De novo-designed transmembrane proteins bind and regulate a cytokine receptor.

Transmembrane (TM) domains as simple as a single span can perform complex biological functions using entirely lipid-embedded chemical features. Computational design has the potential to generate custom tool molecules directly targeting membrane proteins at their functional TM regions. Thus far, designed TM domain-targeting agents have been limited to mimicking the binding modes and motifs of natural TM interaction partners. Here, we demonstrate the design of de novo TM proteins targeting the erythropoietin receptor (EpoR) TM domain in a custom binding topology competitive with receptor homodimerization. The TM proteins expressed in mammalian cells complex with EpoR and inhibit erythropoietin-induced cell proliferation. In vitro, the synthetic TM domain complex outcompetes EpoR homodimerization. Structural characterization reveals that the complex involves the intended amino acids and agrees with our designed molecular model of antiparallel TM helices at 1:1 stoichiometry. Thus, membrane protein TM regions can now be targeted in custom-designed topologies.

Riboexp: an interpretable reinforcement learning framework for ribosome density modeling.

Translation elongation is a crucial phase during protein biosynthesis. In this study, we develop a novel deep reinforcement learning-based framework, named Riboexp, to model the determinants of the uneven distribution of ribosomes on mRNA transcripts during translation elongation. In particular, our model employs a policy network to perform a context-dependent feature selection in the setting of ribosome density prediction. Our extensive tests demonstrated that Riboexp can significantly outperform the state-of-the-art methods in predicting ribosome density by up to 5.9% in terms of per-gene Pearson correlation coefficient on the datasets from three species. In addition, Riboexp can indicate more informative sequence features for the prediction task than other commonly used attribution methods in deep learning. In-depth analyses also revealed the meaningful biological insights generated by the Riboexp framework. Moreover, the application of Riboexp in codon optimization resulted in an increase of protein production by around 31% over the previous state-of-the-art method that models ribosome density. These results have established Riboexp as a powerful and useful computational tool in the studies of translation dynamics and protein synthesis. Availability: The data and code of this study are available on GitHub: https://github.com/Liuxg16/Riboexp. Contact:zengjy321@tsinghua.edu.cn; songsen@tsinghua.edu.cn.

SNAC-tag for sequence-specific chemical protein cleavage.

Site-specific protein cleavage is essential for many protein-production protocols and typically requires proteases. We report the development of a chemical protein-cleavage method that is achieved through the use of a sequence-specific nickel-assisted cleavage (SNAC)-tag. We demonstrate that the SNAC-tag can be inserted before both water-soluble and membrane proteins to achieve fusion protein cleavage under biocompatible conditions with efficiency comparable to that of enzymes, and that the method works even when enzymatic cleavages fail.

Screening of Differentially Expressed Genes and Localization Analysis of Female Gametophyte at the Free Nuclear Mitosis Stage in Pinus tabuliformis Carr.

Female sterility is a common phenomenon in the plant world, and systematic research has not been carried out in gymnosperms. In this study, the ovules of No. 28 sterile line and No. 15 fertile line Pinus tabuliformis were used as materials, and a total of 18 cDNA libraries were sequenced by the HiSeqTM 4000 platform to analyze the differentially expressed genes (DEGs) and simple sequence repeats (SSRs) between the two lines. In addition, this study further analyzed the DEGs involved in the signal transduction of plant hormones, revealing that the signal pathways related to auxin, cytokinin, and gibberellin were blocked in the sterile ovule. Additionally, real-time fluorescent quantitative PCR verified that the expression trend of DEGs related to plant hormones was consistent with the results of high-throughput sequencing. Frozen sections and fluorescence in situ hybridization (FISH) were used to study the temporal and spatial expression patterns of PtRab in the ovules of P. tabuliformis. It was found that PtRab was significantly expressed in female gametophytes and rarely expressed in the surrounding diploid tissues. This study further explained the molecular regulation mechanism of female sterility in P. tabuliformis, preliminarily mining the key factors of ovule abortion in gymnosperms at the transcriptional level.

Secure multiparty computation for privacy-preserving drug discovery.

