AnimalMetaOmics: a multi-omics data resources for exploring animal microbial genomes and microbiomes.

The Animal Meta-omics landscape database (AnimalMetaOmics, https://yanglab.hzau.edu.cn/animalmetaomics#/) is a comprehensive and freely available resource that includes metagenomic, metatranscriptomic, and metaproteomic data from various non-human animal species and provides abundant information on animal microbiomes, including cluster analysis of microbial cognate genes, functional gene annotations, active microbiota composition, gene expression abundance, and microbial protein identification. In this work, 55 898 microbial genomes were annotated from 581 animal species, including 42 924 bacterial genomes, 12 336 virus genomes, 496 archaea genomes and 142 fungi genomes. Moreover, 321 metatranscriptomic datasets were analyzed from 31 animal species and 326 metaproteomic datasets from four animal species, as well as the pan-genomic dynamics and compositional characteristics of 679 bacterial species and 13 archaea species from animal hosts. Researchers can efficiently access and acquire the information of cross-host microbiota through a user-friendly interface, such as species, genomes, activity levels, expressed protein sequences and functions, and pan-genome composition. These valuable resources provide an important reference for better exploring the classification, functional diversity, biological process diversity and functional genes of animal microbiota.

Enhanced neural synchronization during social communications between dyads with high autistic traits.

Autism is characterized by atypical social communication styles. To investigate whether individuals with high autistic traits could still have effective social communication among each other, we compared the behavioral patterns and communication quality within 64 dyads of college students paired with both high, both low, and mixed high-low (HL) autistic traits, with their gender matched. Results revealed that the high-high (HH) autistic dyads exhibited atypical behavioral patterns during conversations, including reduced mutual gaze, communicational turns, and emotional sharing compared with the low-low and/or HL autistic dyads. However, the HH autistic dyads displayed enhanced interpersonal neural synchronization during social communications measured by functional near-infrared spectroscopy, suggesting an effective communication style. Besides, they also provided more positive subjective evaluations of the conversations. These findings highlight the potential for alternative pathways to effectively communicate with the autistic community, contribute to a deeper understanding of how high autistic traits influence social communication dynamics among autistic individuals, and provide important insights for the clinical practices for supporting autistic people.

Switching of brain networks across different cerebral perfusion states: insights from EEG dynamic microstate analyses.

The alteration of neural interactions across different cerebral perfusion states remains unclear. This study aimed to fulfill this gap by examining the longitudinal brain dynamic information interactions before and after cerebral reperfusion. Electroencephalogram in eyes-closed state at baseline and postoperative 7-d and 3-month follow-ups (moyamoya disease: 20, health controls: 23) were recorded. Dynamic network analyses were focused on the features and networks of electroencephalogram microstates across different microstates and perfusion states. Considering the microstate features, the parameters were disturbed of microstate B, C, and D but preserved of microstate A. The transition probabilities of microstates A-B and B-D were increased to play a complementary role across different perfusion states. Moreover, the microstate variability was decreased, but was significantly improved after cerebral reperfusion. Regarding microstate networks, the functional connectivity strengths were declined, mainly within frontal, parietal, and occipital lobes and between parietal and occipital lobes in different perfusion states, but were ameliorated after cerebral reperfusion. This study elucidates how dynamic interaction patterns of brain neurons change after cerebral reperfusion, which allows for the observation of brain network transitions across various perfusion states in a live clinical setting through direct intervention.

CRAMdb: a comprehensive database for composition and roles of microbiome in animals.

CRAMdb (a database for composition and roles of animal microbiome) is a comprehensive resource of curated and consistently annotated metagenomes for non-human animals. It focuses on the composition and roles of the microbiome in various animal species. The main goal of the CRAMdb is to facilitate the reuse of animal metagenomic data, and enable cross-host and cross-phenotype comparisons. To this end, we consistently annotated microbiomes (including 16S, 18S, ITS and metagenomics sequencing data) of 516 animals from 475 projects spanning 43 phenotype pairs to construct the database that is equipped with 9430 bacteria, 278 archaea, 2216 fungi and 458 viruses. CRAMdb provides two main contents: microbiome composition data, illustrating the landscape of the microbiota (bacteria, archaea, fungi, and viruses) in various animal species, and microbiome association data, revealing the relationships between the microbiota and various phenotypes across different animal species. More importantly, users can quickly compare the composition of the microbiota of interest cross-host or body site and the associated taxa that differ between phenotype pairs cross-host or cross-phenotype. CRAMdb is freely available at (http://www.ehbio.com/CRAMdb).

