Phase Separation of Disease-Associated SHP2 Mutants Underlies MAPK Hyperactivation.

The non-receptor protein tyrosine phosphatase (PTP) SHP2, encoded by PTPN11, plays an essential role in RAS-mitogen-activated protein kinase (MAPK) signaling during normal development. It has been perplexing as to why both enzymatically activating and inactivating mutations in PTPN11 result in human developmental disorders with overlapping clinical manifestations. Here, we uncover a common liquid-liquid phase separation (LLPS) behavior shared by these disease-associated SHP2 mutants. SHP2 LLPS is mediated by the conserved well-folded PTP domain through multivalent electrostatic interactions and regulated by an intrinsic autoinhibitory mechanism through conformational changes. SHP2 allosteric inhibitors can attenuate LLPS of SHP2 mutants, which boosts SHP2 PTP activity. Moreover, disease-associated SHP2 mutants can recruit and activate wild-type (WT) SHP2 in LLPS to promote MAPK activation. These results not only suggest that LLPS serves as a gain-of-function mechanism involved in the pathogenesis of SHP2-associated human diseases but also provide evidence that PTP may be regulated by LLPS that can be therapeutically targeted.

Observation and quantification of the pseudogap in unitary Fermi gases.

The microscopic origin of high-temperature superconductivity in cuprates remains unknown. It is widely believed that substantial progress could be achieved by better understanding of the pseudogap phase, a normal non-superconducting state of cuprates1,2. In particular, a central issue is whether the pseudogap could originate from strong pairing fluctuations3. Unitary Fermi gases4,5, in which the pseudogap-if it exists-necessarily arises from many-body pairing, offer ideal quantum simulators to address this question. Here we report the observation of a pair-fluctuation-driven pseudogap in homogeneous unitary Fermi gases of lithium-6 atoms, by precisely measuring the fermion spectral function through momentum-resolved microwave spectroscopy and without spurious effects from final-state interactions. The temperature dependence of the pairing gap, inverse pair lifetime and single-particle scattering rate are quantitatively determined by analysing the spectra. We find a large pseudogap above the superfluid transition temperature. The inverse pair lifetime exhibits a thermally activated exponential behaviour, uncovering the microscopic virtual pair breaking and recombination mechanism. The obtained large, temperature-independent single-particle scattering rate is comparable with that set by the Planckian limit6. Our findings quantitatively characterize the pseudogap in strongly interacting Fermi gases and they lend support for the role of preformed pairing as a precursor to superfluidity.

Structural Insights of Transcriptionally Active, Full-Length Androgen Receptor Coactivator Complexes.

Steroid receptors activate gene transcription by recruiting coactivators to initiate transcription of their target genes. For most nuclear receptors, the ligand-dependent activation function domain-2 (AF-2) is a primary contributor to the nuclear receptor (NR) transcriptional activity. In contrast to other steroid receptors, such as ERα, the activation function of androgen receptor (AR) is largely dependent on its ligand-independent AF-1 located in its N-terminal domain (NTD). It remains unclear why AR utilizes a different AF domain from other receptors despite that NRs share similar domain organizations. Here, we present cryoelectron microscopy (cryo-EM) structures of DNA-bound full-length AR and its complex structure with key coactivators, SRC-3 and p300. AR dimerization follows a unique head-to-head and tail-to-tail manner. Unlike ERα, AR directly contacts a single SRC-3 and p300. The AR NTD is the primary site for coactivator recruitment. The structures provide a basis for understanding assembly of the AR:coactivator complex and its domain contributions for coactivator assembly and transcriptional regulation.

TRAF4-mediated nonproteolytic ubiquitination of androgen receptor promotes castration-resistant prostate cancer.

