Breaking NGF-TrkA immunosuppression in melanoma sensitizes immunotherapy for durable memory T cell protection.

Melanoma cells, deriving from neuroectodermal melanocytes, may exploit the nervous system's immune privilege for growth. Here we show that nerve growth factor (NGF) has both melanoma cell intrinsic and extrinsic immunosuppressive functions. Autocrine NGF engages tropomyosin receptor kinase A (TrkA) on melanoma cells to desensitize interferon γ signaling, leading to T and natural killer cell exclusion. In effector T cells that upregulate surface TrkA expression upon T cell receptor activation, paracrine NGF dampens T cell receptor signaling and effector function. Inhibiting NGF, either through genetic modification or with the tropomyosin receptor kinase inhibitor larotrectinib, renders melanomas susceptible to immune checkpoint blockade therapy and fosters long-term immunity by activating memory T cells with low affinity. These results identify the NGF-TrkA axis as an important suppressor of anti-tumor immunity and suggest larotrectinib might be repurposed for immune sensitization. Moreover, by enlisting low-affinity T cells, anti-NGF reduces acquired resistance to immune checkpoint blockade and prevents melanoma recurrence.

Transient interactions modulate the affinity of NF-κB transcription factors for DNA.

The dimeric nuclear factor kappa B (NF-κB) transcription factors (TFs) regulate gene expression by binding to a variety of κB DNA elements with conserved G:C-rich flanking sequences enclosing a degenerate central region. Toward defining mechanistic principles of affinity regulated by degeneracy, we observed an unusual dependence of the affinity of RelA on the identity of the central base pair, which appears to be noncontacted in the complex crystal structures. The affinity of κB sites with A or T at the central position is ~10-fold higher than with G or C. The crystal structures of neither the complexes nor the free κB DNAs could explain the differences in affinity. Interestingly, differential dynamics of several residues were revealed in molecular dynamics simulation studies, where simulation replicates totaling 148 µs were performed on NF-κB:DNA complexes and free κB DNAs. Notably, Arg187 and Arg124 exhibited selectivity in transient interactions that orchestrated a complex interplay among several DNA-interacting residues in the central region. Binding and simulation studies with mutants supported these observations of transient interactions dictating specificity. In combination with published reports, this work provides insights into the nuanced mechanisms governing the discriminatory binding of NF-κB family TFs to κB DNA elements and sheds light on cancer pathogenesis of cRel, a close homolog of RelA.

An intersectional genetic approach for simultaneous cell type-specific labelling and gene knockout in the mouse.

Precise genome manipulation in specific cell types and subtypes in vivo is crucial for neurobiological research because of the cellular heterogeneity of the brain. Site-specific recombinase systems in the mouse, such as Cre-loxP, improve cell type-specific genome manipulation; however, undesirable expression of cell type-specific Cre can occur. This could be due to transient expression during early development, natural expression in more than one cell type, kinetics of recombinases, sensitivity of the Cre reporter, and disruption in cis-regulatory elements by transgene insertion. Moreover, cell subtypes cannot be distinguished in cell type-specific Cre mice. To address these issues, we applied an intersectional genetic approach in mouse using triple recombination systems (Cre-loxP, Flp-FRT and Dre-rox). As a proof of principle, we labelled heterogeneous cell subtypes and deleted target genes within given cell subtypes by labelling neuropeptide Y (NPY)-, calretinin (calbindin 2) (CR)- and cholecystokinin (CCK)-expressing GABAergic neurons in the brain followed by deletion of RNA-binding Fox-1 homolog 3 (Rbfox3) in our engineered mice. Together, our study applies an intersectional genetic approach in vivo to generate engineered mice serving dual purposes of simultaneous cell subtype-specific labelling and gene knockout.

scCorrector: a robust method for integrating multi-study single-cell data.

