic, scientific experimental design-to optimize the hyperparameters of the ResNet and establish a ResNet with optimal robustness. RESULTS: An open dataset of macular degeneration images ( https://data.mendeley.com/datasets/rscbjbr9sj/3 ) was divided into training, validation, and test datasets. According to accuracy, false negative rate, and signal-to-noise ratio, this study used uniform design to determine the optimal combination of ResNet hyperparameters. The ResNet model was tested and the results compared with results obtained in a previous study using the same dataset. The ResNet model achieved higher optimal accuracy (0.9907), higher mean accuracy (0.9848), and a lower mean false negative rate (0.015) than did the model previously reported. The optimal ResNet hyperparameter combination identified using the uniform design method exhibited excellent performance. CONCLUSION: The high stability of the ResNet model established using uniform design is attributable to the study's strict focus on achieving both high accuracy and low standard deviation. This study optimized the hyperparameters of the ResNet model by using uniform design because the design features uniform distribution of experimental points and facilitates effective determination of the representative parameter combination, reducing the time required for parameter design and fulfilling the requirements of a systematic parameter design process.

When homologous sequences meet structural decoys: Accurate contact prediction by tFold in CASP14-(tFold for CASP14 contact prediction).

In this paper, we report our tFold framework's performance on the inter-residue contact prediction task in the 14th Critical Assessment of protein Structure Prediction (CASP14). Our tFold framework seamlessly combines both homologous sequences and structural decoys under an ultra-deep network architecture. Squeeze-excitation and axial attention mechanisms are employed to effectively capture inter-residue interactions. In CASP14, our best predictor achieves 41.78% in the averaged top-L precision for long-range contacts for all the 22 free-modeling (FM) targets, and ranked 1st among all the 60 participating teams. The tFold web server is now freely available at: https://drug.ai.tencent.com/console/en/tfold.

Deep Residual Neural Networks Resolve Quartet Molecular Phylogenies.

Phylogenetic inference is of fundamental importance to evolutionary as well as other fields of biology, and molecular sequences have emerged as the primary data for this task. Although many phylogenetic methods have been developed to explicitly take into account substitution models of sequence evolution, such methods could fail due to model misspecification or insufficiency, especially in the face of heterogeneities in substitution processes across sites and among lineages. In this study, we propose to infer topologies of four-taxon trees using deep residual neural networks, a machine learning approach needing no explicit modeling of the subject system and having a record of success in solving complex nonlinear inference problems. We train residual networks on simulated protein sequence data with extensive amino acid substitution heterogeneities. We show that the well-trained residual network predictors can outperform existing state-of-the-art inference methods such as the maximum likelihood method on diverse simulated test data, especially under extensive substitution heterogeneities. Reassuringly, residual network predictors generally agree with existing methods in the trees inferred from real phylogenetic data with known or widely believed topologies. Furthermore, when combined with the quartet puzzling algorithm, residual network predictors can be used to reconstruct trees with more than four taxa. We conclude that deep learning represents a powerful new approach to phylogenetic reconstruction, especially when sequences evolve via heterogeneous substitution processes. We present our best trained predictor in a freely available program named Phylogenetics by Deep Learning (PhyDL, https://gitlab.com/ztzou/phydl; last accessed January 3, 2020).

Deep-learning-based image registration for nano-resolution tomographic reconstruction.

Nano-resolution full-field transmission X-ray microscopy has been successfully applied to a wide range of research fields thanks to its capability of non-destructively reconstructing the 3D structure with high resolution. Due to constraints in the practical implementations, the nano-tomography data is often associated with a random image jitter, resulting from imperfections in the hardware setup. Without a proper image registration process prior to the reconstruction, the quality of the result will be compromised. Here a deep-learning-based image jitter correction method is presented, which registers the projective images with high efficiency and accuracy, facilitating a high-quality tomographic reconstruction. This development is demonstrated and validated using synthetic and experimental datasets. The method is effective and readily applicable to a broad range of applications. Together with this paper, the source code is published and adoptions and improvements from our colleagues in this field are welcomed.

Use of residual neural network for the detection of ossification of the posterior longitudinal ligament on plain cervical radiography.

