Circadian clocks are modulated by compartmentalized oscillating translation.

Terrestrial organisms developed circadian rhythms for adaptation to Earth's quasi-24-h rotation. Achieving precise rhythms requires diurnal oscillation of fundamental biological processes, such as rhythmic shifts in the cellular translational landscape; however, regulatory mechanisms underlying rhythmic translation remain elusive. Here, we identified mammalian ATXN2 and ATXN2L as cooperating master regulators of rhythmic translation, through oscillating phase separation in the suprachiasmatic nucleus along circadian cycles. The spatiotemporal oscillating condensates facilitate sequential initiation of multiple cycling processes, from mRNA processing to protein translation, for selective genes including core clock genes. Depleting ATXN2 or 2L induces opposite alterations to the circadian period, whereas the absence of both disrupts translational activation cycles and weakens circadian rhythmicity in mice. Such cellular defect can be rescued by wild type, but not phase-separation-defective ATXN2. Together, we revealed that oscillating translation is regulated by spatiotemporal condensation of two master regulators to achieve precise circadian rhythm in mammals.

Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays.

Spatially resolved transcriptomic technologies are promising tools to study complex biological processes such as mammalian embryogenesis. However, the imbalance between resolution, gene capture, and field of view of current methodologies precludes their systematic application to analyze relatively large and three-dimensional mid- and late-gestation embryos. Here, we combined DNA nanoball (DNB)-patterned arrays and in situ RNA capture to create spatial enhanced resolution omics-sequencing (Stereo-seq). We applied Stereo-seq to generate the mouse organogenesis spatiotemporal transcriptomic atlas (MOSTA), which maps with single-cell resolution and high sensitivity the kinetics and directionality of transcriptional variation during mouse organogenesis. We used this information to gain insight into the molecular basis of spatial cell heterogeneity and cell fate specification in developing tissues such as the dorsal midbrain. Our panoramic atlas will facilitate in-depth investigation of longstanding questions concerning normal and abnormal mammalian development.

A Common Embryonic Origin of Stem Cells Drives Developmental and Adult Neurogenesis.

New neurons arise from quiescent adult neural progenitors throughout life in specific regions of the mammalian brain. Little is known about the embryonic origin and establishment of adult neural progenitors. Here, we show that Hopx+ precursors in the mouse dentate neuroepithelium at embryonic day 11.5 give rise to proliferative Hopx+ neural progenitors in the primitive dentate region, and they, in turn, generate granule neurons, but not other neurons, throughout development and then transition into Hopx+ quiescent radial glial-like neural progenitors during an early postnatal period. RNA-seq and ATAC-seq analyses of Hopx+ embryonic, early postnatal, and adult dentate neural progenitors further reveal common molecular and epigenetic signatures and developmental dynamics. Together, our findings support a "continuous" model wherein a common neural progenitor population exclusively contributes to dentate neurogenesis throughout development and adulthood. Adult dentate neurogenesis may therefore represent a lifelong extension of development that maintains heightened plasticity in the mammalian hippocampus.

Fibroblast growth factor 13 is a microtubule-stabilizing protein regulating neuronal polarization and migration.

Secretory fibroblast growth factors (FGFs) and their receptors are known for their regulatory function in the early stages of neural development. FGF13, a nonsecretory protein of the FGF family, is expressed in cerebral cortical neurons during development and is a candidate gene for syndromal and nonspecific forms of X-chromosome-linked mental retardation (XLMR). However, its function during development remains unclear. We show that FGF13 acts intracellularly as a microtubule-stabilizing protein required for axon and leading process development and neuronal migration in the cerebral cortex. FGF13 is enriched in axonal growth cones and interacts directly with microtubules. Furthermore, FGF13 polymerizes tubulins and stabilizes microtubules. The loss of FGF13 impairs neuronal polarization and increases the branching of axons and leading processes. Genetic deletion of FGF13 in mice results in neuronal migration defects in both the neocortex and the hippocampus. FGF13-deficient mice also exhibit weakened learning and memory, which is correlated to XLMR patients' intellectual disability.

An epigenetic circuit controls neurogenic programs during neocortex development.

