Protein Sequence Features Research Articles

Prediction of the Receptorome for the Human-Infecting Virome.

The virus receptors are key for the viral infection of host cells. Identification of the virus receptors is still challenging at present. Our previous study has shown that human virus receptor proteins have some unique features including high N-glycosylation level, high number of interaction partners and high expression level. Here, a random-forest model was built to identify human virus receptorome from human cell membrane proteins with an accepted accuracy based on the combination of the unique features of human virus receptors and protein sequences. A total of 1424 human cell membrane proteins were predicted to constitute the receptorome of the human-infecting virome. In addition, the combination of the random-forest model with protein–protein interactions between human and viruses predicted in previous studies enabled further prediction of the receptors for 693 human-infecting viruses, such as the enterovirus, norovirus and West Nile virus. Finally, the candidate alternative receptors of the SARS-CoV-2 were also predicted in this study. As far as we know, this study is the first attempt to predict the receptorome for the human-infecting virome and would greatly facilitate the identification of the receptors for viruses.Electronic supplementary materialThe online version of this article (10.1007/s12250-020-00259-6) contains supplementary material, which is available to authorized users.

Use of Chou's 5-steps rule to predict the subcellular localization of gram-negative and gram-positive bacterial proteins by multi-label learning based on gene ontology annotation and profile alignment.

To date, many proteins generated by large-scale genome sequencing projects are still uncharacterized and subject to intensive investigations by both experimental and computational means. Knowledge of protein subcellular localization (SCL) is of key importance for protein function elucidation. However, it remains a challenging task, especially for multiple sites proteins known to shuttle between cell compartments to perform their proper biological functions and proteins which do not have significant homology to proteins of known subcellular locations. Due to their low-cost and reasonable accuracy, machine learning-based methods have gained much attention in this context with the availability of a plethora of biological databases and annotated proteins for analysis and benchmarking. Various predictive models have been proposed to tackle the SCL problem, using different protein sequence features pertaining to the subcellular localization, however, the overwhelming majority of them focuses on single localization and cover very limited cellular locations. The prediction was basically established on sorting signals, amino acids compositions, and homology. To improve the prediction quality, focus is actually on knowledge information extracted from annotation databases, such as protein–protein interactions and Gene Ontology (GO) functional domains annotation which has been recently a widely adopted and essential information for learning systems. To deal with such problem, in the present study, we considered SCL prediction task as a multi-label learning problem and tried to label both single site and multiple sites unannotated bacterial protein sequences by mining proteins homology relationships using both GO terms of protein homologs and PSI-BLAST profiles. The experiments using 5-fold cross-validation tests on the benchmark datasets showed a significant improvement on the results obtained by the proposed consensus multi-label prediction model which discriminates six compartments for Gram-negative and five compartments for Gram-positive bacterial proteins.

An Improved Topology Prediction of Alpha-Helical Transmembrane Protein Based on Deep Multi-Scale Convolutional Neural Network.

Alpha-helical proteins ( αTMPs) are essential in various biological processes. Despite their tertiary structures are crucial for revealing complex functions, experimental structure determination remains challenging and costly. In the past decades, various sequence-based topology prediction methods have been developed to bridge the gap between the sequences and structures by characterizing the structural features, but significant improvements are still required. Deep learning brings a great opportunity for its powerful representation learning capability from limited original data. In this work, we improved our αTMP topology prediction method DMCTOP using deep learning, which composed of two deep convolutional blocks to simultaneously extract local and global contextual features. Consequently, the inputs were simplified to reflect the original features of the sequence, including a protein sequence feature and an evolutionary conservation feature. DMCTOP can efficiently and accurately identify all topological types and the N-terminal orientation for an αTMP sequence. To validate the effectiveness of our method, we benchmarked DMCTOP against 13 peer methods according to the whole sequence, the transmembrane segment and the traditional criterion in testing experiments. All the results reveal that our method achieved the highest prediction accuracy and outperformed all the previous methods. The method is available at https://icdtools.nenu.edu.cn/dmctop.