Hypomethylating agent decitabine sensitizes diffuse large B-cell lymphoma to venetoclax.

Despite recent advances in the therapy of diffuse large B-cell lymphoma (DLBCL), many patients are still not cured. Therefore, new therapeutic strategies are needed. The anti-apoptotic B-cell lymphoma 2 (BCL2) gene is commonly dysregulated in DLBCL due to various mechanisms such as chromosomal translocation t(14;18)(q32;q21) and copy number alterations; however, targeting BCL-2 with the selective inhibitor, venetoclax, led to response in only a minority of patients. Thus, we sought to identify a rational combination partner of venetoclax to improve its activity against DLBCL cells. Utilizing a functional assay, dynamic BH3 profiling, we found that the DNA hypomethylating agent decitabine increased mitochondrial apoptotic priming and BCL-2 dependence in DLBCL cells. RNA-sequencing analysis revealed that decitabine suppressed the pro-survival PI3K-AKT pathway and altered the mitochondria membrane composition in DLBCL cell lines. Additionally, it induced a DNA damage response and increased BAX and BAK activities. The combination of decitabine and venetoclax synergistically suppressed proliferation of DLBCL cells both in vitro and in vivo in a DLBCL cell line-derived xenograft mouse model. Our study suggests that decitabine plus venetoclax is a promising combination to explore clinically in DLBCL.

Prognostic significance of platelet­to­albumin ratio in patients with nasopharyngeal carcinoma receiving concurrent chemoradiotherapy: a retrospective study of 858 cases.

BACKGROUND: Despite evidence supporting the high correlation of the novel platelet-to-albumin ratio (PAR) with survival in diverse malignancies, its prognostic relevance in nasopharyngeal carcinoma (NPC) remains underexplored. This study aimed to examine the link between PAR and overall survival (OS) in NPC and to establish a predictive model based on this biomarker. METHODS: We retrospectively assembled a cohort consisting of 858 NPC patients who underwent concurrent chemoradiotherapy (CCRT). Utilizing the maximally selected log-rank method, we ascertained the optimal cut-off point for the PAR. Subsequently, univariate and multivariate Cox proportional hazards models were employed to discern factors significantly associated with OS and to construct a predictive nomogram. Further, we subjected the nomogram's predictive accuracy to rigorous independent validation. RESULTS: The discriminative optimal PAR threshold was determined to be 4.47, effectively stratifying NPC patients into two prognostically distinct subgroups (hazard ratio [HR] = 0.53; 95% confidence interval [CI]: 0.28-0.98, P = 0.042). A predictive nomogram was formulated using the results from multivariate analysis, which revealed age greater than 45 years, T stage, N stage, and PAR score as independent predictors of OS. The nomogram demonstrated a commendable predictive capability for OS, with a C-index of 0.69 (95% CI: 0.64-0.75), surpassing the performance of the conventional staging system, which had a C-index of 0.56 (95% CI: 0.65-0.74). CONCLUSIONS: In the context of NPC patients undergoing CCRT, the novel nutritional-inflammatory biomarker PAR emerges as a promising, cost-efficient, easily accessible, non-invasive, and potentially valuable predictor of prognosis. The predictive efficacy of the nomogram incorporating the PAR score exceeded that of the conventional staging approach, thereby indicating its potential as an enhanced prognostic tool in this clinical setting.

Taurine ameliorates sensorimotor function by inhibiting apoptosis and activating A2 astrocytes in mice after subarachnoid hemorrhage.

Subarachnoid hemorrhage (SAH) is a form of severe acute stroke with very high mortality and disability rates. Early brain injury (EBI) and delayed cerebral ischemia (DCI) contribute to the poor prognosis of patients with SAH. Currently, some researchers have started to focus on changes in amino acid metabolism that occur in brain tissues after SAH. Taurine is a sulfur-containing amino acid that is semi-essential in animals, and it plays important roles in various processes, such as neurodevelopment, osmotic pressure regulation, and membrane stabilization. In acute stroke, such as cerebral hemorrhage, taurine plays a neuroprotective role. However, the role of taurine after subarachnoid hemorrhage has rarely been reported. In the present study, we established a mouse model of SAH. We found that taurine administration effectively improved the sensorimotor function of these mice. In addition, taurine treatment alleviated sensorimotor neuron damage and reduced the proportion of apoptotic cells. Furthermore, taurine treatment enhanced the polarization of astrocytes toward the neuroprotective phenotype while inhibiting their polarization toward the neurotoxic phenotype. This study is the first to reveal the relationship between taurine and astrocyte polarization and may provide a new strategy for SAH research and clinical treatment.