The production and expansion of intermediate progenitors (IPs) are essential for neocortical neurogenesis during development and over evolution. Here, we have characterized an epigenetic circuit that precisely controls neurogenic programs, particularly properties of IPs, during neocortical development. The circuit comprises a long non-coding RNA (LncBAR) and the BAF (SWI/SNF) chromatin-remodeling complex, which transcriptionally maintains the expression of Zbtb20. LncBAR knockout neocortex contains more deep-layer but fewer upper-layer projection neurons. Intriguingly, loss of LncBAR promotes IP production, but paradoxically prolongs the duration of the cell cycle of IPs during mid-later neocortical neurogenesis. Moreover, in LncBAR knockout mice, depletion of the neural progenitor pool at embryonic stage results in fewer adult neural progenitor cells in the subventricular zone of lateral ventricles, leading to a failure in adult neurogenesis to replenish the olfactory bulb. LncBAR binds to BRG1, the core enzymatic component of the BAF chromatin-remodeling complex. LncBAR depletion enhances association of BRG1 with the genomic locus of, and suppresses the expression of, Zbtb20, a transcription factor gene known to regulate both embryonic and adult neurogenesis. ZBTB20 overexpression in LncBAR-knockout neural precursors reverses compromised cell cycle progressions of IPs.

Boosting Polysulfide Redox Kinetics by Temperature-Induced Metal-Insulator Transition Effect of Tungsten-Doped Vanadium Dioxide for High-Temperature Lithium-Sulfur Batteries.

The practical application of Li-S batteries is still severely restricted by poor cyclic performance caused by the intrinsic polysulfides shuttle effect, which is even more severe under the high-temperature condition owing to the inevitable increase of polysulfides' solubility and diffusion rate. Herein, tungsten-doped vanadium dioxide (W-VO2) micro-flowers are employed with first-order metal-insulator phase transition (MIT) property as a robust and multifunctional modification layer to hamper the shuttle effect and simultaneously improve the thermotolerance of the common separator. Tungsten doping significantly reduces the transition temperature from 68 to 35 °C of vanadium dioxide, which renders the W-VO2 easier to turn from the insulating monoclinic phase into the metallic rutile phase. The systematic experiments and theoretical analysis demonstrate that the temperature-induced in-suit MIT property endows the W-VO2 catalyst with strong chemisorption against polysulfides, low energy barrier for liquid-to-solid conversion, and outstanding diffusion kinetics of Li-ion under high temperatures. Benefiting from these advantages, the Li-S batteries with W-VO2 modified separator exhibit significantly improved rate and long-term cyclic performance under 50 °C. Remarkably, even at an elevated temperature (80 °C), they still exhibit superior electrochemical performance. This work opens a rewarding avenue to use phase-changing materials for high-temperature Li-S batteries.

Optimizing Surface State Electrons of Topological Semi-Metal by Atomic Doping for Enhanced Hydrogen Evolution Reaction.

Topological materials carrying topological surface states (TSSs) have extraordinary carrier mobility and robustness, which provide a new platform for searching for efficient hydrogen evolution reaction (HER) electrocatalysts. However, the majority of these TSSs originate from the sp band of topological quantum catalysts rather than the d band. Here, based on the density functional theory calculation, it is reported a topological semimetal Pd3Sn carrying TSSs mainly derived from d orbital and proposed that optimizing surface state electrons of Pd3Sn by introduction heteroatoms (Ni) can promote hybridization between hydrogen atoms and electrons, thereby reducing the Gibbs free energy (ΔGH) of adsorbed hydrogen and improving its HER performance. Moreover, this is well verified by electrocatalytic experiment results, the Ni-doped Pd3Sn (Ni0.1Pd2.9Sn) show much lower overpotential (-29 mV vs RHE) and Tafel slope (17 mV dec-1) than Pd3Sn (-39 mV vs RHE, 25 mV dec-1) at a current density of 10 mA cm-2. Significantly, the Ni0.1Pd2.9Sn nanoparticles exhibit excellent stability for HER. The electrocatalytic activity of Ni0.1Pd2.9Sn nanoparticles is superior to that of commercial Pt. This work provides an accurate guide for manipulating surface state electrons to improve the HER performance of catalysts.

Comparative Proteome and Cis-Regulatory Element Analysis Reveals Specific Molecular Pathways Conserved in Dog and Human Brains.