Sequence analysis and protein interactions of Arabidopsis CIA2 and CIL proteins

BackgroundA previous screening of Arabidopsis thaliana for mutants exhibiting dysfunctional chloroplast protein transport identified the chloroplast import apparatus (cia) gene. The cia2 mutant has a pale green phenotype and reduced rate of protein import into chloroplasts, but leaf shape and size are similar to wild-type plants of the same developmental stage. Microarray analysis showed that nuclear CIA2 protein enhances expression of the Toc75, Toc33, CPN10 and cpRPs genes, thereby up-regulating protein import and synthesis efficiency in chloroplasts. CIA2-like (CIL) shares 65% sequence identity to CIA2, suggesting that CIL and CIA2 are homologous proteins in Arabidopsis. Here, we further assess the protein interactions and sequence features of CIA2 and CIL.ResultsSubcellular localizations of truncated CIA2 protein fragments in our onion transient assay demonstrate that CIA2 contains two nuclear localization signals (NLS) located at amino acids (aa) 62-65 and 291-308, whereas CIL has only one NLS at aa 47-50. We screened a yeast two-hybrid (Y2H) Arabidopsis cDNA library to search for putative CIA2-interacting proteins and identified 12 nuclear proteins, including itself, CIL, and flowering-control proteins (such as CO, NF-YB1, NF-YC1, NF-YC9 and ABI3). Additional Y2H experiments demonstrate that CIA2 and CIL mainly interact with flowering-control proteins via their N-termini, but preferentially form homo- or hetero-dimers through their C-termini. Moreover, sequence alignment showed that the N-terminal sequences of CIA2, CIL and NF-YA are highly conserved. Therefore, NF-YA in the NF-Y complex could be substituted by CIA2 or CIL.ConclusionsWe show that Arabidopsis CIA2 and CIL can interact with CO and NF-Y complex, so not only may they contribute to regulate chloroplast function but also to modulate flower development.

Subcellular location prediction of apoptosis proteins using two novel feature extraction methods based on evolutionary information and LDA

BackgroundApoptosis, also called programmed cell death, refers to the spontaneous and orderly death of cells controlled by genes in order to maintain a stable internal environment. Identifying the subcellular location of apoptosis proteins is very helpful in understanding the mechanism of apoptosis and designing drugs. Therefore, the subcellular localization of apoptosis proteins has attracted increased attention in computational biology. Effective feature extraction methods play a critical role in predicting the subcellular location of proteins.ResultsIn this paper, we proposed two novel feature extraction methods based on evolutionary information. One of the features obtained the evolutionary information via the transition matrix of the consensus sequence (CTM). And the other utilized the evolutionary information from PSSM based on absolute entropy correlation analysis (AECA-PSSM). After fusing the two kinds of features, linear discriminant analysis (LDA) was used to reduce the dimension of the proposed features. Finally, the support vector machine (SVM) was adopted to predict the protein subcellular locations. The proposed CTM-AECA-PSSM-LDA subcellular location prediction method was evaluated using the CL317 dataset and ZW225 dataset. By jackknife test, the overall accuracy was 99.7% (CL317) and 95.6% (ZW225) respectively.ConclusionsThe experimental results show that the proposed method which is hopefully to be a complementary tool for the existing methods of subcellular localization, can effectively extract more abundant features of protein sequence and is feasible in predicting the subcellular location of apoptosis proteins.

