The Chinese pine genome and methylome unveil key features of conifer evolution.

Conifers dominate the world's forest ecosystems and are the most widely planted tree species. Their giant and complex genomes present great challenges for assembling a complete reference genome for evolutionary and genomic studies. We present a 25.4-Gb chromosome-level assembly of Chinese pine (Pinus tabuliformis) and revealed that its genome size is mostly attributable to huge intergenic regions and long introns with high transposable element (TE) content. Large genes with long introns exhibited higher expressions levels. Despite a lack of recent whole-genome duplication, 91.2% of genes were duplicated through dispersed duplication, and expanded gene families are mainly related to stress responses, which may underpin conifers' adaptation, particularly in cold and/or arid conditions. The reproductive regulation network is distinct compared with angiosperms. Slow removal of TEs with high-level methylation may have contributed to genomic expansion. This study provides insights into conifer evolution and resources for advancing research on conifer adaptation and development.

Single-cell transcriptomic analysis identifies an immune-prone population in erythroid precursors during human ontogenesis.

Nonimmune cells can have immunomodulatory roles that contribute to healthy development. However, the molecular and cellular mechanisms underlying the immunomodulatory functions of erythroid cells during human ontogenesis remain elusive. Here, integrated, single-cell transcriptomic studies of erythroid cells from the human yolk sac, fetal liver, preterm umbilical cord blood (UCB), term UCB and adult bone marrow (BM) identified classical and immune subsets of erythroid precursors with divergent differentiation trajectories. Immune-erythroid cells were present from the yolk sac to the adult BM throughout human ontogenesis but failed to be generated in vitro from human embryonic stem cells. Compared with classical-erythroid precursors, these immune-erythroid cells possessed dual erythroid and immune regulatory networks, showed immunomodulatory functions and interacted more frequently with various innate and adaptive immune cells. Our findings provide important insights into the nature of immune-erythroid cells and their roles during development and diseases.

Lysosomal cyst(e)ine storage potentiates tolerance to oxidative stress in cancer cells.

Cyst(e)ine is a key precursor for the synthesis of glutathione (GSH), which protects cancer cells from oxidative stress. Cyst(e)ine is stored in lysosomes, but its role in redox regulation is unclear. Here, we show that breast cancer cells upregulate major facilitator superfamily domain containing 12 (MFSD12) to increase lysosomal cyst(e)ine storage, which is released by cystinosin (CTNS) to maintain GSH levels and buffer oxidative stress. We find that mTORC1 regulates MFSD12 by directly phosphorylating residue T254, while mTORC1 inhibition enhances lysosome acidification that activates CTNS. This switch modulates lysosomal cyst(e)ine levels in response to oxidative stress, fine-tuning redox homeostasis to enhance cell fitness. MFSD12-T254A mutant inhibits MFSD12 function and suppresses tumor progression. Moreover, MFSD12 overexpression correlates with poor neoadjuvant chemotherapy response and prognosis in breast cancer patients. Our findings reveal the critical role of lysosomal cyst(e)ine storage in adaptive redox homeostasis and suggest that MFSD12 is a potential therapeutic target.

Integration of light and temperature sensing by liquid-liquid phase separation of phytochrome B.

Light and temperature in plants are perceived by a common receptor, phytochrome B (phyB). How phyB distinguishes these signals remains elusive. Here, we report that phyB spontaneously undergoes phase separation to assemble liquid-like droplets. This capacity is driven by its C terminus through self-association, whereas the intrinsically disordered N-terminal extension (NTE) functions as a biophysical modulator of phase separation. Light exposure triggers a conformational change to subsequently alter phyB condensate assembly, while temperature sensation is directly mediated by the NTE to modulate the phase behavior of phyB droplets. Multiple signaling components are selectively incorporated into phyB droplets to form concentrated microreactors, allowing switch-like control of phyB signaling activity through phase transitions. Therefore, light and temperature cues are separately read out by phyB via allosteric changes and spontaneous phase separation, respectively. We provide a conceptual framework showing how the distinct but highly correlated physical signals are interpreted and sorted by one receptor.

Structural insight into the allosteric inhibition of human sodium-calcium exchanger NCX1 by XIP and SEA0400.