Brain development and function are governed by precisely regulated protein expressions in different regions. To date, multiregional brain proteomes have been systematically analyzed only for adult human and mouse brains. To understand the underpinnings of brain development and function, we generated proteomes from six regions of the postnatal brain at three developmental stages of domestic dogs (Canis familiaris), which are special among animals in terms of their remarkable human-like social cognitive abilities. Quantitative analysis of the spatiotemporal proteomes identified region-enriched synapse types at different developmental stages and differential myelination progression in different brain regions. Through integrative analysis of inter-regional expression patterns of orthologous proteins and genome-wide cis-regulatory element frequencies, we found that proteins related with myelination and hippocampus were highly correlated between dog and human but not between mouse and human, although mouse is phylogenetically closer to human. Moreover, the global expression patterns of neurodegenerative disease and autism spectrum disorder-associated proteins in dog brain more resemble human brain than in mouse brain. The high similarity of myelination and hippocampus-related pathways in dog and human at both proteomic and genetic levels may contribute to their shared social cognitive abilities. The inter-regional expression patterns of disease-associated proteins in the brain of different species provide important information to guide mechanistic and translational study using appropriate animal models.

Effects of ambient temperature and humidity on COPD mortality in Ganzhou city, China.

This study used the time series data of Ganzhou city to explore the individual and interaction effects of temperature and humidity on COPD death, and identify vulnerable subgroups of the population. We collected daily COPD mortality and meteorological data in Ganzhou from 2016 to 2019. The nonlinear distribution lag model was used to examine the associations and interaction between daily mean temperature and humidity and COPD mortality. For the total population, male and 65 years old or above, the relative risk (RR) for COPD mortality could be significant at extremely low temperature (3.3 ℃), reaching 1.799 (95% confidence interval [CI]: 1.216, 2.662), 1.894 (95% CI: 1.164, 3.084) and 1.779 (95% CI:1.185, 2.670). Also, at extremely low humidity (47.8%), the risk reached 1.888 (95% CI: 1.217, 2.930), 1.837 (95% CI: 1.066, 3.165) and 2.166 (95% CI: 1.375, 3.414). The cumulative COPD death risk for females was 3.524 (95% CI: 1.340, 9.267) at high temperature (30.7 ℃), 1.953(95% CI: 1.036, 3.683) at low humidity (47.8%) and 1.726 (95% CI: 1.048, 2.845) at high humidity (96.7%). For the total COPD deaths and subgroups, the interaction effects between daily temperature and humidity were not significant (p > 0.05). Both extremely low temperature and low humidity increased the risk of COPD death in Ganzhou city, especially for males and people over 65 years old. Females were more sensitive to extremely high temperature and humidity. Patients with COPD should pay attention to self-protection under extreme temperature and humidity weather conditions.

Modulating d-Band Electronic Structures of Molybdenum Disulfide via p/n Doping to Boost Polysulfide Conversion in Lithium-Sulfur Batteries.

Polysulfide shuttle effect and sluggish sulfur reaction kinetics severely impede the cycling stability and sulfur utilization of lithium-sulfur (Li-S) batteries. Modulating d-band electronic structures of molybdenum disulfide electrocatalysts via p/n doping is promising to boost polysulfide conversion and suppress polysulfide migration in lithium-sulfur batteries. Herein, p-type V-doped MoS2 (V-MoS2 ) and n-type Mn-doped MoS2 (Mn-MoS2 ) catalysts are well-designed. Experimental results and theoretical analyses reveal that both of them significantly increase the binding energy of polysulfides on the catalysts' surface and accelerate the sluggish conversion kinetics of sulfur species. Particularly, the p-type V-MoS2 catalyst exhibits a more obvious bidirectional catalytic effect. Electronic structure analysis further demonstrates that the superior anchoring and electrocatalytic activities are originated from the upward shift of the d-band center and the optimized electronic structure induced by duplex metal coupling. As a result, the Li-S batteries with V-MoS2 modified separator exhibit a high initial capacity of 1607.2 mAh g-1 at 0.2 C and excellent rate and cycling performance. Moreover, even at a high sulfur loading of 6.84 mg cm-2 , a favorable initial areal capacity of 8.98 mAh cm-2 is achieved at 0.1 C. This work may bring widespread attention to atomic engineering in catalyst design for high-performance Li-S batteries.

Spectral optimization of trichromatic white LEDs based on age of lighting user and application scene.