Castration-resistant prostate cancer (CRPC) poses a major clinical challenge with the androgen receptor (AR) remaining to be a critical oncogenic player. Several lines of evidence indicate that AR induces a distinct transcriptional program after androgen deprivation in CRPCs. However, the mechanism triggering AR binding to a distinct set of genomic loci in CRPC and how it promotes CRPC development remain unclear. We demonstrate here that atypical ubiquitination of AR mediated by an E3 ubiquitin ligase TRAF4 plays an important role in this process. TRAF4 is highly expressed in CRPCs and promotes CRPC development. It mediates K27-linked ubiquitination at the C-terminal tail of AR and increases its association with the pioneer factor FOXA1. Consequently, AR binds to a distinct set of genomic loci enriched with FOXA1- and HOXB13-binding motifs to drive different transcriptional programs including an olfactory transduction pathway. Through the surprising upregulation of olfactory receptor gene transcription, TRAF4 increases intracellular cAMP levels and boosts E2F transcription factor activity to promote cell proliferation under androgen deprivation conditions. Altogether, these findings reveal a posttranslational mechanism driving AR-regulated transcriptional reprogramming to provide survival advantages for prostate cancer cells under castration conditions.

High-accuracy protein model quality assessment using attention graph neural networks.

Great improvement has been brought to protein tertiary structure prediction through deep learning. It is important but very challenging to accurately rank and score decoy structures predicted by different models. CASP14 results show that existing quality assessment (QA) approaches lag behind the development of protein structure prediction methods, where almost all existing QA models degrade in accuracy when the target is a decoy of high quality. How to give an accurate assessment to high-accuracy decoys is particularly useful with the available of accurate structure prediction methods. Here we propose a fast and effective single-model QA method, QATEN, which can evaluate decoys only by their topological characteristics and atomic types. Our model uses graph neural networks and attention mechanisms to evaluate global and amino acid level scores, and uses specific loss functions to constrain the network to focus more on high-precision decoys and protein domains. On the CASP14 evaluation decoys, QATEN performs better than other QA models under all correlation coefficients when targeting average LDDT. QATEN shows promising performance when considering only high-accuracy decoys. Compared to the embedded evaluation modules of predicted ${C}_{\alpha^{-}} RMSD$ (pRMSD) in RosettaFold and predicted LDDT (pLDDT) in AlphaFold2, QATEN is complementary and capable of achieving better evaluation on some decoy structures generated by AlphaFold2 and RosettaFold. These results suggest that the new QATEN approach can be used as a reliable independent assessment algorithm for high-accuracy protein structure decoys.

Leveraging scaffold information to predict protein-ligand binding affinity with an empirical graph neural network.

Protein-ligand binding affinity prediction is an important task in structural bioinformatics for drug discovery and design. Although various scoring functions (SFs) have been proposed, it remains challenging to accurately evaluate the binding affinity of a protein-ligand complex with the known bound structure because of the potential preference of scoring system. In recent years, deep learning (DL) techniques have been applied to SFs without sophisticated feature engineering. Nevertheless, existing methods cannot model the differential contribution of atoms in various regions of proteins, and the relationship between atom properties and intermolecular distance is also not fully explored. We propose a novel empirical graph neural network for accurate protein-ligand binding affinity prediction (EGNA). Graphs of protein, ligand and their interactions are constructed based on different regions of each bound complex. Proteins and ligands are effectively represented by graph convolutional layers, enabling the EGNA to capture interaction patterns precisely by simulating empirical SFs. The contributions of different factors on binding affinity can thus be transparently investigated. EGNA is compared with the state-of-the-art machine learning-based SFs on two widely used benchmark data sets. The results demonstrate the superiority of EGNA and its good generalization capability.

MINDG: a drug-target interaction prediction method based on an integrated learning algorithm.

MOTIVATION: Drug-target interaction (DTI) prediction refers to the prediction of whether a given drug molecule will bind to a specific target and thus exert a targeted therapeutic effect. Although intelligent computational approaches for drug target prediction have received much attention and made many advances, they are still a challenging task that requires further research. The main challenges are manifested as follows: (i) most graph neural network-based methods only consider the information of the first-order neighboring nodes (drug and target) in the graph, without learning deeper and richer structural features from the higher-order neighboring nodes. (ii) Existing methods do not consider both the sequence and structural features of drugs and targets, and each method is independent of each other, and cannot combine the advantages of sequence and structural features to improve the interactive learning effect. RESULTS: To address the above challenges, a Multi-view Integrated learning Network that integrates Deep learning and Graph Learning (MINDG) is proposed in this study, which consists of the following parts: (i) a mixed deep network is used to extract sequence features of drugs and targets, (ii) a higher-order graph attention convolutional network is proposed to better extract and capture structural features, and (iii) a multi-view adaptive integrated decision module is used to improve and complement the initial prediction results of the above two networks to enhance the prediction performance. We evaluate MINDG on two dataset and show it improved DTI prediction performance compared to state-of-the-art baselines. AVAILABILITY AND IMPLEMENTATION: https://github.com/jnuaipr/MINDG.