Sodium-calcium exchanger proteins influence calcium homeostasis in many cell types and participate in a wide range of physiological and pathological processes. Here, we elucidate the cryo-EM structure of the human Na+/Ca2+ exchanger NCX1.3 in the presence of a specific inhibitor, SEA0400. Conserved ion-coordinating residues are exposed on the cytoplasmic face of NCX1.3, indicating that the observed structure is stabilized in an inward-facing conformation. We show how regulatory calcium-binding domains (CBDs) assemble with the ion-translocation transmembrane domain (TMD). The exchanger-inhibitory peptide (XIP) is trapped within a groove between the TMD and CBD2 and predicted to clash with gating helices TMs1/6 at the outward-facing state, thus hindering conformational transition and promoting inactivation of the transporter. A bound SEA0400 molecule stiffens helix TM2ab and affects conformational rearrangements of TM2ab that are associated with the ion-exchange reaction, thus allosterically attenuating Ca2+-uptake activity of NCX1.3.

Dual action of ketamine confines addiction liability.

Ketamine is used clinically as an anaesthetic and a fast-acting antidepressant, and recreationally for its dissociative properties, raising concerns of addiction as a possible side effect. Addictive drugs such as cocaine increase the levels of dopamine in the nucleus accumbens. This facilitates synaptic plasticity in the mesolimbic system, which causes behavioural adaptations and eventually drives the transition to compulsion1-4. The addiction liability of ketamine is a matter of much debate, in part because of its complex pharmacology that among several targets includes N-methyl-D-aspartic acid (NMDA) receptor (NMDAR) antagonism5,6. Here we show that ketamine does not induce the synaptic plasticity that is typically observed with addictive drugs in mice, despite eliciting robust dopamine transients in the nucleus accumbens. Ketamine nevertheless supported reinforcement through the disinhibition of dopamine neurons in the ventral tegmental area (VTA). This effect was mediated by NMDAR antagonism in GABA (γ-aminobutyric acid) neurons of the VTA, but was quickly terminated by type-2 dopamine receptors on dopamine neurons. The rapid off-kinetics of the dopamine transients along with the NMDAR antagonism precluded the induction of synaptic plasticity in the VTA and the nucleus accumbens, and did not elicit locomotor sensitization or uncontrolled self-administration. In summary, the dual action of ketamine leads to a unique constellation of dopamine-driven positive reinforcement, but low addiction liability.

Generation of complex bone marrow organoids from human induced pluripotent stem cells.

The human bone marrow (BM) niche sustains hematopoiesis throughout life. We present a method for generating complex BM-like organoids (BMOs) from human induced pluripotent stem cells (iPSCs). BMOs consist of key cell types that self-organize into spatially defined three-dimensional structures mimicking cellular, structural and molecular characteristics of the hematopoietic microenvironment. Functional properties of BMOs include the presence of an in vivo-like vascular network, the presence of multipotent mesenchymal stem/progenitor cells, the support of neutrophil differentiation and responsiveness to inflammatory stimuli. Single-cell RNA sequencing revealed a heterocellular composition including the presence of a hematopoietic stem/progenitor (HSPC) cluster expressing genes of fetal HSCs. BMO-derived HSPCs also exhibited lymphoid potential and a subset demonstrated transient engraftment potential upon xenotransplantation in mice. We show that the BMOs could enable the modeling of hematopoietic developmental aspects and inborn errors of hematopoiesis, as shown for human VPS45 deficiency. Thus, iPSC-derived BMOs serve as a physiologically relevant in vitro model of the human BM microenvironment to study hematopoietic development and BM diseases.

Cbl Ubiquitin Ligases Control B Cell Exit from the Germinal-Center Reaction.

Selective expansion of high-affinity antigen-specific B cells in germinal centers (GCs) is a key event in antibody affinity maturation. GC B cells with improved affinity can either continue affinity-driven selection or exit the GC to differentiate into plasma cells (PCs) or memory B cells. Here we found that deleting E3 ubiquitin ligases Cbl and Cbl-b (Cbls) in GC B cells resulted in the early exit of high-affinity antigen-specific B cells from the GC reaction and thus impaired clonal expansion. Cbls were highly expressed in GC light zone (LZ) B cells, where they promoted the ubiquitination and degradation of Irf4, a transcription factor facilitating PC fate choice. Strong CD40 and BCR stimulation triggered the Cbl degradation, resulting in increased Irf4 expression and exit from GC affinity selection. Thus, a regulatory cascade that is centered on the Cbl ubiquitin ligases ensures affinity-driven clonal expansion by connecting BCR affinity signals with differentiation programs.