Efficacy of H2O2 inactivated bovine virus diarrhoea virus (BVDV) type 1 vaccine in mice.

BACKGROUND: Bovine viral diarrhea (BVD) is an acute febrile infectious disease caused by the bovine viral diarrhea virus (BVDV), which has brought huge economic losses to the world's cattle industry. At present, commercial inactivated BVDV vaccines may cause some adverse reactions during use. This study aims to develop a safer and more efficient inactivated BVDV vaccine. METHODS: Here, we described the generation and preclinical efficacy of a hydrogen peroxide (H2O2) inactivated BVDV type 1 vaccine in mice, and administered it separately with commercial vaccine (formaldehyde inactivated) in mice to study its efficacy. RESULTS: The BVDV type 1 IgG, IFN- γ, IL-4 and neutralizing antibody in the serum of the H2O2 inactivated vaccine group can be maintained in mice for 70 days. The IgG level reached its maximum value of 0.67 on the 42nd day, significantly higher than the commercial formaldehyde inactivated BVDV type 1 vaccine. IFN- γ and IL-4 reached their maximum values on the 28th day after immunization, at 123.16 pg/ml and 143.80 pg/ml, respectively, slightly higher than commercial vaccines, but the effect was not significant. At the same time the BVDV-1 neutralizing antibody titer reached a maximum of 12 Nu on the 42nd day post vaccination. CONCLUSIONS: The H2O2 inactivated BVDV vaccine has good safety and immunogenicity, which provides a potential solution for the further development of an efficient and safe BVDV vaccine.

Distributed lag effects and vulnerable groups of PM and active pulmonary TB in Qingdao, China.

There has been a gap regarding current knowledge of the effect of PM on pulmonary TB, such as the exposure-time-response between them. This study aimed to explore the distributed lag effects of particulate matter (PM) on active pulmonary tuberculosis (TB) and identify the vulnerable groups. A generalized additive mixed model combined with a distributed lag non-linear model was applied to quantify the association between PM and active pulmonary TB with adjustment for potential confounders. Relative risk (RR) and cumulative RR with 95% confidence interval (CI) were calculated to quantify the exposure-time-response. A total of 16,486 cases of active pulmonary TB were notified. Results suggested that a unit 10 µg/m3 increase of daily PM2.5 concentration was positively associated with active pulmonary TB morbidity at 36-115 lag day and RR reached maximum at 66 lag day (1.0076; 95%CI, 1.0031-1.0122), and the cumulative RR was 2.1940 (95%CI, 1.2292-3.9161). For PM10, this association was significantly positive at 73-117 lag day, and RR reached maximum at 100 lag day (1.0036; 95%CI, 1.0003-1.0067), and the cumulative RR was not significant. This study provides evidence that PM significantly associate with active pulmonary TB. Vulnerability to PM2.5 was identified in male, female, 0-18 ages, 19-64 ages, workers, and students. Our findings have significant implications for developing local strategies to prevent and reduce health impact in PM polluted areas.

Cross-Domain Knowledge Transfer for Sustainable Heterogeneous Industrial Internet-of-Things Networks.

In this article, a novel cross-domain knowledge transfer method is implemented to optimize the tradeoff between energy consumption and information freshness for all pieces of equipment powered by heterogeneous energy sources within smart factory. Three distinct groups of use cases are considered, each utilizing a different energy source: grid power, green energy source, and mixed energy sources. Differing from mainstream algorithms that require consistency among groups, the proposed method enables knowledge transfer even across varying state and/or action spaces. With the advantage of multiple layers of knowledge extraction, a lightweight knowledge transfer is achieved without the need for neural networks. This facilitates broader applications in self-sustainable wireless networks. Simulation results reveal a notable improvement in the 'warm start' policy for each equipment, manifesting as a 51.32% increase in initial reward compared to a random policy approach.

Nanosensor-Driven Detection of Neuron-Derived Exosomal Aß42 with Graphene Electrolyte-Gated Transistor for Alzheimer's Disease Diagnosis.