Comprehensive molecular classification predicted microenvironment profiles and therapy response for HCC.

BACKGROUND AND AIMS: Tumor microenvironment (TME) heterogeneity leads to a discrepancy in survival prognosis and clinical treatment response for patients with HCC. The clinical applications of documented molecular subtypes are constrained by several issues. APPROACH AND RESULTS: We integrated 3 single-cell data sets to describe the TME landscape and identified 6 prognosis-related cell subclusters. Unsupervised clustering of subcluster-specific markers was performed to generate transcriptomic subtypes. The predictive value of these molecular subtypes for prognosis and treatment response was explored in multiple external HCC cohorts and the Xiangya HCC cohort. TME features were estimated using single-cell immune repertoire sequencing, mass cytometry, and multiplex immunofluorescence. The prognosis-related score was constructed based on a machine-learning algorithm. Comprehensive single-cell analysis described TME heterogeneity in HCC. The 5 transcriptomic subtypes possessed different clinical prognoses, stemness characteristics, immune landscapes, and therapeutic responses. Class 1 exhibited an inflamed phenotype with better clinical outcomes, while classes 2 and 4 were characterized by a lack of T-cell infiltration. Classes 5 and 3 indicated an inhibitory tumor immune microenvironment. Analysis of multiple therapeutic cohorts suggested that classes 5 and 3 were sensitive to immune checkpoint blockade and targeted therapy, whereas classes 1 and 2 were more responsive to transcatheter arterial chemoembolization treatment. Class 4 displayed resistance to all conventional HCC therapies. Four potential therapeutic agents and 4 targets were further identified for high prognosis-related score patients with HCC. CONCLUSIONS: Our study generated a clinically valid molecular classification to guide precision medicine in patients with HCC.

UPK3A+ umbrella cell damage mediated by TLR3-NR2F6 triggers programmed destruction of urothelium in Hunner-type interstitial cystitis/painful bladder syndrome.

Urothelial damage and barrier dysfunction emerge as the foremost mechanisms in Hunner-type interstitial cystitis/bladder pain syndrome (HIC). Although treatments aimed at urothelial regeneration and repair have been employed, their therapeutic effectiveness remains limited due to the inadequate understanding of specific cell types involved in damage and the lack of specific molecular targets within these mechanisms. Therefore, we harnessed single-cell RNA sequencing to elucidate the heterogeneity and developmental trajectory of urothelial cells within HIC bladders. Through reclustering, we identified eight distinct clusters of urothelial cells. There was a significant reduction in UPK3A+ umbrella cells and a simultaneous increase in progenitor-like pluripotent cells (PPCs) within the HIC bladder. Pseudotime analysis of the urothelial cells in the HIC bladder revealed that cells faced challenges in differentiating into UPK3A+ umbrella cells, while PPCs exhibited substantial proliferation to compensate for the loss of UPK3A+ umbrella cells. The urothelium in HIC remains unrepaired, despite the substantial proliferation of PPCs. Thus, we propose that inhibiting the pivotal signaling pathways responsible for the injury to UPK3A+ umbrella cells is paramount for restoring the urothelial barrier and alleviating lower urinary tract symptoms in HIC patients. Subsequently, we identified key molecular pathways (TLR3 and NR2F6) associated with the injury of UPK3A+ umbrella cells in HIC urothelium. Finally, we conducted in vitro and in vivo experiments to confirm the potential of the TLR3-NR2F6 axis as a promising therapeutic target for HIC. These findings hold the potential to inhibit urothelial injury, providing promising clues for early diagnosis and functional bladder self-repair strategies for HIC patients. © 2024 The Pathological Society of Great Britain and Ireland.