Thymus aging and immune reconstitution, progresses and challenges.

Thymus is a primary lymphoid organ essential for the development of T lymphocytes. Age-related thymic involution is a prominent feature of immune senescence. The thymus undergoes rapid growth during fetal and neonatal development, peaks in size before puberty and then begins to undergo a decrease in cellularity with age. Dramatic changes occur with age-associated thymic involution. The most prominent features of thymic involution include: (i) epithelial structure disruption, (ii) adipogenesis, and (iii) thymocyte development arrest. There is a sex disparity in thymus aging. It is a multifactorial process controlled and regulated by a series of molecules, including the transcription factor FOXN1, fibroblast and keratinocyte growth factors (FGF and KGF, respectively), sex steroids, Notch signaling, WNT signaling, and microRNAs. Nevertheless, there is still no satisfactory evolutionary or physiological explanation for age-associated thymic involution, and understanding the precise mechanism(s) for thymus aging remains challenging. Sustained thymic regeneration has yet to be achieved by sex steroid ablation. Recent preclinical studies indicate that long-term thymic reconstitution can be achieved via adoptive transfer of in vitro-generated progenitor T (proT) cells, and improvements in the methods for the generation of human proT cells make this an attractive approach. Future clinical applications may rely on new applications integrating proT cells, cytokine support and sex-steroid inhibition treatments.

Fast and accurate Bayesian polygenic risk modeling with variational inference.

The advent of large-scale genome-wide association studies (GWASs) has motivated the development of statistical methods for phenotype prediction with single-nucleotide polymorphism (SNP) array data. These polygenic risk score (PRS) methods use a multiple linear regression framework to infer joint effect sizes of all genetic variants on the trait. Among the subset of PRS methods that operate on GWAS summary statistics, sparse Bayesian methods have shown competitive predictive ability. However, most existing Bayesian approaches employ Markov chain Monte Carlo (MCMC) algorithms, which are computationally inefficient and do not scale favorably to higher dimensions, for posterior inference. Here, we introduce variational inference of polygenic risk scores (VIPRS), a Bayesian summary statistics-based PRS method that utilizes variational inference techniques to approximate the posterior distribution for the effect sizes. Our experiments with 36 simulation configurations and 12 real phenotypes from the UK Biobank dataset demonstrated that VIPRS is consistently competitive with the state-of-the-art in prediction accuracy while being more than twice as fast as popular MCMC-based approaches. This performance advantage is robust across a variety of genetic architectures, SNP heritabilities, and independent GWAS cohorts. In addition to its competitive accuracy on the "White British" samples, VIPRS showed improved transferability when applied to other ethnic groups, with up to 1.7-fold increase in R2 among individuals of Nigerian ancestry for low-density lipoprotein (LDL) cholesterol. To illustrate its scalability, we applied VIPRS to a dataset of 9.6 million genetic markers, which conferred further improvements in prediction accuracy for highly polygenic traits, such as height.

Image-based molecular representation learning for drug development: a survey.

Artificial intelligence (AI) powered drug development has received remarkable attention in recent years. It addresses the limitations of traditional experimental methods that are costly and time-consuming. While there have been many surveys attempting to summarize related research, they only focus on general AI or specific aspects such as natural language processing and graph neural network. Considering the rapid advance on computer vision, using the molecular image to enable AI appears to be a more intuitive and effective approach since each chemical substance has a unique visual representation. In this paper, we provide the first survey on image-based molecular representation for drug development. The survey proposes a taxonomy based on the learning paradigms in computer vision and reviews a large number of corresponding papers, highlighting the contributions of molecular visual representation in drug development. Besides, we discuss the applications, limitations and future directions in the field. We hope this survey could offer valuable insight into the use of image-based molecular representation learning in the context of drug development.

The innate immunity protein IFITM3 modulates γ-secretase in Alzheimer's disease.