The advent of single-cell sequencing technologies has revolutionized cell biology studies. However, integrative analyses of diverse single-cell data face serious challenges, including technological noise, sample heterogeneity, and different modalities and species. To address these problems, we propose scCorrector, a variational autoencoder-based model that can integrate single-cell data from different studies and map them into a common space. Specifically, we designed a Study Specific Adaptive Normalization for each study in decoder to implement these features. scCorrector substantially achieves competitive and robust performance compared with state-of-the-art methods and brings novel insights under various circumstances (e.g. various batches, multi-omics, cross-species, and development stages). In addition, the integration of single-cell data and spatial data makes it possible to transfer information between different studies, which greatly expand the narrow range of genes covered by MERFISH technology. In summary, scCorrector can efficiently integrate multi-study single-cell datasets, thereby providing broad opportunities to tackle challenges emerging from noisy resources.

iCRBP-LKHA: Large convolutional kernel and hybrid channel-spatial attention for identifying circRNA-RBP interaction sites.

Circular RNAs (circRNAs) play vital roles in transcription and translation. Identification of circRNA-RBP (RNA-binding protein) interaction sites has become a fundamental step in molecular and cell biology. Deep learning (DL)-based methods have been proposed to predict circRNA-RBP interaction sites and achieved impressive identification performance. However, those methods cannot effectively capture long-distance dependencies, and cannot effectively utilize the interaction information of multiple features. To overcome those limitations, we propose a DL-based model iCRBP-LKHA using deep hybrid networks for identifying circRNA-RBP interaction sites. iCRBP-LKHA adopts five encoding schemes. Meanwhile, the neural network architecture, which consists of large kernel convolutional neural network (LKCNN), convolutional block attention module with one-dimensional convolution (CBAM-1D) and bidirectional gating recurrent unit (BiGRU), can explore local information, global context information and multiple features interaction information automatically. To verify the effectiveness of iCRBP-LKHA, we compared its performance with shallow learning algorithms on 37 circRNAs datasets and 37 circRNAs stringent datasets. And we compared its performance with state-of-the-art DL-based methods on 37 circRNAs datasets, 37 circRNAs stringent datasets and 31 linear RNAs datasets. The experimental results not only show that iCRBP-LKHA outperforms other competing methods, but also demonstrate the potential of this model in identifying other RNA-RBP interaction sites.

Neuronal splicing regulator RBFOX3 mediates seizures via regulating Vamp1 expression preferentially in NPY-expressing GABAergic neurons.

Epilepsy is a common neurological disorder, which has been linked to mutations or deletions of RNA binding protein, fox-1 homolog (Caenorhabditis elegans) 3 (RBFOX3)/NeuN, a neuronal splicing regulator. However, the mechanism of seizure mediation by RBFOX3 remains unknown. Here, we show that mice with deletion of Rbfox3 in gamma-aminobutyric acid (GABA) ergic neurons exhibit spontaneous seizures and high premature mortality due to increased presynaptic release, postsynaptic potential, neuronal excitability, and synaptic transmission in hippocampal dentate gyrus granule cells (DGGCs). Attenuating early excitatory gamma-aminobutyric acid (GABA) action by administering bumetanide, an inhibitor of early GABA depolarization, rescued premature mortality. Rbfox3 deletion reduced hippocampal expression of vesicle-associated membrane protein 1 (VAMP1), a GABAergic neuron-specific presynaptic protein. Postnatal restoration of VAMP1 rescued premature mortality and neuronal excitability in DGGCs. Furthermore, Rbfox3 deletion in GABAergic neurons showed fewer neuropeptide Y (NPY)-expressing GABAergic neurons. In addition, deletion of Rbfox3 in NPY-expressing GABAergic neurons lowered intrinsic excitability and increased seizure susceptibility. Our results establish RBFOX3 as a critical regulator and possible treatment path for epilepsy.

Graph representation learning in bioinformatics: trends, methods and applications.