The crystal structure of human forkhead box N1 in complex with DNA reveals the structural basis for forkhead box family specificity

Forkhead box N1 (FOXN1) is a member of the forkhead box family of transcription factors and plays an important role in thymic epithelial cell differentiation and development. FOXN1 mutations in humans and mice give rise to the “nude” phenotype, which is marked by athymia. FOXN1 belongs to a subset of the FOX family that recognizes an alternative forkhead-like (FHL) consensus sequence (GACGC) that is different from the more widely recognized forkhead (FKH) sequence RYAAAYA (where R is purine, and Y is pyrimidine). Here, we present the FOXN1 structure in complex with DNA containing an FHL motif at 1.6 Å resolution, in which the DNA sequence is recognized by a mixture of direct and water-mediated contacts provided by residues in an α-helix inserted in the DNA major groove (the recognition helix). Comparisons with the structure of other FOX family members revealed that the FKH and FHL DNA sequences are bound in two distinct modes, with partially different registers for the protein DNA contacts. We identified a single alternative rotamer within the recognition helix itself as an important determinant of DNA specificity and found protein sequence features in the recognition helix that could be used to predict the specificity of other FOX family members. Finally, we demonstrate that the C-terminal region of FOXN1 is required for high-affinity DNA binding and that FOXN1 has a significantly reduced affinity for DNA that contains 5′-methylcytosine, which may have implications for the role of FOXN1 in thymic involution.

RPI-SE: a stacking ensemble learning framework for ncRNA-protein interactions prediction using sequence information

BackgroundThe interactions between non-coding RNAs (ncRNA) and proteins play an essential role in many biological processes. Several high-throughput experimental methods have been applied to detect ncRNA-protein interactions. However, these methods are time-consuming and expensive. Accurate and efficient computational methods can assist and accelerate the study of ncRNA-protein interactions.ResultsIn this work, we develop a stacking ensemble computational framework, RPI-SE, for effectively predicting ncRNA-protein interactions. More specifically, to fully exploit protein and RNA sequence feature, Position Weight Matrix combined with Legendre Moments is applied to obtain protein evolutionary information. Meanwhile, k-mer sparse matrix is employed to extract efficient feature of ncRNA sequences. Finally, an ensemble learning framework integrated different types of base classifier is developed to predict ncRNA-protein interactions using these discriminative features. The accuracy and robustness of RPI-SE was evaluated on three benchmark data sets under five-fold cross-validation and compared with other state-of-the-art methods.ConclusionsThe results demonstrate that RPI-SE is competent for ncRNA-protein interactions prediction task with high accuracy and robustness. It’s anticipated that this work can provide a computational prediction tool to advance ncRNA-protein interactions related biomedical research.

Exploring the H2H genes in 3D view

It is estimated to contain hundreds of head-to-head (H2H) gene in the eukaryotic genomes, often with two genes in an H2H pair been prone to coexpress and o-function. Therefore, h2h plays a crucial role in human disease control and there have been many studies on H2H gene and its related bidirectional promoters. Recent chromosome conformation capture techniques, such as Hi-C, and ChIA-PET have provided us with new opportunities to study H2H in 3D view. This paper proposes a powerful machine learning algorithm LightGBM to predict h2h cluster. Two sets of features, namely protein features and sequence features, are extracted. Then these two sets of features are used to train a classifier to predict h2h cluster. Experimental results show that this method can effectively predict h2h cluster. Our results show a large fraction of long-range H2H (TSS >1 kb) exist and help us to understand the H2H at 3D level.

A Deep Learning Framework for Gene Ontology Annotations With Sequence- and Network-Based Information.

Knowledge of protein functions plays an important role in biology and medicine. With the rapid development of high-throughput technologies, a huge number of proteins have been discovered. However, there are a great number of proteins without functional annotations. A protein usually has multiple functions and some functions or biological processes require interactions of a plurality of proteins. Additionally, Gene Ontology provides a useful classification for protein functions and contains more than 40,000 terms. We propose a deep learning framework called DeepGOA to predict protein functions with protein sequences and protein-protein interaction (PPI) networks. For protein sequences, we extract two types of information: sequence semantic information and subsequence-based features. We use the word2vec technique to numerically represent protein sequences, and utilize a Bi-directional Long and Short Time Memory (Bi-LSTM) and multi-scale convolutional neural network (multi-scale CNN) to obtain the global and local semantic features of protein sequences, respectively. Additionally, we use the InterPro tool to scan protein sequences for extracting subsequence-based information, such as domains and motifs. Then, the information is plugged into a neural network to generate high-quality features. For the PPI network, the Deepwalk algorithm is applied to generate its embedding information of PPI. Then the two types of features are concatenated together to predict protein functions. To evaluate the performance of DeepGOA, several different evaluation methods and metrics are utilized. The experimental results show that DeepGOA outperforms DeepGO and BLAST.