Innate immunity is associated with Alzheimer's disease1, but the influence of immune activation on the production of amyloid-ß is unknown2,3. Here we identify interferon-induced transmembrane protein 3 (IFITM3) as a γ-secretase modulatory protein, and establish a mechanism by which inflammation affects the generation of amyloid-ß. Inflammatory cytokines induce the expression of IFITM3 in neurons and astrocytes, which binds to γ-secretase and upregulates its activity, thereby increasing the production of amyloid-ß. The expression of IFITM3 is increased with ageing and in mouse models that express familial Alzheimer's disease genes. Furthermore, knockout of IFITM3 reduces γ-secretase activity and the formation of amyloid plaques in a transgenic mouse model (5xFAD) of early amyloid deposition. IFITM3 protein is upregulated in tissue samples from a subset of patients with late-onset Alzheimer's disease that exhibit higher γ-secretase activity. The amount of IFITM3 in the γ-secretase complex has a strong and positive correlation with γ-secretase activity in samples from patients with late-onset Alzheimer's disease. These findings reveal a mechanism in which γ-secretase is modulated by neuroinflammation via IFITM3 and the risk of Alzheimer's disease is thereby increased.

GDPF: a data resource for the distribution of prokaryotic protein families across the global biosphere.

Microorganisms encode most of the functions of life on Earth. However, conventional research has primarily focused on specific environments such as humans, soil and oceans, leaving the distribution of functional families throughout the global biosphere poorly comprehended. Here, we present the database of the global distribution of prokaryotic protein families (GDPF, http://bioinfo.qd.sdu.edu.cn/GDPF/), a data resource on the distribution of functional families across the global biosphere. GDPF provides global distribution information for 36 334 protein families, 19 734 superfamilies and 12 089 KEGG (Kyoto Encyclopedia of Genes and Genomes) orthologs from multiple source databases, covering typical environments such as soil, oceans, animals, plants and sediments. Users can browse, search and download the distribution data of each entry in 10 000 global microbial communities, as well as conduct comparative analysis of distribution disparities among multiple entries across various environments. The GDPF data resource contributes to uncovering the geographical distribution patterns, key influencing factors and macroecological principles of microbial functions at a global level, thereby promoting research in Earth ecology and human health.

SparsePro: An efficient fine-mapping method integrating summary statistics and functional annotations.

Identifying causal variants from genome-wide association studies (GWAS) is challenging due to widespread linkage disequilibrium (LD) and the possible existence of multiple causal variants in the same genomic locus. Functional annotations of the genome may help to prioritize variants that are biologically relevant and thus improve fine-mapping of GWAS results. Classical fine-mapping methods conducting an exhaustive search of variant-level causal configurations have a high computational cost, especially when the underlying genetic architecture and LD patterns are complex. SuSiE provided an iterative Bayesian stepwise selection algorithm for efficient fine-mapping. In this work, we build connections between SuSiE and a paired mean field variational inference algorithm through the implementation of a sparse projection, and propose effective strategies for estimating hyperparameters and summarizing posterior probabilities. Moreover, we incorporate functional annotations into fine-mapping by jointly estimating enrichment weights to derive functionally-informed priors. We evaluate the performance of SparsePro through extensive simulations using resources from the UK Biobank. Compared to state-of-the-art methods, SparsePro achieved improved power for fine-mapping with reduced computation time. We demonstrate the utility of SparsePro through fine-mapping of five functional biomarkers of clinically relevant phenotypes. In summary, we have developed an efficient fine-mapping method for integrating summary statistics and functional annotations. Our method can have wide utility in understanding the genetics of complex traits and increasing the yield of functional follow-up studies of GWAS. SparsePro software is available on GitHub at https://github.com/zhwm/SparsePro.

Age-dependent seasonal growth cessation in Populus.

In temperate and boreal regions, perennial plants adapt their annual growth cycle to the change of seasons. In natural forests, juvenile seedlings usually display longer growth seasons compared to adult trees to ensure their establishment and survival under canopy shade. However, how trees adjust their annual growth according to their age is not known. In this study, we show that age-dependent seasonal growth cessation is genetically controlled and found that the miR156-SPL3/5 module, a key regulon of vegetative phase change (VPC), also triggers age-dependent growth cessation in Populus trees. We show that miR156 promotes shoot elongation during vegetative growth, and its targets SPL3/5s function in the same pathway but as repressors. We find that the miR156-SPL3/5s regulon controls growth cessation in both leaves and shoot apices and through multiple pathways, but with a different mechanism compared to how the miR156-SPL regulon controls VPC in annual plants. Taken together, our results reveal an age-dependent genetic network in mediating seasonal growth cessation, a key phenological process in the climate adaptation of perennial trees.

GLP-1 in the Hypothalamic Paraventricular Nucleus Promotes Sympathetic Activation and Hypertension.