Structural and Functional Impacts of ER Coactivator Sequential Recruitment.

Nuclear receptors recruit multiple coactivators sequentially to activate transcription. This "ordered" recruitment allows different coactivator activities to engage the nuclear receptor complex at different steps of transcription. Estrogen receptor (ER) recruits steroid receptor coactivator-3 (SRC-3) primary coactivator and secondary coactivators, p300/CBP and CARM1. CARM1 recruitment lags behind the binding of SRC-3 and p300 to ER. Combining cryo-electron microscopy (cryo-EM) structure analysis and biochemical approaches, we demonstrate that there is a close crosstalk between early- and late-recruited coactivators. The sequential recruitment of CARM1 not only adds a protein arginine methyltransferase activity to the ER-coactivator complex, it also alters the structural organization of the pre-existing ERE/ERα/SRC-3/p300 complex. It induces a p300 conformational change and significantly increases p300 HAT activity on histone H3K18 residues, which, in turn, promotes CARM1 methylation activity on H3R17 residues to enhance transcriptional activity. This study reveals a structural role for a coactivator sequential recruitment and biochemical process in ER-mediated transcription.

DeepQs: Local quality assessment of cryo-EM density map by deep learning map-model fit score.

Cryogenic electron microscopy maps are valuable for determining macromolecule structures. A proper quality assessment method is essential for cryo-EM map selection or revision. This article presents DeepQs, a novel approach to estimate local quality for 3D cryo-EM density maps, using a deep-learning algorithm based on map-model fit score. DeepQs is a parameter-free method for users and incorporates structural information between map and its related atomic model into well-trained models by deep learning. More specifically, the DeepQs approach leverages the interplay between map and atomic model through predefined map-model fit score, Q-score. DeepQs can get close results to the ground truth map-model fit scores with only cryo-EM map as input. In experiments, DeepQs demonstrates the lowest root mean square error with standard method Fourier shell correlation metric and high correlation with map-model fit score, Q-score, when compared with other local quality estimation methods in high-resolution dataset (<=5 Å). DeepQs can also be applied to evaluate the quality of the post-processed maps. In both cases, DeepQs runs faster by using GPU acceleration. Our program is available at http://www.csbio.sjtu.edu.cn/bioinf/DeepQs for academic use.

UBC9-mediated SUMOylation of Lamin B1 enhances DNA-damage-induced nuclear DNA leakage and autophagy after spinal cord injury.

Recent studies have shown that nucleophagy can mitigate DNA damage by selectively degrading nuclear components protruding from the nucleus. However, little is known about the role of nucleophagy in neurons after spinal cord injury (SCI). Western blot analysis and immunofluorescence were performed to evaluate the nucleophagy after nuclear DNA damage and leakage in SCI neurons in vivo and NSC34 expression in primary neurons cultured with oxygen-glucose deprivation (OGD) in vitro, as well as the interaction and colocalization of autophagy protein LC3 with nuclear lamina protein Lamin B1. The effect of UBC9, a Small ubiquitin-related modifier (SUMO) E2 ligase, on Lamin B1 SUMOylation and nucleophagy was examined by siRNA transfection or 2-D08 (a small-molecule inhibitor of UBC9), immunoprecipitation, and immunofluorescence. In SCI and OGD injured NSC34 or primary cultured neurons, neuronal nuclear DNA damage induced the SUMOylation of Lamin B1, which was required by the nuclear Lamina accumulation of UBC9. Furthermore, LC3/Atg8, an autophagy-related protein, directly bound to SUMOylated Lamin B1, and delivered Lamin B1 to the lysosome. Knockdown or suppression of UBC9 with siRNA or 2-D08 inhibited SUMOylation of Lamin B1 and subsequent nucleophagy and protected against neuronal death. Upon neuronal DNA damage and leakage after SCI, SUMOylation of Lamin B1 is induced by nuclear Lamina accumulation of UBC9. Furthermore, it promotes LC3-Lamin B1 interaction to trigger nucleophagy that protects against neuronal DNA damage.