Using Random Forest Model Combined With Gabor Feature to Predict Protein-Protein Interaction From Protein Sequence.

Protein-protein interactions (PPIs) play a crucial role in the life cycles of living cells. Thus, it is important to understand the underlying mechanisms of PPIs. Although many high-throughput technologies have generated large amounts of PPI data in different organisms, the experiments for detecting PPIs are still costly and time-consuming. Therefore, novel computational methods are urgently needed for predicting PPIs. For this reason, developing a new computational method for predicting PPIs is drawing more and more attention. In this study, we proposed a novel computational method based on texture feature of protein sequence for predicting PPIs. Especially, the Gabor feature is used to extract texture feature and protein evolutionary information from Position-Specific Scoring Matrix, which is generated by Position-Specific Iterated Basic Local Alignment Search Tool. Then, random forest–based classifiers are used to infer the protein interactions. When performed on PPI data sets of yeast, human, and Helicobacter pylori, we obtained good results with average accuracies of 92.10%, 97.03%, and 86.45%, respectively. To better evaluate the proposed method, we compared Gabor feature, Discrete Cosine Transform, and Local Phase Quantization. Our results show that the proposed method is both feasible and stable and the Gabor feature descriptor is reliable in extracting protein sequence information. Furthermore, additional experiments have been conducted to predict PPIs of other 4 species data sets. The promising results indicate that our proposed method is both powerful and robust.

Deep learning based prediction of reversible HAT/HDAC-specific lysine acetylation.

Protein lysine acetylation regulation is an important molecular mechanism for regulating cellular processes and plays critical physiological and pathological roles in cancers and diseases. Although massive acetylation sites have been identified through experimental identification and high-throughput proteomics techniques, their enzyme-specific regulation remains largely unknown. Here, we developed the deep learning-based protein lysine acetylation modification prediction (Deep-PLA) software for histone acetyltransferase (HAT)/histone deacetylase (HDAC)-specific acetylation prediction based on deep learning. Experimentally identified substrates and sites of several HATs and HDACs were curated from the literature to generate enzyme-specific data sets. We integrated various protein sequence features with deep neural network and optimized the hyperparameters with particle swarm optimization, which achieved satisfactory performance. Through comparisons based on cross-validations and testing data sets, the model outperformed previous studies. Meanwhile, we found that protein-protein interactions could enrich enzyme-specific acetylation regulatory relations and visualized this information in the Deep-PLA web server. Furthermore, a cross-cancer analysis of acetylation-associated mutations revealed that acetylation regulation was intensively disrupted by mutations in cancers and heavily implicated in the regulation of cancer signaling. These prediction and analysis results might provide helpful information to reveal the regulatory mechanism of protein acetylation in various biological processes to promote the research on prognosis and treatment of cancers. Therefore, the Deep-PLA predictor and protein acetylation interaction networks could provide helpful information for studying the regulation of protein acetylation. The web server of Deep-PLA could be accessed at http://deeppla.cancerbio.info.

Convolution Neural Network-Based Prediction of Protein Thermostability.