Graph is a natural data structure for describing complex systems, which contains a set of objects and relationships. Ubiquitous real-life biomedical problems can be modeled as graph analytics tasks. Machine learning, especially deep learning, succeeds in vast bioinformatics scenarios with data represented in Euclidean domain. However, rich relational information between biological elements is retained in the non-Euclidean biomedical graphs, which is not learning friendly to classic machine learning methods. Graph representation learning aims to embed graph into a low-dimensional space while preserving graph topology and node properties. It bridges biomedical graphs and modern machine learning methods and has recently raised widespread interest in both machine learning and bioinformatics communities. In this work, we summarize the advances of graph representation learning and its representative applications in bioinformatics. To provide a comprehensive and structured analysis and perspective, we first categorize and analyze both graph embedding methods (homogeneous graph embedding, heterogeneous graph embedding, attribute graph embedding) and graph neural networks. Furthermore, we summarize their representative applications from molecular level to genomics, pharmaceutical and healthcare systems level. Moreover, we provide open resource platforms and libraries for implementing these graph representation learning methods and discuss the challenges and opportunities of graph representation learning in bioinformatics. This work provides a comprehensive survey of emerging graph representation learning algorithms and their applications in bioinformatics. It is anticipated that it could bring valuable insights for researchers to contribute their knowledge to graph representation learning and future-oriented bioinformatics studies.

Computational prediction and characterization of cell-type-specific and shared binding sites.

MOTIVATION: Cell-type-specific gene expression is maintained in large part by transcription factors (TFs) selectively binding to distinct sets of sites in different cell types. Recent research works have provided evidence that such cell-type-specific binding is determined by TF's intrinsic sequence preferences, cooperative interactions with co-factors, cell-type-specific chromatin landscapes and 3D chromatin interactions. However, computational prediction and characterization of cell-type-specific and shared binding sites is rarely studied. RESULTS: In this article, we propose two computational approaches for predicting and characterizing cell-type-specific and shared binding sites by integrating multiple types of features, in which one is based on XGBoost and another is based on convolutional neural network (CNN). To validate the performance of our proposed approaches, ChIP-seq datasets of 10 binding factors were collected from the GM12878 (lymphoblastoid) and K562 (erythroleukemic) human hematopoietic cell lines, each of which was further categorized into cell-type-specific (GM12878- and K562-specific) and shared binding sites. Then, multiple types of features for these binding sites were integrated to train the XGBoost- and CNN-based models. Experimental results show that our proposed approaches significantly outperform other competing methods on three classification tasks. Moreover, we identified independent feature contributions for cell-type-specific and shared sites through SHAP values and explored the ability of the CNN-based model to predict cell-type-specific and shared binding sites by excluding or including DNase signals. Furthermore, we investigated the generalization ability of our proposed approaches to different binding factors in the same cellular environment. AVAILABILITY AND IMPLEMENTATION: The source code is available at: https://github.com/turningpoint1988/CSSBS. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Interstitial fibrosis increases the risk of end-stage kidney disease in patients with lupus nephritis.

OBJECTIVE: To evaluate the risk of end-stage kidney disease (ESKD) in LN patients using tubulointerstitial lesion scores. METHODS: Clinical profiles and histopathological presentations of 151 biopsy-proven LN patients were retrospectively examined. Risk factors of ESKD based on characteristics and scoring of their tubulointerstitial lesions [e.g. interstitial inflammation (II), tubular atrophy (TA) and interstitial fibrosis (IF)] were analysed. RESULTS: The mean age of 151 LN patients was 36 years old, and 136 (90.1%) were female. The LN cases examined included: class I/II (n = 3, 2%), class III/IV (n = 119, 78.8%), class V (n = 23, 15.2%) and class VI (n = 6, 4.0%). The mean serum creatinine level was 1.4 mg/dl. Tubulointerstitial lesions were recorded in 120 (79.5%) patients. Prior to receiving renal biopsy, nine (6.0%) patients developed ESKD. During the follow-up period (mean, 58 months), an additional 47 patients (31.1%) progressed to ESKD. Multivariate analyses identified serum creatinine [hazard ratio (HR): 1.7, 95% CI: 1.42-2.03, P < 0.001] and IF (HR: 3.2, 95% CI: 1.58-6.49, P = 0.001) as independent risk factors of ESKD. Kaplan-Meier analysis further confirmed a heightened risk of ESKD associated with IF. CONCLUSION: Tubulointerstitial involvement is commonly observed in the histopathological presentation of LN. However, IF, rather than II or TA, was found to increase the risk of ESKD in our cohort. Therefore, to predict renal outcome in LN patients prior to adjusting immunosuppressive treatment, the degree of IF should be reviewed.