MOTIVATION: Quantitative structure-activity relationship (QSAR) and drug-target interaction (DTI) prediction are both commonly used in drug discovery. Collaboration among pharmaceutical institutions can lead to better performance in both QSAR and DTI prediction. However, the drug-related data privacy and intellectual property issues have become a noticeable hindrance for inter-institutional collaboration in drug discovery. RESULTS: We have developed two novel algorithms under secure multiparty computation (MPC), including QSARMPC and DTIMPC, which enable pharmaceutical institutions to achieve high-quality collaboration to advance drug discovery without divulging private drug-related information. QSARMPC, a neural network model under MPC, displays good scalability and performance and is feasible for privacy-preserving collaboration on large-scale QSAR prediction. DTIMPC integrates drug-related heterogeneous network data and accurately predicts novel DTIs, while keeping the drug information confidential. Under several experimental settings that reflect the situations in real drug discovery scenarios, we have demonstrated that DTIMPC possesses significant performance improvement over the baseline methods, generates novel DTI predictions with supporting evidence from the literature and shows the feasible scalability to handle growing DTI data. All these results indicate that QSARMPC and DTIMPC can provide practically useful tools for advancing privacy-preserving drug discovery. AVAILABILITY AND IMPLEMENTATION: The source codes of QSARMPC and DTIMPC are available on the GitHub: https://github.com/rongma6/QSARMPC_DTIMPC.git. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Evaluation of work resumption strategies after COVID-19 reopening in the Chinese city of Shenzhen: a mathematical modeling study.

OBJECTIVES: As China is facing a potential second wave of the epidemic, we reviewed and evaluated the intervention measures implemented in a major metropolitan city, Shenzhen, during the early phase of Wuhan lockdown. STUDY DESIGN: Based on the classic SEITR model and combined with population mobility, a compartmental model was constructed to simulate the transmission of COVID-19 and disease progression in the Shenzhen population. METHODS: Based on published epidemiological data on COVID-19 and population mobility data from Baidu Qianxi, we constructed a compartmental model to evaluate the impact of work and traffic resumption on the epidemic in Shenzhen in various scenarios. RESULTS: Imported cases account for most (58.6%) of the early reported cases in Shenzhen. We demonstrated that with strict inflow population control and a high level of mask usage after work resumption, various resumptions resulted in only an insignificant difference in the number of cumulative infections. Shenzhen may experience this second wave of infections approximately two weeks after the traffic resumption if the incidence risk in Hubei is high at the moment of resumption. CONCLUSION: Regardless of the work resumption strategy adopted in Shenzhen, the risk of a resurgence of COVID-19 after its reopening was limited. The strict control of imported cases and extensive use of facial masks play a key role in COVID-19 prevention.

Combined Transcriptome Analysis Reveals the Ovule Abortion Regulatory Mechanisms in the Female Sterile Line of Pinus tabuliformis Carr.

Ovule abortion is a common phenomenon in plants that has an impact on seed production. Previous studies of ovule and female gametophyte (FG) development have mainly focused on angiosperms, especially in Arabidopsis thaliana. However, because it is difficult to acquire information about ovule development in gymnosperms, this remains unclear. Here, we investigated the transcriptomic data of natural ovule abortion mutants (female sterile line, STE) and the wild type (female fertile line, FER) of Pinus tabuliformis Carr. to evaluate the mechanism of ovule abortion during the process of free nuclear mitosis (FNM). Using single-molecule real-time (SMRT) sequencing and next-generation sequencing (NGS), 18 cDNA libraries via Illumina and two normalized libraries via PacBio, with a total of almost 400,000 reads, were obtained. Our analysis showed that the numbers of isoforms and alternative splicing (AS) patterns were significantly variable between FER and STE. The functional annotation results demonstrate that genes involved in the auxin response, energy metabolism, signal transduction, cell division, and stress response were differentially expressed in different lines. In particular, AUX/IAA, ARF2, SUS, and CYCB had significantly lower expression in STE, showing that auxin might be insufficient in STE, thus hindering nuclear division and influencing metabolism. Apoptosis in STE might also have affected the expression levels of these genes. To confirm the transcriptomic analysis results, nine pairs were confirmed by quantitative real-time PCR. Taken together, these results provide new insights into ovule abortion in gymnosperms and further reveal the regulatory mechanisms of ovule development.

ACME: pan-specific peptide-MHC class I binding prediction through attention-based deep neural networks.