Most natural proteins exhibit poor thermostability, which limits their industrial application. Computer-aided rational design is an efficient purpose-oriented method that can improve protein thermostability. Numerous machine-learning-based methods have been designed to predict the changes in protein thermostability induced by mutations. However, all of these methods have certain limitations due to existing mutation coding methods that overlook protein sequence features. Here we propose a method to predict protein thermostability using convolutional neural networks based on an in-depth study of thermostability-related protein properties. This method comprises a three-dimensional coding algorithm, including protein mutation information and a strategy to extract neighboring features at protein mutation sites based on multiscale convolution. The accuracies on the S1615 and S388 data sets, which are widely used for protein thermostability predictions, reached 86.4 and 87%, respectively. The Matthews correlation coefficient was nearly double those produced using other methods. Furthermore, a model was constructed to predict the thermostability of Rhizomucor miehei lipase mutants based on the S3661 data set, a single amino acid mutation data set screened from the ProTherm protein thermodynamics database. Compared with the RIF strategy, which consists of three algorithms, i.e., Rosetta ddg monomer, I Mutant 3.0, and FoldX, the accuracy of the proposed method was higher (75.0 vs 66.7%), and the negative sample resolution was simultaneously enhanced. These results indicate that our prediction method more effectively assessed the protein thermostability and distinguished its features, making it a powerful tool to devise mutations that enhance the thermostability of proteins, particularly enzymes.

SubMito-XGBoost: predicting protein submitochondrial localization by fusing multiple feature information and eXtreme gradient boosting.

Mitochondria are an essential organelle in most eukaryotes. They not only play an important role in energy metabolism but also take part in many critical cytopathological processes. Abnormal mitochondria can trigger a series of human diseases, such as Parkinson's disease, multifactor disorder and Type-II diabetes. Protein submitochondrial localization enables the understanding of protein function in studying disease pathogenesis and drug design. We proposed a new method, SubMito-XGBoost, for protein submitochondrial localization prediction. Three steps are included: (i) the g-gap dipeptide composition (g-gap DC), pseudo-amino acid composition (PseAAC), auto-correlation function (ACF) and Bi-gram position-specific scoring matrix (Bi-gram PSSM) are employed to extract protein sequence features, (ii) Synthetic Minority Oversampling Technique (SMOTE) is used to balance samples, and the ReliefF algorithm is applied for feature selection and (iii) the obtained feature vectors are fed into XGBoost to predict protein submitochondrial locations. SubMito-XGBoost has obtained satisfactory prediction results by the leave-one-out-cross-validation (LOOCV) compared with existing methods. The prediction accuracies of the SubMito-XGBoost method on the two training datasets M317 and M983 were 97.7% and 98.9%, which are 2.8-12.5% and 3.8-9.9% higher than other methods, respectively. The prediction accuracy of the independent test set M495 was 94.8%, which is significantly better than the existing studies. The proposed method also achieves satisfactory predictive performance on plant and non-plant protein submitochondrial datasets. SubMito-XGBoost also plays an important role in new drug design for the treatment of related diseases. The source codes and data are publicly available at https://github.com/QUST-AIBBDRC/SubMito-XGBoost/. Supplementary data are available at Bioinformatics online.

PhaSepDB: a database of liquid-liquid phase separation related proteins.

It's widely appreciated that liquid–liquid phase separation (LLPS) underlies the formation of membraneless organelles, which function to concentrate proteins and nucleic acids. In the past few decades, major efforts have been devoted to identify the phase separation associated proteins and elucidate their functions. To better utilize the knowledge dispersed in published literature, we developed PhaSepDB (http://db.phasep.pro/), a manually curated database of phase separation associated proteins. Currently, PhaSepDB includes 2914 non-redundant proteins localized in different organelles curated from published literature and database. PhaSepDB provides protein summary, publication reference and sequence features of phase separation associated proteins. The sequence features which reflect the LLPS behavior are also available for other human protein candidates. The online database provides a convenient interface for the research community to easily browse, search and download phase separation associated proteins. As a centralized resource, we believe PhaSepDB will facilitate the future study of phase separation.