Risk Factors and Incidence of Serious Infections in Patients With Systemic Lupus Erythematosus Undergoing Rituximab Therapy.

OBJECTIVE: To evaluate the risk and protective factors of serious infection (SI) in patients with systemic lupus erythematosus (SLE) within 180 days of rituximab (RTX) treatment. METHODS: Patients with SLE treated with RTX were analyzed. SI was defined as any infectious disease requiring hospitalization. The clinical characteristics, laboratory profiles, medications, and incidence rate (IR) are presented. Multivariate Cox proportional hazards models and Kaplan-Meier analysis for risk factors of SI were performed. RESULTS: A total of 174 patients with SLE receiving RTX treatment were enrolled. The overall IR of SIs was 51.0/100 patient-years (PYs). Pneumonia (30.4/100 PYs), followed by soft tissue infections, intra-abdominal infections, and Pneumocystis jiroveci pneumonia (all 6.1/100 PYs) were the leading types of SIs. Twelve patients died during the 180-day follow-up (crude mortality rate: 14.6/100 PYs). Chronic kidney disease (CKD), defined as an estimated glomerular filtration rate < 60 mL/min/1.73 m2 (hazard ratio [HR] 2.88, 95% CI 1.30-6.38), and a background prednisolone (PSL) equivalent dosage ≥ 15 mg/day (HR 3.50, 95% CI 1.57-7.78) were risk factors for SIs among all patients with SLE. Kaplan-Meier analysis confirmed the risk of SI for patients with SLE with CKD and a background PSL equivalent dosage ≥ 15 mg/day (log-rank P = 0.001 and 0.02, respectively). Hydroxychloroquine (HCQ) reduced the risk of SIs in patients with SLE (HR 0.35, 95% CI 0.15-0.82; log-rank P = 0.003). CONCLUSION: SI was prevalent in patients with SLE after RTX treatment. Patients with SLE with CKD and high-dose glucocorticoid use required constant vigilance. HCQ may reduce the risk of SI among patients with SLE administered RTX.

iCircDA-NEAE: Accelerated attribute network embedding and dynamic convolutional autoencoder for circRNA-disease associations prediction.

Accumulating evidence suggests that circRNAs play crucial roles in human diseases. CircRNA-disease association prediction is extremely helpful in understanding pathogenesis, diagnosis, and prevention, as well as identifying relevant biomarkers. During the past few years, a large number of deep learning (DL) based methods have been proposed for predicting circRNA-disease association and achieved impressive prediction performance. However, there are two main drawbacks to these methods. The first is these methods underutilize biometric information in the data. Second, the features extracted by these methods are not outstanding to represent association characteristics between circRNAs and diseases. In this study, we developed a novel deep learning model, named iCircDA-NEAE, to predict circRNA-disease associations. In particular, we use disease semantic similarity, Gaussian interaction profile kernel, circRNA expression profile similarity, and Jaccard similarity simultaneously for the first time, and extract hidden features based on accelerated attribute network embedding (AANE) and dynamic convolutional autoencoder (DCAE). Experimental results on the circR2Disease dataset show that iCircDA-NEAE outperforms other competing methods significantly. Besides, 16 of the top 20 circRNA-disease pairs with the highest prediction scores were validated by relevant literature. Furthermore, we observe that iCircDA-NEAE can effectively predict new potential circRNA-disease associations.