Simultaneous Stereoinvertive and Stereoselective C(sp3)-C(sp3) Cross-Coupling of Boronic Esters and Allylic Carbonates.

With increasing interest in constructing more three-dimensional entities, there has been growing interest in cross-coupling reactions that forge C(sp3)-C(sp3) bonds, which leads to additional challenges as it is not just a more difficult bond to construct but issues of stereocontrol also arise. Herein, we report the stereocontrolled cross-coupling of enantioenriched boronic esters with racemic allylic carbonates enabled by iridium catalysis, leading to the formation of C(sp3)-C(sp3) bonds with single or vicinal stereogenic centers. The method shows broad substrate scope, enabling primary, secondary, and even tertiary boronic esters to be employed, and can be used to prepare any of the four possible stereoisomers of a coupled product with vicinal chiral centers. The new method, which combines the simultaneous enantiospecific reaction of a chiral nucleophile with the enantioselective reaction of a chiral electrophile in a single process, offers a solution for stereodivergent cross-coupling of two C(sp3) fragments.

A Pilot Study To Assess the Suitability of Riboflavin As a Surrogate Marker of Breast Cancer Resistance Protein in Healthy Participants.

We recently showed that riboflavin is a selected substrate of breast cancer resistance protein (BCRP) over P-glycoprotein (P-gp) and demonstrated its prediction performance in preclinical drug-drug interaction (DDI) studies. The aim of this study was to investigate the suitability of riboflavin to assess BCRP inhibition in humans. First, we assessed the substrate potential of riboflavin toward other major drug transporters using established transfected cell systems. Riboflavin is a substrate for organic anion transporter (OAT)1, OAT3, and multidrug and toxin extrusion protein (MATE)2-K, with uptake ratios ranging from 2.69 to 11.6, but riboflavin is not a substrate of organic anion-transporting polypeptide (OATP)1B1, OATP1B3, organic cation transporter (OCT)2, and MATE1. The effects of BMS-986371, a potent in vitro inhibitor of BCRP (IC 50 0.40 µM), on the pharmacokinetics of riboflavin, isobutyryl carnitine, and arginine were then examined in healthy male adults (N = 14 or 16) after oral administration of methotrexate (MTX) (7.5 mg) and enteric-coated (EC) sulfasalazine (SSZ) (1000 mg) alone or in combination with BMS-986371 (150 mg). Oral administration of BMS-986371 increased the area under the plasma concentration-time curves (AUCs) of rosuvastatin and immediate-release (IR) SSZ to 1.38- and 1.51-fold, respectively, and significantly increased AUC(0-4h), AUC(0-24h), and C max of riboflavin by 1.25-, 1.14-, and 1.11-fold (P-values of 0.003, 0.009, and 0.025, respectively) compared with the MTX/SSZ EC alone group. In contrast, BMS-986371 did not significantly influence the AUC(0-24h) and C max values of isobutyryl carnitine and arginine (0.96- to 1.07-fold, respectively; P > 0.05). Overall, these data indicate that plasma riboflavin is a promising biomarker of BCRP that may offer a possibility to assess drug candidate as a BCRP modulator in early drug development. SIGNIFICANCE STATEMENT: Endogenous compounds that serve as biomarkers for clinical inhibition of breast cancer resistance protein (BCRP) are not currently available. This study provides the initial evidence that riboflavin is a promising BCRP biomarker in humans. For the first time, the value of leveraging the substrate of BCRP with acceptable prediction performance in clinical studies is shown. Additional clinical investigations with known BCRP inhibitors are needed to fully validate and showcase the utility of this biomarker.

SIFLoc: a self-supervised pre-training method for enhancing the recognition of protein subcellular localization in immunofluorescence microscopic images.