Increased expression of Phytophthora sojae genes encoding membrane-degrading enzymes appears to suggest an early onset of necrotrophy during Glycine max infection

Phytophthora sojae is an oomycete pathogen that causes root, stem, and leaf rot in soybean plants, frequently leading to massive economic losses. Despite its importance, the mechanism by which P. sojae penetrates the host is not yet fully understood. Evidence indicates that P. sojae is not capable of penetrating the plant cell wall via mechanical force, suggesting that alternative factors facilitate breakdown of the host cell wall and membrane. Members of the carbohydrate esterase (CE) family 10 (carboxylesterases, arylesterases, sterol esterases and acetylcholine esterases, collectively known as CE10), are thought to be important for this penetration process. To gain insight into the potential role of CE10-coding genes in P. sojae pathogenesis, the newly revised version of the P. sojae genome was searched for putative CE10-coding genes, and various bioinformatic analyses were conducted using their amino acid and nucleotide sequences. In addition, in planta infection assays were conducted with P. sojae Race 4 and soybean cultivars Williams and Williams 82, and the transcriptional activity of P. sojae CE10-coding genes was evaluated at different time points during infection. Results suggest that these genes are important for both the biotrophic and necrotrophic stages of the P. sojae infection process and provide molecular evidence for stage distinction during infection progression. Furthermore, bioinformatic analyses have identified several conserved gene and protein sequence features that appear to have a significant impact on observed levels of expression during infection. Results agree with previous reports implicating other carbohydrate-active enzymes in P. sojae infection.

Protein–protein interaction site prediction through combining local and global features with deep neural networks

Protein-protein interactions (PPIs) play important roles in many biological processes. Conventional biological experiments for identifying PPI sites are costly and time-consuming. Thus, many computational approaches have been proposed to predict PPI sites. Existing computational methods usually use local contextual features to predict PPI sites. Actually, global features of protein sequences are critical for PPI site prediction. A new end-to-end deep learning framework, named DeepPPISP, through combining local contextual and global sequence features, is proposed for PPI site prediction. For local contextual features, we use a sliding window to capture features of neighbors of a target amino acid as in previous studies. For global sequence features, a text convolutional neural network is applied to extract features from the whole protein sequence. Then the local contextual and global sequence features are combined to predict PPI sites. By integrating local contextual and global sequence features, DeepPPISP achieves the state-of-the-art performance, which is better than the other competing methods. In order to investigate if global sequence features are helpful in our deep learning model, we remove or change some components in DeepPPISP. Detailed analyses show that global sequence features play important roles in DeepPPISP. The DeepPPISP web server is available at http://bioinformatics.csu.edu.cn/PPISP/. The source code can be obtained from https://github.com/CSUBioGroup/DeepPPISP. Supplementary data are available at Bioinformatics online.

In-silico Structural Annotation of Methylthioadenosine Nucleosidase Protein Zm00014a_031618 in Maize (Zea mays L.)

Aims: The aim of this study was In-Silico structural annotation of an amino acid sequence of Methylthioadenosine Nucleosidase Protein Zm00014a_031618 in Maize (Zea mays) retrieved from NCBI with the accession number PWZ58979.
 Study Design: The use of In-Silico studies for the structural annotation of Methylthioadenosine Nucleosidase protein.
 Place and Duration of Study: The research was conducted at the Bioinformatics Laboratory, Chevron Biotechnology Centre, Modibbo Adama University of Yola, Nigeria. Between June 2018 to July 2018.
 Methodology: The Methylthioadenosine Nucleosidase protein was retrieved from NCBI, physical and chemical parameters were calculated using ExPASy - ProtParam tool, the server SOPMA was used for secondary structure analysis (helix, sheets, and coils) and I-TASSER was used to obtain the 3D structure.
 Results: ExPASy - Prot Param tool computated the various physical and chemical parameters such as molecular weight (MW) 30117.97, total number of positively (+R) 27, negatively charged residues (-R) 30, theoretical isoelectric point (pI) 5.96, aliphatic index (AI) 103.67 and grand average hydropathy (GRAVY) 0.293. The SOPMA server was used for calculating the secondary structural features of protein sequences as Alpha helix 39.16%, Extended strand 14.69%, Beta turn 6.64% and Random coil 39.51%. I-Tasser was used for predicting the 3D structure where 2qttA from PDB was used as the template.
 Conclusion: This study helped in understanding the structural analysis of the Methylthioadenosine Nucleosidase Protein Zm00014a_031618 in maize (Z. mays).