MOTIVATION: Prediction of peptide binding to the major histocompatibility complex (MHC) plays a vital role in the development of therapeutic vaccines for the treatment of cancer. Algorithms with improved correlations between predicted and actual binding affinities are needed to increase precision and reduce the number of false positive predictions. RESULTS: We present ACME (Attention-based Convolutional neural networks for MHC Epitope binding prediction), a new pan-specific algorithm to accurately predict the binding affinities between peptides and MHC class I molecules, even for those new alleles that are not seen in the training data. Extensive tests have demonstrated that ACME can significantly outperform other state-of-the-art prediction methods with an increase of the Pearson correlation coefficient between predicted and measured binding affinities by up to 23 percentage points. In addition, its ability to identify strong-binding peptides has been experimentally validated. Moreover, by integrating the convolutional neural network with attention mechanism, ACME is able to extract interpretable patterns that can provide useful and detailed insights into the binding preferences between peptides and their MHC partners. All these results have demonstrated that ACME can provide a powerful and practically useful tool for the studies of peptide-MHC class I interactions. AVAILABILITY AND IMPLEMENTATION: ACME is available as an open source software at https://github.com/HYsxe/ACME. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

DeepHINT: understanding HIV-1 integration via deep learning with attention.

MOTIVATION: Human immunodeficiency virus type 1 (HIV-1) genome integration is closely related to clinical latency and viral rebound. In addition to human DNA sequences that directly interact with the integration machinery, the selection of HIV integration sites has also been shown to depend on the heterogeneous genomic context around a large region, which greatly hinders the prediction and mechanistic studies of HIV integration. RESULTS: We have developed an attention-based deep learning framework, named DeepHINT, to simultaneously provide accurate prediction of HIV integration sites and mechanistic explanations of the detected sites. Extensive tests on a high-density HIV integration site dataset showed that DeepHINT can outperform conventional modeling strategies by automatically learning the genomic context of HIV integration from primary DNA sequence alone or together with epigenetic information. Systematic analyses on diverse known factors of HIV integration further validated the biological relevance of the prediction results. More importantly, in-depth analyses of the attention values output by DeepHINT revealed intriguing mechanistic implications in the selection of HIV integration sites, including potential roles of several DNA-binding proteins. These results established DeepHINT as an effective and explainable deep learning framework for the prediction and mechanistic study of HIV integration. AVAILABILITY AND IMPLEMENTATION: DeepHINT is available as an open-source software and can be downloaded from https://github.com/nonnerdling/DeepHINT. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Reconstructing spatial organizations of chromosomes through manifold learning.

Decoding the spatial organizations of chromosomes has crucial implications for studying eukaryotic gene regulation. Recently, chromosomal conformation capture based technologies, such as Hi-C, have been widely used to uncover the interaction frequencies of genomic loci in a high-throughput and genome-wide manner and provide new insights into the folding of three-dimensional (3D) genome structure. In this paper, we develop a novel manifold learning based framework, called GEM (Genomic organization reconstructor based on conformational Energy and Manifold learning), to reconstruct the three-dimensional organizations of chromosomes by integrating Hi-C data with biophysical feasibility. Unlike previous methods, which explicitly assume specific relationships between Hi-C interaction frequencies and spatial distances, our model directly embeds the neighboring affinities from Hi-C space into 3D Euclidean space. Extensive validations demonstrated that GEM not only greatly outperformed other state-of-art modeling methods but also provided a physically and physiologically valid 3D representations of the organizations of chromosomes. Furthermore, we for the first time apply the modeled chromatin structures to recover long-range genomic interactions missing from original Hi-C data.

[Xiongcanyishen Prescription increases mitochondrial autophagy of Leydig cells in LOH rats via PINK1 and Parkin pathways].

OBJECTIVE: To investigate the effects of PINK1 and Parkin pathways mediated by Xiongcanyishen Prescription (XP) on the mitochondrial autophagy of Leydig cells in male rats with late-onset hypogonadism (LOH). METHODS: Twenty 18-month-old male SD rats were randomly divided into an LOH model control, a low-dose XP, a medium-dose XP and a high-dose XP group, and another 5 two-month-old male SD rats were included as normal controls. The animals in the low-, medium- and high-dose XP groups were treated intragastrically with XP granules at 10.4, 20.8 and 41.6 g/kg, while the normal and LOH model controls with the same volume of distilled water, all for 28 successive days. Then the testis tissues of the rats were harvested for observation of the ultrastructure of the Leydig cells under the electron microscope, and the expressions of PINK1, Parkin and p62 proteins were detected by Western blot. RESULTS: The mitochondria in the Leydig cells of the normal controls were basically normal in morphology, with evident autophagy, those of the model controls showed less autophagy, and those in the XP intervention groups all exhibited autophagy. The expressions of PINK1 and Parkin proteins in the testis tissue were significantly lower and that of p62 markedly higher in the LOH model than in the normal controls (P < 0.05). Compared with the rats in the model control group, those treated with XP showed remarkable elevation in the expression of PINK1 in the low-, medium- and high-dose groups and that of Parkin in the medium- and high-dose groups (P < 0.05), but a significantly down-regulated expression of p62 in all the three XP groups (P < 0.05). CONCLUSIONS: Xiongcanyishen Prescription can enhance the decreased mitochondrial autophagy of Leydig cells in LOH rats, which may be related to its ability of up-regulating the expressions of PINK1 and Parkin and down-regulating that of p62 and its influence on the ultrastructure of Leydig cells.