Glucagon-like peptide-1 (GLP-1) and its analogs are widely used for diabetes treatment. The paraventricular nucleus (PVN) is crucial for regulating cardiovascular activity. This study aims to determine the roles of GLP-1 and its receptors (GLP-1R) in the PVN in regulating sympathetic outflow and blood pressure. Experiments were carried out in male normotensive rats and spontaneously hypertensive rats (SHR). Renal sympathetic nerve activity (RSNA) and mean arterial pressure (MAP) were recorded. GLP-1 and GLP-1R expressions were present in the PVN. PVN microinjection of GLP-1R agonist recombinant human GLP-1 (rhGLP-1) or EX-4 increased RSNA and MAP, which were prevented by GLP-1R antagonist exendin 9-39 (EX9-39) or GLP-1R antagonist 1, superoxide scavenger tempol, antioxidant N-acetylcysteine, NADPH oxidase (NOX) inhibitor apocynin, adenylyl cyclase (AC) inhibitor SQ22536 or protein kinase A (PKA) inhibitor H89. PVN microinjection of rhGLP-1 increased superoxide production, NADPH oxidase activity, cAMP level, AC, and PKA activity, which were prevented by SQ22536 or H89. GLP-1 and GLP-1R were upregulated in the PVN of SHR. PVN microinjection of GLP-1 agonist increased RSNA and MAP in both WKY and SHR, but GLP-1 antagonists caused greater effects in reducing RSNA and MAP in SHR than in WKY. The increased superoxide production and NADPH oxidase activity in the PVN of SHR were augmented by GLP-1R agonists but attenuated by GLP-1R antagonists. These results indicate that activation of GLP-1R in the PVN increased sympathetic outflow and blood pressure via cAMP-PKA-mediated NADPH oxidase activation and subsequent superoxide production. GLP-1 and GLP-1R upregulation in the PVN partially contributes to sympathetic overactivity and hypertension.

Spexin Diminishes Atrial Fibrillation Vulnerability by Acting on Galanin Receptor 2.

BACKGROUND: G protein-coupled receptors play a critical role in atrial fibrillation (AF). Spexin is a novel ligand of galanin receptors (GALRs). In this study, we investigated the regulation of spexin and GALRs on AF and the underlying mechanisms. METHODS: Global spexin knockout (SPX-KO) and cardiomyocyte-specific GALRs knockout (GALR-cKO) mice underwent burst pacing electrical stimulation. Optical mapping was used to determine atrial conduction velocity and action potential duration. Atrial myocyte action potential duration and inward rectifying K+ current (IK1) were recorded using whole-cell patch clamps. Isolated cardiomyocytes were stained with Fluo-3/AM dye, and intracellular Ca2+ handling was examined by CCD camera. A mouse model of AF was established by Ang-II (angiotensin II) infusion. RESULTS: Spexin plasma levels in patients with AF were lower than those in subjects without AF, and knockout of spexin increased AF susceptibility in mice. In the atrium of SPX-KO mice, potassium inwardly rectifying channel subfamily J member 2 (KCNJ2) and sarcolipin (SLN) were upregulated; meanwhile, IK1 current was increased and Ca2+ handling was impaired in isolated atrial myocytes of SPX-KO mice. GALR2-cKO mice, but not GALR1-cKO and GALR3-cKO mice, had a higher incidence of AF, which was associated with higher IK1 current and intracellular Ca2+ overload. The phosphorylation level of CREB (cyclic AMP responsive element binding protein 1) was upregulated in atrial tissues of SPX-KO and GALR2-cKO mice. Chromatin immunoprecipitation confirmed the recruitment of p-CREB to the proximal promoter regions of KCNJ2 and SLN. Finally, spexin treatment suppressed CREB signaling, decreased IK1 current and intracellular Ca2+ overload, which thus reduced the inducibility of AF in Ang-II-infused mice. CONCLUSIONS: Spexin reduces atrial fibrillation susceptibility by inhibiting CREB phosphorylation and thus downregulating KCNJ2 and SLN transcription by GALR2 receptor. The spexin/GALR2/CREB signaling pathway represents a novel therapeutic avenue in the development of agents against atrial fibrillation.

Incorporating the image formation process into deep learning improves network performance.