DeepACLSTM: deep asymmetric convolutional long short-term memory neural models for protein secondary structure prediction

BackgroundProtein secondary structure (PSS) is critical to further predict the tertiary structure, understand protein function and design drugs. However, experimental techniques of PSS are time consuming and expensive, and thus it’s very urgent to develop efficient computational approaches for predicting PSS based on sequence information alone. Moreover, the feature matrix of a protein contains two dimensions: the amino-acid residue dimension and the feature vector dimension. Existing deep learning based methods have achieved remarkable performances of PSS prediction, but the methods often utilize the features from the amino-acid dimension. Thus, there is still room to improve computational methods of PSS prediction.ResultsWe propose a novel deep neural network method, called DeepACLSTM, to predict 8-category PSS from protein sequence features and profile features. Our method efficiently applies asymmetric convolutional neural networks (ACNNs) combined with bidirectional long short-term memory (BLSTM) neural networks to predict PSS, leveraging the feature vector dimension of the protein feature matrix. In DeepACLSTM, the ACNNs extract the complex local contexts of amino-acids; the BLSTM neural networks capture the long-distance interdependencies between amino-acids. Furthermore, the prediction module predicts the category of each amino-acid residue based on both local contexts and long-distance interdependencies. To evaluate performances of DeepACLSTM, we conduct experiments on three publicly available datasets: CB513, CASP10 and CASP12. Results indicate that the performance of our method is superior to the state-of-the-art baselines on three publicly datasets.ConclusionsExperiments demonstrate that DeepACLSTM is an efficient predication method for predicting 8-category PSS and has the ability to extract more complex sequence-structure relationships between amino-acid residues. Moreover, experiments also indicate the feature vector dimension contains the useful information for improving PSS prediction.

DeepConv-DTI: Prediction of drug-target interactions via deep learning with convolution on protein sequences

Identification of drug-target interactions (DTIs) plays a key role in drug discovery. The high cost and labor-intensive nature of in vitro and in vivo experiments have highlighted the importance of in silico-based DTI prediction approaches. In several computational models, conventional protein descriptors have been shown to not be sufficiently informative to predict accurate DTIs. Thus, in this study, we propose a deep learning based DTI prediction model capturing local residue patterns of proteins participating in DTIs. When we employ a convolutional neural network (CNN) on raw protein sequences, we perform convolution on various lengths of amino acids subsequences to capture local residue patterns of generalized protein classes. We train our model with large-scale DTI information and demonstrate the performance of the proposed model using an independent dataset that is not seen during the training phase. As a result, our model performs better than previous protein descriptor-based models. Also, our model performs better than the recently developed deep learning models for massive prediction of DTIs. By examining pooled convolution results, we confirmed that our model can detect binding sites of proteins for DTIs. In conclusion, our prediction model for detecting local residue patterns of target proteins successfully enriches the protein features of a raw protein sequence, yielding better prediction results than previous approaches. Our code is available at https://github.com/GIST-CSBL/DeepConv-DTI.

PRECOG: PREdicting COupling probabilities of G-protein coupled receptors.

G-protein coupled receptors (GPCRs) control multiple physiological states by transducing a multitude of extracellular stimuli into the cell via coupling to intra-cellular heterotrimeric G-proteins. Deciphering which G-proteins couple to each of the hundreds of GPCRs present in a typical eukaryotic organism is therefore critical to understand signalling. Here, we present PRECOG (precog.russelllab.org): a web-server for predicting GPCR coupling, which allows users to: (i) predict coupling probabilities for GPCRs to individual G-proteins instead of subfamilies; (ii) visually inspect the protein sequence and structural features that are responsible for a particular coupling; (iii) suggest mutations to rationally design artificial GPCRs with new coupling properties based on predetermined coupling features.