DeepShape: estimating isoform-level ribosome abundance and distribution with Ribo-seq data.

BACKGROUND: Ribosome profiling brings insight to the process of translation. A basic step in profile construction at transcript level is to map Ribo-seq data to transcripts, and then assign a huge number of multiple-mapped reads to similar isoforms. Existing methods either discard the multiple mapped-reads, or allocate them randomly, or assign them proportionally according to transcript abundance estimated from RNA-seq data. RESULTS: Here we present DeepShape, an RNA-seq free computational method to estimate ribosome abundance of isoforms, and simultaneously compute their ribosome profiles using a deep learning model. Our simulation results demonstrate that DeepShape can provide more accurate estimations on both ribosome abundance and profiles when compared to state-of-the-art methods. We applied DeepShape to a set of Ribo-seq data from PC3 human prostate cancer cells with and without PP242 treatment. In the four cell invasion/metastasis genes that are translationally regulated by PP242 treatment, different isoforms show very different characteristics of translational efficiency and regulation patterns. Transcript level ribosome distributions were analyzed by "Codon Residence Index (CRI)" proposed in this study to investigate the relative speed that a ribosome moves on a codon compared to its synonymous codons. We observe consistent CRI patterns in PC3 cells. We found that the translation of several codons could be regulated by PP242 treatment. CONCLUSION: In summary, we demonstrate that DeepShape can serve as a powerful tool for Ribo-seq data analysis.

Observation of anterior chamber volume after cataract surgery with swept-source optical coherence tomography.

PURPOSE: To investigate the changes in the anterior chamber volume (ACV) with swept-source optical coherence tomography (SS-OCT) after cataract surgery and the factors that influence these ACV changes. METHODS: This was a prospective cohort study. Fifty-one patients who underwent cataract surgery were enrolled. Their ACV, anterior chamber depth, and angle widths were measured with SS-OCT before and 1 day, 1 week, and 1 month after surgery. The associations between the changes in ACV and posterior vitreous detachment (PVD) and axial length (AXL) were determined. RESULTS: Compared with the preoperative volume, ACV increased significantly at all three time points after surgery (all p < 0.001). ACV was greater at 1 week after surgery than at 1 day after surgery (p < 0.001). Both AXL and the presence of PVD were significantly associated with the change in ACV at 1 day after surgery (p = 0.005). However, neither PVD nor AXL affected the change in ACV between 1 day and 1 week after surgery. CONCLUSIONS: ACV stabilized in the first week after cataract surgery. The absorption of irrigation fluid and balanced salt solution in the vitreous cavity contributed to the change in ACV 1 week after surgery. Eyes with longer AXL and PVD tended to show less change in ACV at 1 day after surgery.

TITER: predicting translation initiation sites by deep learning.