Parkin coordinates mitochondrial lipid remodeling to execute mitophagy.

Autophagy has emerged as the prime machinery for implementing organelle quality control. In the context of mitophagy, the ubiquitin E3 ligase Parkin tags impaired mitochondria with ubiquitin to activate autophagic degradation. Although ubiquitination is essential for mitophagy, it is unclear how ubiquitinated mitochondria activate autophagosome assembly locally to ensure efficient destruction. Here, we report that Parkin activates lipid remodeling on mitochondria targeted for autophagic destruction. Mitochondrial Parkin induces the production of phosphatidic acid (PA) and its subsequent conversion to diacylglycerol (DAG) by recruiting phospholipase D2 and activating the PA phosphatase, Lipin-1. The production of DAG requires mitochondrial ubiquitination and ubiquitin-binding autophagy receptors, NDP52 and optineurin (OPTN). Autophagic receptors, via Golgi-derived vesicles, deliver an autophagic activator, EndoB1, to ubiquitinated mitochondria. Inhibition of Lipin-1, NDP52/OPTN, or EndoB1 results in a failure to produce mitochondrial DAG, autophagosomes, and mitochondrial clearance, while exogenous cell-permeable DAG can induce autophagosome production. Thus, mitochondrial DAG production acts downstream of Parkin to enable the local assembly of autophagosomes for the efficient disposal of ubiquitinated mitochondria.

Individualized detection of TMPRSS2-ERG fusion status in prostate cancer: a rank-based qualitative transcriptome signature.

BACKGROUND: TMPRSS2-ERG (T2E) fusion is highly related to aggressive clinical features in prostate cancer (PC), which guides individual therapy. However, current fusion prediction tools lacked enough accuracy and biomarkers were unable to be applied to individuals across different platforms due to their quantitative nature. This study aims to identify a transcriptome signature to detect the T2E fusion status of PC at the individual level. METHODS: Based on 272 high-throughput mRNA expression profiles from the Sboner dataset, we developed a rank-based algorithm to identify a qualitative signature to detect T2E fusion in PC. The signature was validated in 1223 samples from three external datasets (Setlur, Clarissa, and TCGA). RESULTS: A signature, composed of five mRNAs coupled to ERG (five ERG-mRNA pairs, 5-ERG-mRPs), was developed to distinguish T2E fusion status in PC. 5-ERG-mRPs reached 84.56% accuracy in Sboner dataset, which was verified in Setlur dataset (n = 455, accuracy = 82.20%) and Clarissa dataset (n = 118, accuracy = 81.36%). Besides, for 495 samples from TCGA, two subtypes classified by 5-ERG-mRPs showed a higher level of significance in various T2E fusion features than subtypes obtained through current fusion prediction tools, such as STAR-Fusion. CONCLUSIONS: Overall, 5-ERG-mRPs can robustly detect T2E fusion in PC at the individual level, which can be used on any gene measurement platform without specific normalization procedures. Hence, 5-ERG-mRPs may serve as an auxiliary tool for PC patient management.

[Fatal Familial Insomnia With Significant Correlations Between Involuntary Movements and Postural Changes:Report of One Case].

Fatal familial insomnia,an autosomal dominant prion disease,is rare.We reported the clinical symptoms,examination results,diagnosis,treatment,and prognosis of a patient who was diagnosed with fatal familial insomnia.Furthermore,we described the unique clinical manifestations that involuntary movements and laryngeal stridor were significantly correlated with postural changes,aiming to provide reference for the clinical diagnosis,treatment,and research of the disease in the future.

[Tumor-Like Primary Central Nervous System Vasculitis With Spina Involvement:Report of One Case].