Blood-based tests have sparked tremendous attention in non-invasive early diagnosis of Alzheimer's disease (AD), a most prevalent neurodegenerative malady worldwide. Despite significant progress in the methodologies for detecting AD core biomarkers such as Aß42 from serum/plasma, there remains cautious optimism going forward due to its controversial diagnostic value and disease relevance. Here, a graphene electrolyte-gated transistor biosensor is reported for the detection of serum neuron-derived exosomal Aß42 (NDE-Aß42), which is an emerging, compelling trove of blood biomarker for AD. Assisted by the antifouling strategy with the dual-blocking process, the noise against complex biological background was considerably reduced, forging an impressive sensitivity gain with a limit of detection of 447 ag/mL. An accurate detection of SH-SY5Y-derived exosomal Aß42 was also achieved with highly conformable enzyme-linked immunosorbent assay results. Importantly, the clinical analysis for 27 subjects revealed the immense diagnostic value of NDE-Aß42, which can outclass that of serum Aß42. The developed electronic assay demonstrates, for the first time, nanosensor-driven NDE-Aß42 detection, which enables a reliable discrimination of AD patients from non-AD individuals and even the differential diagnosis between AD and vascular dementia patients, with an accuracy of 100% and a Youden index of 1. This NDE-Aß42 biosensor defines a robust approach for blood-based confident AD ascertain.

Shot-noise-limited photon-counting precoding scheme for MIMO ultraviolet communication in atmospheric turbulence.

Photon counting is a promising solution to detecting low-power optical signals for ultraviolet (UV) communications in the forthcoming sixth-generation (6G) network. Different from the conventional additive white Gaussian noise (AWGN) model, the discrete signal-dependent Poisson shot noise poses challenges to the signal processing of photon-counting systems. In this paper, a joint design of precoder and equalizer is proposed for photon-counting multiple-input multiple-output (PhC-MIMO) UV systems. To circumvent the impasse arising from the signal-dependent shot noise, we propose an alternating optimization algorithm based on the minimum mean squared error (MMSE) criterion. The algorithm decomposes the joint design into convex subproblems solved in an alternating manner, and guarantees at least a stationary point solution. Numerical results corroborate that the proposed system exhibits robustness to turbulence fading and offers high throughput while mitigating the adverse effect of background radiation noise. Specifically, the 32 × 8 system can achieve a bit error rate (BER) of 10-5 at the signal energy of -154.0 dBJ per bit under strong Gamma-Gamma turbulence with the scintillation index (S.I.) of 3.

Quercetin alleviates tubulointerstitial inflammation by inhibiting exosomes-mediated crosstalk between tubular epithelial cells and macrophages.

BACKGROUND: Tubulointerstitial inflammation (TII) is a critical pathological feature of kidney disease leading to renal fibrosis, and its treatment remains a major clinical challenge. We sought to explore the role of quercetin, a potential exosomes inhibitor, in exosomes release and TII. METHODS: The effects of quercetin on exosomes release and TII were examined by two TII mouse models: the unilateral ureteral obstruction (UUO) models and the LPS-induced mouse models. In vitro, exosomes-mediated crosstalk between tubular epithelial cells (TECs) and macrophages was performed to investigate the mechanisms by which quercetin inhibited exosomes and TII. RESULTS: In this study, we found that exosomes-mediated crosstalk between TECs and macrophages contributed to the development of TII. In vitro, exosomes released from LPS-stimulated TECs induced increased expression of inflammatory cytokines and fibrotic markers in Raw264·7 cells and vice versa. Interestingly, heat shock protein 70 (Hsp70) or Hsp90 proteins could control exosomes release from TECs and macrophages both in vivo and in vitro. Importantly, quercetin, a previously recognized heat shock protein inhibitor, could significantly reduce exosomes release in TII models by down-regulating Hsp70 or Hsp90. Quercetin abrogated exosomes-mediated intercellular communication, which attenuated TII and renal fibrosis accordingly. CONCLUSION: Quercetin could serve as a novel strategy for treatment of tubulointerstitial inflammation by inhibiting the exosomes-mediated crosstalk between tubules and macrophages.

Comprehensive analysis of the roles of fatty acid transport related proteins in clear cell renal cell carcinoma.