We present Richardson-Lucy network (RLN), a fast and lightweight deep learning method for three-dimensional fluorescence microscopy deconvolution. RLN combines the traditional Richardson-Lucy iteration with a fully convolutional network structure, establishing a connection to the image formation process and thereby improving network performance. Containing only roughly 16,000 parameters, RLN enables four- to 50-fold faster processing than purely data-driven networks with many more parameters. By visual and quantitative analysis, we show that RLN provides better deconvolution, better generalizability and fewer artifacts than other networks, especially along the axial dimension. RLN outperforms classic Richardson-Lucy deconvolution on volumes contaminated with severe out of focus fluorescence or noise and provides four- to sixfold faster reconstructions of large, cleared-tissue datasets than classic multi-view pipelines. We demonstrate RLN's performance on cells, tissues and embryos imaged with widefield-, light-sheet-, confocal- and super-resolution microscopy.

NS7a of SADS-CoV promotes viral infection via inducing apoptosis to suppress type III interferon production.

Swine acute diarrhea syndrome coronavirus (SADS-CoV) is a newly discovered swine coronavirus with potential cross-species transmission risk. Although SADS-CoV-induced host cell apoptosis and innate immunity antagonization has been revealed, underlying signaling pathways remain obscure. Here, we demonstrated that infection of SADS-CoV induced apoptosis in vivo and in vitro, and that viral protein NS7a is mainly responsible for SADS-CoV-induced apoptosis in host cells. Furthermore, we found that NS7a interacted with apoptosis-inducing factor mitochondria associated 1 (AIFM1) to activate caspase-3 via caspase-6 in SADS-CoV-infected cells, and enhanced SADS-CoV replication. Importantly, NS7a suppressed poly(I:C)-induced expression of type III interferon (IFN-λ) via activating caspase-3 to cleave interferon regulatory factor 3 (IRF3), and caspase-3 inhibitor protects piglets against SADS-CoV infection in vivo. These findings reveal how SADS-CoV induced apoptosis to inhibit innate immunity and provide a valuable clue to the development of effective drugs for the clinical control of SADS-CoV infection.IMPORTANCEOver the last 20 years, multiple animal-originated coronaviruses, including severe acute respiratory syndrome coronavirus (SARS-CoV), middle east respiratory syndrome coronavirus (MERS-CoV), and SARS-CoV-2, have caused millions of deaths, seriously jeopardized human health, and hindered social development, indicating that the study of animal-originated coronaviruses with potential for cross-species transmission is particularly important. Bat-originated swine acute diarrhea syndrome coronavirus (SADS-CoV), discovered in 2017, can not only cause fatal diarrhea in piglets, but also infect multiple human cells, with a potential risk of cross-species transmission, but its pathogenesis is unclear. In this study, we demonstrated that NS7a of SADS-CoV suppresses IFN-λ production via apoptosis-inducing factor mitochondria associated 1 (AIFM1)-caspase-6-caspase-3-interferon regulatory factor 3 (IRF3) pathway, and caspase-3 inhibitor (Z-DEVD-FMK) can effectively inhibit SADS-CoV replication and protect infected piglets. Our findings in this study contribute to a better understanding of SADS-CoV-host interactions as a part of the coronaviruses pathogenesis and using apoptosis-inhibitor as a drug as potential therapeutic approaches for prevention and control of SADS-CoV infection.

A feature extraction method based on noise reduction for circRNA-miRNA interaction prediction combining multi-structure features in the association networks.

MOTIVATION: A large number of studies have shown that circular RNA (circRNA) affects biological processes by competitively binding miRNA, providing a new perspective for the diagnosis, and treatment of human diseases. Therefore, exploring the potential circRNA-miRNA interactions (CMIs) is an important and urgent task at present. Although some computational methods have been tried, their performance is limited by the incompleteness of feature extraction in sparse networks and the low computational efficiency of lengthy data. RESULTS: In this paper, we proposed JSNDCMI, which combines the multi-structure feature extraction framework and Denoising Autoencoder (DAE) to meet the challenge of CMI prediction in sparse networks. In detail, JSNDCMI integrates functional similarity and local topological structure similarity in the CMI network through the multi-structure feature extraction framework, then forces the neural network to learn the robust representation of features through DAE and finally uses the Gradient Boosting Decision Tree classifier to predict the potential CMIs. JSNDCMI produces the best performance in the 5-fold cross-validation of all data sets. In the case study, seven of the top 10 CMIs with the highest score were verified in PubMed. AVAILABILITY: The data and source code can be found at https://github.com/1axin/JSNDCMI.