MOTIVATION: Translation initiation is a key step in the regulation of gene expression. In addition to the annotated translation initiation sites (TISs), the translation process may also start at multiple alternative TISs (including both AUG and non-AUG codons), which makes it challenging to predict TISs and study the underlying regulatory mechanisms. Meanwhile, the advent of several high-throughput sequencing techniques for profiling initiating ribosomes at single-nucleotide resolution, e.g. GTI-seq and QTI-seq, provides abundant data for systematically studying the general principles of translation initiation and the development of computational method for TIS identification. METHODS: We have developed a deep learning-based framework, named TITER, for accurately predicting TISs on a genome-wide scale based on QTI-seq data. TITER extracts the sequence features of translation initiation from the surrounding sequence contexts of TISs using a hybrid neural network and further integrates the prior preference of TIS codon composition into a unified prediction framework. RESULTS: Extensive tests demonstrated that TITER can greatly outperform the state-of-the-art prediction methods in identifying TISs. In addition, TITER was able to identify important sequence signatures for individual types of TIS codons, including a Kozak-sequence-like motif for AUG start codon. Furthermore, the TITER prediction score can be related to the strength of translation initiation in various biological scenarios, including the repressive effect of the upstream open reading frames on gene expression and the mutational effects influencing translation initiation efficiency. AVAILABILITY AND IMPLEMENTATION: TITER is available as an open-source software and can be downloaded from https://github.com/zhangsaithu/titer . CONTACT: lzhang20@mail.tsinghua.edu.cn or zengjy321@tsinghua.edu.cn. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

A deep learning framework for modeling structural features of RNA-binding protein targets.

RNA-binding proteins (RBPs) play important roles in the post-transcriptional control of RNAs. Identifying RBP binding sites and characterizing RBP binding preferences are key steps toward understanding the basic mechanisms of the post-transcriptional gene regulation. Though numerous computational methods have been developed for modeling RBP binding preferences, discovering a complete structural representation of the RBP targets by integrating their available structural features in all three dimensions is still a challenging task. In this paper, we develop a general and flexible deep learning framework for modeling structural binding preferences and predicting binding sites of RBPs, which takes (predicted) RNA tertiary structural information into account for the first time. Our framework constructs a unified representation that characterizes the structural specificities of RBP targets in all three dimensions, which can be further used to predict novel candidate binding sites and discover potential binding motifs. Through testing on the real CLIP-seq datasets, we have demonstrated that our deep learning framework can automatically extract effective hidden structural features from the encoded raw sequence and structural profiles, and predict accurate RBP binding sites. In addition, we have conducted the first study to show that integrating the additional RNA tertiary structural features can improve the model performance in predicting RBP binding sites, especially for the polypyrimidine tract-binding protein (PTB), which also provides a new evidence to support the view that RBPs may own specific tertiary structural binding preferences. In particular, the tests on the internal ribosome entry site (IRES) segments yield satisfiable results with experimental support from the literature and further demonstrate the necessity of incorporating RNA tertiary structural information into the prediction model. The source code of our approach can be found in https://github.com/thucombio/deepnet-rbp.

Efficient and accountable anti-leakage attribute-based encryption scheme for cloud storage.

To ensure secure and flexible data sharing in cloud storage, attribute-based encryption (ABE) is introduced to meet the requirements of fine-grained access control and secure one-to-many data sharing. However, the computational burden imposed by attribute encryption renders it unsuitable for resource-constrained environments such as the Internet of Things (IoT) and edge computing. Furthermore, the issue of accountability for illegal keys is crucial, as authorized users may actively disclose or sell authorization keys for personal gain, and keys may also passively leak due to management negligence or hacking incidents. Additionally, since all authorization keys are generated by the attribute authorization center, there is a potential risk of unauthorized key forgery. In response to these challenges, this paper proposes an efficient and accountable leakage-resistant scheme based on attribute encryption. The scheme adopts more secure online/offline encryption mechanisms and cloud server-assisted decryption to alleviate the computational burden on resource-constrained devices. For illegal keys, the scheme supports accountability for both users and the authorization center, allowing the revocation of decryption privileges for malicious users. In the case of passively leaked keys, timely key updates and revocation of decryption capabilities for leaked keys are implemented. Finally, the paper provides selective security and accountability proofs for the scheme under standard models. Efficiency analysis and experimental results demonstrate that the proposed scheme enhances encryption/decryption efficiency, and the storage overhead for accountability is also extremely low.

Photocurrent Enhancement by Copper Incorporation in Chemical-Solution-Synthesized Inorganic Lead Perovskite Thin Films.