Primary central nervous system vasculitis (PACNS) is a vasculitic disorder affecting small to medium-sized blood vessels primarily in the central nervous system,involving the brain,spinal cord,and meninges.Tumor-like PNCAS,a rare subtype of PACNS,is often misdiagnosed as intracranial malignancy,and that with spinal cord involvement is even more uncommon.The lack of specific clinical symptoms and imaging manifestations poses a challenge to the diagnosis of PACNS.This report presents a case of tumor-like PACNS with spinal cord involvement based on the pathological evidence,aiming to enrich the knowledge about this condition.

[Clinical Features of 17 Patients With Primary Angiitis of the Central Nervous System Confirmed by Brain Biopsy].

Objective To analyze the clinical features of 17 patients with primary angiitis of the central nervous system (PACNS) and thus facilitate the early diagnosis and treatment,reduce the recurrence and mortality,and improve the prognoses of this disease. Methods We collected the data of patients with PACNS diagnosed by brain biopsy from January 2009 to June 2023 and analyzed their clinical presentations,laboratory and imaging manifestations,electrophysiological and pathological changes,and treatment regimens and prognosis. Results The 17 patients diagnosed with PACNS via brain biopsy included one child and 16 adults.The subtyping results showed that 10,2,3,2,1,and 1 patients had tumorous,spinal cord-involved,angiography-positive,rapidly progressive,hemorrhagic,and amyloid ß-related PACNS,respectively.Eleven (64.7%) of the patients were complicated with secondary epilepsy.All the patients exhibited abnormal manifestations in head MRI,with 94.1% showing lesions with uneven enhancement around the lesions or in the leptomeninges. Magnetic resonance angiography revealed large vessel abnormalities in 3 patients,and spinal cord involvement was observed in 2 patients.Histopathological typing revealed 7 (43.7%) patients with lymphocytic vasculitis and 5 (31.2%) patients with necrotizing vasculitis.Eleven patients were treated with glucocorticoids and cyclophosphamide,which resulted in partial lesion disappearance and symptom amelioration in 6 patients upon reevaluation with head MRI after 3 months of maintenance therapy.Two,1,and 3 patients experienced rapid disease progression,death,and recurrence within 1 year,respectively.Three patients showed insensitivity to hormonotherapy and residual disabilities.Two patients received rituximab after relapse and remained clinically stable during a follow-up period of 0.5-1 year. Conclusion Tumorous PACNS was more prone to epilepsy,mainly occurring in males.The most common histopathological type was necrotizing vasculitis,which responded to hormonotherapy and had favorable outcomes.Therefore,for the young patients with epilepsy and intracranial tumorous lesions,the possibility of PACNS should be considered.Spinal cord involvement in PACNS was often located in the thoracic and cervical cords,suggesting a poorer prognosis.Electromyography commonly revealed neural conduction abnormalities in the anterior horn or roots,providing clues for differential diagnosis.For suspected spinal cord involvement,comprehensive electromyography is recommended.Rapidly progressive PACNS often presented infratentorial lesions,such as lesions in the pons and medulla,with a higher mortality rate.Hemorrhagic PACNS was rare,and a multifocal hemorrhagic lesion with enhancement in the intracranial region,particularly in young patients,should raise suspicion.For the patients with recurrent or progressive disease,rituximab is a recommended therapeutic option.

A survey on deep learning in DNA/RNA motif mining.

DNA/RNA motif mining is the foundation of gene function research. The DNA/RNA motif mining plays an extremely important role in identifying the DNA- or RNA-protein binding site, which helps to understand the mechanism of gene regulation and management. For the past few decades, researchers have been working on designing new efficient and accurate algorithms for mining motif. These algorithms can be roughly divided into two categories: the enumeration approach and the probabilistic method. In recent years, machine learning methods had made great progress, especially the algorithm represented by deep learning had achieved good performance. Existing deep learning methods in motif mining can be roughly divided into three types of models: convolutional neural network (CNN) based models, recurrent neural network (RNN) based models, and hybrid CNN-RNN based models. We introduce the application of deep learning in the field of motif mining in terms of data preprocessing, features of existing deep learning architectures and comparing the differences between the basic deep learning models. Through the analysis and comparison of existing deep learning methods, we found that the more complex models tend to perform better than simple ones when data are sufficient, and the current methods are relatively simple compared with other fields such as computer vision, language processing (NLP), computer games, etc. Therefore, it is necessary to conduct a summary in motif mining by deep learning, which can help researchers understand this field.