With the rapid growth of high-resolution microscopy imaging data, revealing the subcellular map of human proteins has become a central task in the spatial proteome. The cell atlas of the Human Protein Atlas (HPA) provides precious resources for recognizing subcellular localization patterns at the cell level, and the large-scale annotated data enable learning via advanced deep neural networks. However, the existing predictors still suffer from the imbalanced class distribution and the lack of labeled data for minor classes. Thus, it is necessary to develop new methods for coping with these issues. We leverage the self-supervised learning protocol to address these problems. Especially, we propose a pre-training scheme to enhance the conventional supervised learning framework called SIFLoc. The pre-training is featured by a hybrid data augmentation method and a modified contrastive loss function, aiming to learn good feature representations from microscopic images. The experiments are performed on a large-scale immunofluorescence microscopic image dataset collected from the HPA database. Using the same deep neural networks as the classifier, the model pre-trained via SIFLoc not only outperforms the model without pre-training by a large margin but also shows advantages over the state-of-the-art self-supervised learning methods. Especially, SIFLoc improves the prediction accuracy for minor organelles significantly.

Learning protein subcellular localization multi-view patterns from heterogeneous data of imaging, sequence and networks.

Location proteomics seeks to provide automated high-resolution descriptions of protein location patterns within cells. Many efforts have been undertaken in location proteomics over the past decades, thereby producing plenty of automated predictors for protein subcellular localization. However, most of these predictors are trained solely from high-throughput microscopic images or protein amino acid sequences alone. Unifying heterogeneous protein data sources has yet to be exploited. In this paper, we present a pipeline called sequence, image, network-based protein subcellular locator (SIN-Locator) that constructs a multi-view description of proteins by integrating multiple data types including images of protein expression in cells or tissues, amino acid sequences and protein-protein interaction networks, to classify the patterns of protein subcellular locations. Proteins were encoded by both handcrafted features and deep learning features, and multiple combining methods were implemented. Our experimental results indicated that optimal integrations can considerately enhance the classification accuracy, and the utility of SIN-Locator has been demonstrated through applying to new released proteins in the human protein atlas. Furthermore, we also investigate the contribution of different data sources and influence of partial absence of data. This work is anticipated to provide clues for reconciliation and combination of multi-source data for protein location analysis.

Accurate flexible refinement for atomic-level protein structure using cryo-EM density maps and deep learning.

With the rapid progress of deep learning in cryo-electron microscopy and protein structure prediction, improving the accuracy of the protein structure model by using a density map and predicted contact/distance map through deep learning has become an urgent need for robust methods. Thus, designing an effective protein structure optimization strategy based on the density map and predicted contact/distance map is critical to improving the accuracy of structure refinement. In this article, a protein structure optimization method based on the density map and predicted contact/distance map by deep-learning technology was proposed in accordance with the result of matching between the density map and the initial model. Physics- and knowledge-based energy functions, integrated with Cryo-EM density map data and deep-learning data, were used to optimize the protein structure in the simulation. The dynamic confidence score was introduced to the iterative process for choosing whether it is a density map or a contact/distance map to dominate the movement in the simulation to improve the accuracy of refinement. The protocol was tested on a large set of 224 non-homologous membrane proteins and generated 214 structural models with correct folds, where 4.5% of structural models were generated from structural models with incorrect folds. Compared with other state-of-the-art methods, the major advantage of the proposed methods lies in the skills for using density map and contact/distance map in the simulation, as well as the new energy function in the re-assembly simulations. Overall, the results demonstrated that this strategy is a valuable approach and ready to use for atomic-level structure refinement using cryo-EM density map and predicted contact/distance map.

circRNA-binding protein site prediction based on multi-view deep learning, subspace learning and multi-view classifier.