Perovskite thin films are at the forefront of highly promising photovoltaic technologies due to their remarkable optoelectronic properties. Herein, we explore a low-cost, reproducible, and industry-scalable methodology to synthesize an all-inorganic CsPbI1.5Br1.5 perovskite thin film with additional incorporation of copper and chloride ions into the lattice structure. The synthesis process involves chemical bath deposition of PbS, followed by a gas-solid iodination reaction to yield PbI2. Subsequently, dip-coating incorporates Cs+, Cu2+, Br-, and Cl- ions into PbI2, and annealing at 270 °C produces perovskite thin films. The results show a large coverage area and a uniform thickness of each perovskite thin film. Comprehensive characterization, including X-ray diffraction, Raman spectroscopy, X-ray photoelectron spectroscopy, scanning electron microscopy, and photoluminescence, provides the structural, chemical, and optical properties of the synthesized thin films. To evaluate the practical implications of our methodology, we fabricated photodetectors employing CsPbI1.5Br1.5 and (Cs0.95:Cu0.01)PbI1.5Br1.3Cl0.1 perovskite films. A comparative analysis unequivocally demonstrates a significant increase in photodetector performance when utilizing (Cs0.95:Cu0.01)PbI1.5Br1.3Cl0.1 perovskite films. While our findings quantitatively assess the tangible enhancement in photocurrent, we acknowledge the potential for improvement in device fabrication to enhance the overall performance. This study not only affirms the successful low-cost synthesis of perovskite thin films but also emphasizes the pivotal role of Cu2+ and Cl- ions in enhancing the performance of perovskite-based optoelectronic devices.

Designed Transmembrane Proteins Inhibit the Erythropoietin Receptor in a Custom Binding Topology.

Transmembrane (TM) domains as simple as a single span can perform complex biological functions using entirely lipid-embedded chemical features. Computational design has potential to generate custom tool molecules directly targeting membrane proteins at their functional TM regions. Thus far, designed TM domain-targeting agents have been limited to mimicking binding modes and motifs of natural TM interaction partners. Here, we demonstrate the design of de novo TM proteins targeting the erythropoietin receptor (EpoR) TM domain in a custom binding topology competitive with receptor homodimerization. The TM proteins expressed in mammalian cells complex with EpoR and inhibit erythropoietin-induced cell proliferation. In vitro, the synthetic TM domain complex outcompetes EpoR homodimerization. Structural characterization reveals that the complex involves the intended amino acids and agrees with our designed molecular model of antiparallel TM helices at 1:1 stoichiometry. Thus, membrane protein TM regions can now be targeted in custom designed topologies.

Interfacial and structural modifications in perovskite solar cells.

The rapid and continuous progress made in perovskite solar cell (PSC) technology has drawn considerable attention from the photovoltaic research community, and the application of perovskites in other electronic devices (such as photodetectors, light-emitting diodes, and batteries) has become imminent. Because of the diversity in device configurations, optimization of film deposition, and exploration of material systems, the power conversion efficiency (PCE) of PSCs has been certified to be as high as 25.2%, making this type of solar cells the fastest advancing technology until now. As demonstrated by researchers worldwide, controlling the morphology and defects in perovskite films is essential for attaining high-performance PSCs. In this regard, interface engineering has proven to be a very efficient way to address these issues, obtaining better charge collection efficiency, and reducing recombination losses. In this review, the interfacial modification between perovskite films and charge-transport layers (CTLs) as well as CTLs and electrodes of PSCs has been widely summarized. Grain boundary (GB) engineering and stress engineering are also included since they are closely related to the improvement in device performance and stability.

DeepCPI: A Deep Learning-based Framework for Large-scale in silico Drug Screening.

Accurate identification of compound-protein interactions (CPIs) in silico may deepen our understanding of the underlying mechanisms of drug action and thus remarkably facilitate drug discovery and development. Conventional similarity- or docking-based computational methods for predicting CPIs rarely exploit latent features from currently available large-scale unlabeled compound and protein data and often limit their usage to relatively small-scale datasets. In the present study, we propose DeepCPI, a novel general and scalable computational framework that combines effective feature embedding (a technique of representation learning) with powerful deep learning methods to accurately predict CPIs at a large scale. DeepCPI automatically learns the implicit yet expressive low-dimensional features of compounds and proteins from a massive amount of unlabeled data. Evaluations of the measured CPIs in large-scale databases, such as ChEMBL and BindingDB, as well as of the known drug-target interactions from DrugBank, demonstrated the superior predictive performance of DeepCPI. Furthermore, several interactions among small-molecule compounds and three G protein-coupled receptor targets (glucagon-like peptide-1 receptor, glucagon receptor, and vasoactive intestinal peptide receptor) predicted using DeepCPI were experimentally validated. The present study suggests that DeepCPI is a useful and powerful tool for drug discovery and repositioning. The source code of DeepCPI can be downloaded from https://github.com/FangpingWan/DeepCPI.