Locating transcription factor binding sites by fully convolutional neural network.

Transcription factors (TFs) play an important role in regulating gene expression, thus identification of the regions bound by them has become a fundamental step for molecular and cellular biology. In recent years, an increasing number of deep learning (DL) based methods have been proposed for predicting TF binding sites (TFBSs) and achieved impressive prediction performance. However, these methods mainly focus on predicting the sequence specificity of TF-DNA binding, which is equivalent to a sequence-level binary classification task, and fail to identify motifs and TFBSs accurately. In this paper, we developed a fully convolutional network coupled with global average pooling (FCNA), which by contrast is equivalent to a nucleotide-level binary classification task, to roughly locate TFBSs and accurately identify motifs. Experimental results on human ChIP-seq datasets show that FCNA outperforms other competing methods significantly. Besides, we find that the regions located by FCNA can be used by motif discovery tools to further refine the prediction performance. Furthermore, we observe that FCNA can accurately identify TF-DNA binding motifs across different cell lines and infer indirect TF-DNA bindings.

MeSHHeading2vec: a new method for representing MeSH headings as vectors based on graph embedding algorithm.

Effectively representing Medical Subject Headings (MeSH) headings (terms) such as disease and drug as discriminative vectors could greatly improve the performance of downstream computational prediction models. However, these terms are often abstract and difficult to quantify. In this paper, we converted the MeSH tree structure into a relationship network and applied several graph embedding algorithms on it to represent these terms. Specifically, the relationship network consisting of nodes (MeSH headings) and edges (relationships), which can be constructed by the tree num. Then, five graph embedding algorithms including DeepWalk, LINE, SDNE, LAP and HOPE were implemented on the relationship network to represent MeSH headings as vectors. In order to evaluate the performance of the proposed methods, we carried out the node classification and relationship prediction tasks. The results show that the MeSH headings characterized by graph embedding algorithms can not only be treated as an independent carrier for representation, but also can be utilized as additional information to enhance the representation ability of vectors. Thus, it can serve as an input and continue to play a significant role in any computational models related to disease, drug, microbe, etc. Besides, our method holds great hope to inspire relevant researchers to study the representation of terms in this network perspective.

Base-resolution prediction of transcription factor binding signals by a deep learning framework.

Transcription factors (TFs) play an important role in regulating gene expression, thus the identification of the sites bound by them has become a fundamental step for molecular and cellular biology. In this paper, we developed a deep learning framework leveraging existing fully convolutional neural networks (FCN) to predict TF-DNA binding signals at the base-resolution level (named as FCNsignal). The proposed FCNsignal can simultaneously achieve the following tasks: (i) modeling the base-resolution signals of binding regions; (ii) discriminating binding or non-binding regions; (iii) locating TF-DNA binding regions; (iv) predicting binding motifs. Besides, FCNsignal can also be used to predict opening regions across the whole genome. The experimental results on 53 TF ChIP-seq datasets and 6 chromatin accessibility ATAC-seq datasets show that our proposed framework outperforms some existing state-of-the-art methods. In addition, we explored to use the trained FCNsignal to locate all potential TF-DNA binding regions on a whole chromosome and predict DNA sequences of arbitrary length, and the results show that our framework can find most of the known binding regions and accept sequences of arbitrary length. Furthermore, we demonstrated the potential ability of our framework in discovering causal disease-associated single-nucleotide polymorphisms (SNPs) through a series of experiments.