Circular RNAs (circRNAs) generally bind to RNA-binding proteins (RBPs) to play an important role in the regulation of autoimmune diseases. Thus, it is crucial to study the binding sites of RBPs on circRNAs. Although many methods, including traditional machine learning and deep learning, have been developed to predict the interactions between RNAs and RBPs, and most of them are focused on linear RNAs. At present, few studies have been done on the binding relationships between circRNAs and RBPs. Thus, in-depth research is urgently needed. In the existing circRNA-RBP binding site prediction methods, circRNA sequences are the main research subjects, but the relevant characteristics of circRNAs have not been fully exploited, such as the structure and composition information of circRNA sequences. Some methods have extracted different views to construct recognition models, but how to efficiently use the multi-view data to construct recognition models is still not well studied. Considering the above problems, this paper proposes a multi-view classification method called DMSK based on multi-view deep learning, subspace learning and multi-view classifier for the identification of circRNA-RBP interaction sites. In the DMSK method, first, we converted circRNA sequences into pseudo-amino acid sequences and pseudo-dipeptide components for extracting high-dimensional sequence features and component features of circRNAs, respectively. Then, the structure prediction method RNAfold was used to predict the secondary structure of the RNA sequences, and the sequence embedding model was used to extract the context-dependent features. Next, we fed the above four views' raw features to a hybrid network, which is composed of a convolutional neural network and a long short-term memory network, to obtain the deep features of circRNAs. Furthermore, we used view-weighted generalized canonical correlation analysis to extract four views' common features by subspace learning. Finally, the learned subspace common features and multi-view deep features were fed to train the downstream multi-view TSK fuzzy system to construct a fuzzy rule and fuzzy inference-based multi-view classifier. The trained classifier was used to predict the specific positions of the RBP binding sites on the circRNAs. The experiments show that the prediction performance of the proposed method DMSK has been improved compared with the existing methods. The code and dataset of this study are available at https://github.com/Rebecca3150/DMSK.

MDGF-MCEC: a multi-view dual attention embedding model with cooperative ensemble learning for CircRNA-disease association prediction.

Circular RNA (circRNA) is closely involved in physiological and pathological processes of many diseases. Discovering the associations between circRNAs and diseases is of great significance. Due to the high-cost to verify the circRNA-disease associations by wet-lab experiments, computational approaches for predicting the associations become a promising research direction. In this paper, we propose a method, MDGF-MCEC, based on multi-view dual attention graph convolution network (GCN) with cooperative ensemble learning to predict circRNA-disease associations. First, MDGF-MCEC constructs two disease relation graphs and two circRNA relation graphs based on different similarities. Then, the relation graphs are fed into a multi-view GCN for representation learning. In order to learn high discriminative features, a dual-attention mechanism is introduced to adjust the contribution weights, at both channel level and spatial level, of different features. Based on the learned embedding features of diseases and circRNAs, nine different feature combinations between diseases and circRNAs are treated as new multi-view data. Finally, we construct a multi-view cooperative ensemble classifier to predict the associations between circRNAs and diseases. Experiments conducted on the CircR2Disease database demonstrate that the proposed MDGF-MCEC model achieves a high area under curve of 0.9744 and outperforms the state-of-the-art methods. Promising results are also obtained from experiments on the circ2Disease and circRNADisease databases. Furthermore, the predicted associated circRNAs for hepatocellular carcinoma and gastric cancer are supported by the literature. The code and dataset of this study are available at https://github.com/ABard0/MDGF-MCEC.

De novo drug design by iterative multiobjective deep reinforcement learning with graph-based molecular quality assessment.

MOTIVATION: Generating molecules of high quality and drug-likeness in the vast chemical space is a big challenge in the drug discovery. Most existing molecule generative methods focus on diversity and novelty of molecules, but ignoring drug potentials of the generated molecules during the generation process. RESULTS: In this study, we present a novel de novo multiobjective quality assessment-based drug design approach (QADD), which integrates an iterative refinement framework with a novel graph-based molecular quality assessment model on drug potentials. QADD designs a multiobjective deep reinforcement learning pipeline to generate molecules with multiple desired properties iteratively, where a graph neural network-based model for accurate molecular quality assessment on drug potentials is introduced to guide molecule generation. Experimental results show that QADD can jointly optimize multiple molecular properties with a promising performance and the quality assessment module is capable of guiding the generated molecules with high drug potentials. Furthermore, applying QADD to generate novel molecules binding to a biological target protein DRD2 also demonstrates the algorithm's efficacy. AVAILABILITY AND IMPLEMENTATION: QADD is freely available online for academic use at https://github.com/yifang000/QADD or http://www.csbio.sjtu.edu.cn/bioinf/QADD.