Ossification of the posterior longitudinal ligament (OPLL) causes serious problems, such as myelopathy and acute spinal cord injury. The early and accurate diagnosis of OPLL would hence prevent the miserable prognoses. Plain lateral radiography is an essential method for the evaluation of OPLL. Therefore, minimizing the diagnostic errors of OPLL on radiography is crucial. Image identification based on a residual neural network (RNN) has been recognized to be potentially effective as a diagnostic strategy for orthopedic diseases; however, the accuracy of detecting OPLL using RNN has remained unclear. An RNN was trained with plain lateral cervical radiography images of 2,318 images from 672 patients (535 images from 304 patients with OPLL and 1,773 images from 368 patients of Negative). The accuracy, sensitivity, specificity, false positive rate, and false negative rate of diagnosis of the RNN were calculated. The mean accuracy, sensitivity, specificity, false positive rate, and false negative rate of the model were 98.9%, 97.0%, 99.4%, 2.2%, and 1.0%, respectively. The model achieved an overall area under the curve of 0.99 (95% confidence interval, 0.97-1.00) in which AUC in each fold estimated was 0.99, 0.99, 0.98, 0.98, and 0.99, respectively. An algorithm trained by an RNN could make binary classification of OPLL on cervical lateral X-ray images. RNN may hence be useful as a screening tool to assist physicians in identifying patients with OPLL in future setting. To achieve accurate identification of OPLL patients clinically, RNN has to be trained with other cause of myelopathy.

Deep Cooperative Spectrum Sensing Based on Residual Neural Network Using Feature Extraction and Random Forest Classifier.

Some bands in the frequency spectrum have become overloaded and others underutilized due to the considerable increase in demand and user allocation policy. Cognitive radio applies detection techniques to dynamically allocate unlicensed users. Cooperative spectrum sensing is currently showing promising results. Therefore, in this work, we propose a cooperative spectrum detection system based on a residual neural network architecture combined with feature extractor and random forest classifier. The objective of this paper is to propose a cooperative spectrum sensing approach that can achieve high accuracy in higher levels of noise power density with less unlicensed users cooperating in the system. Therefore, we propose to extract features of the sensing information of each unlicensed user, then we use a random forest to classify if there is a presence of a licensed user in each band analyzed by the unlicensed user. Then, information from several unlicensed users are shared to a fusion center, where the decision about the presence or absence of a licensed user is accomplished by a model trained by a residual neural network. In our work, we achieved a high level of accuracy even when the noise power density is high, which means that our proposed approach is able to recognize the presence of a licensed user in 98% of the cases when the evaluated channel suffers a high level of noise power density (-134 dBm/Hz). This result was achieved with the cooperation of 10 unlicensed users.

RDb2C2: an improved method to identify the residue-residue pairing in ß strands.

BACKGROUND: Despite the great advance of protein structure prediction, accurate prediction of the structures of mainly ß proteins is still highly challenging, but could be assisted by the knowledge of residue-residue pairing in ß strands. Previously, we proposed a ridge-detection-based algorithm RDb2C that adopted a multi-stage random forest framework to predict the ß-ß pairing given the amino acid sequence of a protein. RESULTS: In this work, we developed a second version of this algorithm, RDb2C2, by employing the residual neural network to further enhance the prediction accuracy. In the benchmark test, this new algorithm improves the F1-score by > 10 percentage points, reaching impressively high values of ~ 72% and ~ 73% in the BetaSheet916 and BetaSheet1452 sets, respectively. CONCLUSION: Our new method promotes the prediction accuracy of ß-ß pairing to a new level and the prediction results could better assist the structure modeling of mainly ß proteins. We prepared an online server of RDb2C2 at http://structpred.life.tsinghua.edu.cn/rdb2c2.html.

Template-based prediction of protein structure with deep learning.

BACKGROUND: Accurate prediction of protein structure is fundamentally important to understand biological function of proteins. Template-based modeling, including protein threading and homology modeling, is a popular method for protein tertiary structure prediction. However, accurate template-query alignment and template selection are still very challenging, especially for the proteins with only distant homologs available. RESULTS: We propose a new template-based modelling method called ThreaderAI to improve protein tertiary structure prediction. ThreaderAI formulates the task of aligning query sequence with template as the classical pixel classification problem in computer vision and naturally applies deep residual neural network in prediction. ThreaderAI first employs deep learning to predict residue-residue aligning probability matrix by integrating sequence profile, predicted sequential structural features, and predicted residue-residue contacts, and then builds template-query alignment by applying a dynamic programming algorithm on the probability matrix. We evaluated our methods both in generating accurate template-query alignment and protein threading. Experimental results show that ThreaderAI outperforms currently popular template-based modelling methods HHpred, CNFpred, and the latest contact-assisted method CEthreader, especially on the proteins that do not have close homologs with known structures. In particular, in terms of alignment accuracy measured with TM-score, ThreaderAI outperforms HHpred, CNFpred, and CEthreader by 56, 13, and 11%, respectively, on template-query pairs at the similarity of fold level from SCOPe data. And on CASP13's TBM-hard data, ThreaderAI outperforms HHpred, CNFpred, and CEthreader by 16, 9 and 8% in terms of TM-score, respectively. CONCLUSIONS: These results demonstrate that with the help of deep learning, ThreaderAI can significantly improve the accuracy of template-based structure prediction, especially for distant-homology proteins.