OBJECTIVE: This study aimed to explore the clinical significance of fatty acid transport-related protein (FATRP) in patients with clear cell renal cell carcinoma(ccRCC). METHODS: RNA-seq data and corresponding clinical data of ccRCC were obtained from TCGA data portal. Seventeen key FATRP genes were comprehensively investigated using bioinformatics approaches to systematically investigate their expression patterns in ccRCC. In addition, the correlation between the expression levels of these genes and clinicopathological features in ccRCC was further explored. RESULTS: Among the 17 key FATRP genes, only FABP5, FABP6, and FABP7 could be regarded as ideal biomarkers for ccRCC, as they were highly expressed in ccRCC tumor tissues, and positively correlates with tumor progression and poor prognosis. FABP6 had the highest copy number variations (CNV) events (63.07 %), and ccRCC patients with FABP6 amplification had a better prognosis than the unaltered group. DNA methylation levels of FABP6 and FABP7 were downregulated in ccRCC tumor tissues compared to those in normal tissues. FABP5 showed the opposite results. Moreover, a novel four FATRP gene (FABP1, FABP5, FABP7, FATP2) and three clinical parameter (age, stage, and grade) prediction model was constructed and that comprised a significant independent prognostic signature. CONCLUSIONS: Only a few FATRP genes are upregulated in ccRCC tumor tissue, and positively correlate with tumor progression and poor prognosis. The accuracy of a single gene of these FATRP genes as predictors of progression and prognosis of ccRCC is limited. The performance of the novel prediction model proposed by this study was much better than that of any single gene.

Isolation and identification of hemostatic steroidal glycosides from Ypsilandra thibetica.

The phytoconstituents of the fraction with hemostatic activity of the 70% aqueous ethanol extract of Ypsilandra thibetica Franch. were investigated. As a result, fourteen previously unreported spirostanol saponins, ypsilandrosides Z1-Z14, and nine known analogues were isolated and characterized by MS, NMR, and chemical methods. Among them, ypsilandrosides Z1-Z4 (1-4) have a rare 12-O-ß-d-glucopyranosyl group, while ypsilandrosides Z5-Z8 (5-8) possess a rare double bond between C-4 and C-5, and a hydroxyl or carbonyl located at the C-6. All isolates were further tested for their hemostatic activity. The results suggested that five spirostanol tetraglycosides show favorable inducing platelet aggregation activities. Among them, ypsilandroside G (16) displayed significant inducing platelet aggregation activity with an EC50 value of 57.17 µM. Furthermore, the preliminary structure-activity relationship of these spirostanol glycosides' hemostatic activity was discussed.

Tubular-specific CDK12 knockout causes a defect in urine concentration due to premature cleavage of the slc12a1 gene.

Cyclin-dependent kinase 12 (CDK12) plays a critical role in regulating gene transcription. CDK12 inhibition is a potential anticancer therapeutic strategy. However, several clinical trials have shown that CDK inhibitors might cause renal dysfunction and electrolyte disorders. CDK12 is abundant in renal tubular epithelial cells (RTECs), but the exact role of CDK12 in renal physiology remains unclear. Genetic knockout of CDK12 in mouse RTECs causes polydipsia, polyuria, and hydronephrosis. This phenotype is caused by defects in water reabsorption that are the result of reduced Na-K-2Cl cotransporter 2 (NKCC2) levels in the kidney. In addition, CKD12 knockout causes an increase in Slc12a1 (which encodes NKCC2) intronic polyadenylation events, which results in Slc12a1 truncated transcript production and NKCC2 downregulation. These findings provide novel insight into CDK12 being necessary for maintaining renal homeostasis by regulating NKCC2 transcription, which explains the critical water and electrolyte disturbance that occurs during the application of CDK12 inhibitors for cancer treatment. Therefore, there are safety concerns about the clinical use of these new anticancer drugs.

Dynamics of consumer-resource reaction-diffusion models: single and multiple consumer species.

A consumer-resource reaction-diffusion model with a single consumer species was proposed and experimentally studied by Zhang et al.(Ecol Lett 20:1118-1128, 2017). Analytical study on its dynamics was further performed by He et al.(J Math Biol 78:1605-1636, 2019). In this work, we completely settle the conjecture proposed by He et al.(J Math Biol 78:1605-1636, 2019) about the global dynamics of the consumer-resource model for small yield rate. We then study a multi-species consumer-resource model where all the consumer species compete with each other through depression of the limited resources by consumption and there is no direct competition between them. We show that in this case, all consumer species persist uniformly, which implies that "competition exclusion" phenomenon will never happen. We also clarify its dynamics in both homogeneous and heterogeneous environments under various circumstances.

Post-translational modification of RNA m6A demethylase ALKBH5 regulates ROS-induced DNA damage response.

Faithful genome integrity maintenance plays an essential role in cell survival. Here, we identify the RNA demethylase ALKBH5 as a key regulator that protects cells from DNA damage and apoptosis during reactive oxygen species (ROS)-induced stress. We find that ROS significantly induces global mRNA N6-methyladenosine (m6A) levels by modulating ALKBH5 post-translational modifications (PTMs), leading to the rapid and efficient induction of thousands of genes involved in a variety of biological processes including DNA damage repair. Mechanistically, ROS promotes ALKBH5 SUMOylation through activating ERK/JNK signaling, leading to inhibition of ALKBH5 m6A demethylase activity by blocking substrate accessibility. Moreover, ERK/JNK/ALKBH5-PTMs/m6A axis is activated by ROS in hematopoietic stem/progenitor cells (HSPCs) in vivo in mice, suggesting a physiological role of this molecular pathway in the maintenance of genome stability in HSPCs. Together, our study uncovers a molecular mechanism involving ALKBH5 PTMs and increased mRNA m6A levels that protect genomic integrity of cells in response to ROS.

Extracting laboratory test information from paper-based reports.

BACKGROUND: In the healthcare domain today, despite the substantial adoption of electronic health information systems, a significant proportion of medical reports still exist in paper-based formats. As a result, there is a significant demand for the digitization of information from these paper-based reports. However, the digitization of paper-based laboratory reports into a structured data format can be challenging due to their non-standard layouts, which includes various data types such as text, numeric values, reference ranges, and units. Therefore, it is crucial to develop a highly scalable and lightweight technique that can effectively identify and extract information from laboratory test reports and convert them into a structured data format for downstream tasks. METHODS: We developed an end-to-end Natural Language Processing (NLP)-based pipeline for extracting information from paper-based laboratory test reports. Our pipeline consists of two main modules: an optical character recognition (OCR) module and an information extraction (IE) module. The OCR module is applied to locate and identify text from scanned laboratory test reports using state-of-the-art OCR algorithms. The IE module is then used to extract meaningful information from the OCR results to form digitalized tables of the test reports. The IE module consists of five sub-modules, which are time detection, headline position, line normalization, Named Entity Recognition (NER) with a Conditional Random Fields (CRF)-based method, and step detection for multi-column. Finally, we evaluated the performance of the proposed pipeline on 153 laboratory test reports collected from Peking University First Hospital (PKU1). RESULTS: In the OCR module, we evaluate the accuracy of text detection and recognition results at three different levels and achieved an averaged accuracy of 0.93. In the IE module, we extracted four laboratory test entities, including test item name, test result, test unit, and reference value range. The overall F1 score is 0.86 on the 153 laboratory test reports collected from PKU1. With a single CPU, the average inference time of each report is only 0.78 s. CONCLUSION: In this study, we developed a practical lightweight pipeline to digitalize and extract information from paper-based laboratory test reports in diverse types and with different layouts that can be adopted in real clinical environments with the lowest possible computing resources requirements. The high evaluation performance on the real-world hospital dataset validated the feasibility of the proposed pipeline.

Learning spatiotemporal features of DSA using 3D CNN and BiConvGRU for ischemic moyamoya disease detection.

BACKGROUND: Moyamoya disease (MMD) is a serious intracranial cerebrovascular disease. Cerebral hemorrhage caused by MMD will bring life risk to patients. Therefore, MMD detection is of great significance in the prevention of cerebral hemorrhage. In order to improve the accuracy of digital subtraction angiography (DSA) in the diagnosis of ischemic MMD, in this paper, a deep network architecture combined with 3D convolutional neural network (3D CNN) and bidirectional convolutional gated recurrent unit (BiConvGRU) is proposed to learn the spatiotemporal features for ischemic MMD detection. METHODS: Firstly, 2D convolutional neural network (2D CNN) is utilized to extract spatial features for each frame of DSA. Secondly, the long-term spatiotemporal features of DSA sequence are extracted by BiConvGRU. Thirdly, the short-term spatiotemporal features of DSA are further extracted by 3D convolutional neural network (3D CNN). In addition, different features are extracted when gray images and optical flow images pass through the network, and multiple features are extracted by features fusion. Finally, the fused features are utilized to classify. RESULTS: The proposed method was quantitatively evaluated on a data sets of 630 cases. The experimental results showed a detection accuracy of 0.9788, sensitivity and specificity were 0.9780 and 0.9796, respectively, and area under curve (AUC) was 0.9856. Compared with other methods, we can get the highest accuracy and AUC. CONCLUSIONS: The experimental results show that the proposed method is stable and reliable for ischemic MMD detection, which provides an option for doctors to accurately diagnose ischemic MMD.