The optimization of trichromatic white light emitting diodes (LEDs) spectrum for application scenes related to the age of lighting users is proposed and demonstrated. Based on the spectral transmissivity of human eyes at different ages, the visual and non-visual responses of human eyes to different wavelengths of light, we have built the blue light hazards (BLH) and circadian action factor (CAF) related to the age of the lighting user. The BLH and CAF are used to evaluate the spectral combinations of high color rendering index (CRI) white LEDs obtained from different radiation flux ratios of red, green, and blue monochrome spectrum. The best spectra of white LEDs for lighting users at different ages in work and leisure scenes are achieved due to the optimization criterion of BLH proposed by us. This research provides a solution for intelligent health lighting design applicable to light users of different ages and application scenes.

Inactivating Celsr2 promotes motor axon fasciculation and regeneration in mouse and human.

Understanding new modulators of axon regeneration is central to neural repair. Our previous work demonstrated critical roles of atypical cadherin Celsr2 during neural development, including cilia organization, neuron migration and axon navigation. Here, we address its role in axon regeneration. We show that Celsr2 is highly expressed in both mouse and human spinal motor neurons. Celsr2 knockout promotes axon regeneration and fasciculation in mouse cultured spinal explants. Similarly, cultured Celsr2 mutant motor neurons extend longer neurites and larger growth cones, with increased expression of end-binding protein 3 and higher potassium-induced calcium influx. Mice with Celsr2 conditional knockout in spinal motor neurons do not exhibit any behavioural deficits; however, after branchial plexus injury, axon regeneration and functional forelimb locomotor recovery are significantly improved. Similarly, knockdown of CELSR2 using shRNA interference in cultured human spinal motor explants and motor neurons increases axonal fasciculation and growth. In mouse adult spinal cord after root avulsion, in mouse embryonic spinal cords, and in cultured human motor neurons, Celsr2 downregulation is accompanied by increased levels of GTP-bound Rac1 and Cdc42, and of JNK and c-Jun. In conclusion, Celsr2 negatively regulates motor axon regeneration and is a potential target to improve neural repair.

Twist1 regulates macrophage plasticity to promote renal fibrosis through galectin-3.

Renal interstitial fibrosis is the pathological basis of end-stage renal disease, in which the heterogeneity of macrophages in renal microenvironment plays an important role. However, the molecular mechanisms of macrophage plasticity during renal fibrosis progression remain unclear. In this study, we found for the first time that increased expression of Twist1 in macrophages was significantly associated with the severity of renal fibrosis in IgA nephropathy patients and mice with unilateral ureteral obstruction (UUO). Ablation of Twist1 in macrophages markedly alleviated renal tubular injury and renal fibrosis in UUO mice, accompanied by a lower extent of macrophage infiltration and M2 polarization in the kidney. The knockdown of Twist1 inhibited the chemotaxis and migration of macrophages, at least partially, through the CCL2/CCR2 axis. Twist1 downregulation inhibited M2 macrophage polarization and reduced the secretion of the profibrotic factors Arg-1, MR (CD206), IL-10, and TGF-ß. Galectin-3 was decreased in the macrophages of the conditional Twist1-deficient mice, and Twist1 was shown to directly activate galectin-3 transcription. Up-regulation of galectin-3 recovered Twist1-mediated M2 macrophage polarization. In conclusion, Twist1/galectin-3 signaling regulates macrophage plasticity (M2 phenotype) and promotes renal fibrosis. This study could suggest new strategies for delaying kidney fibrosis in patients with chronic kidney disease.

Association between sleep disorders during pregnancy and risk of postpartum depression: a systematic review and meta-analysis.

Sleep disorders are common among pregnant females. However, its association with postpartum depression (PPD) is unknown. We aimed to assess if sleep disorders during pregnancy increase the risk of PPD by a systematic review. The databases of PubMed, CENTRAL, ScienceDirect, Embase, and Google Scholar were searched for studies reporting the association between any type of sleep disorder during pregnancy and the risk of PPD. Effect sizes were pooled in a random-effects model. Sixteen studies with data of 12,614 women were included. Meta-analysis indicated that sleep disorders during pregnancy resulted in a statistically significant increased risk of PPD (OR: 2.36 95% CI: 1.72, 2.32). The overall result had high heterogeneity (I2 = 84%). Sub-group analysis based on study location (Asian vs Western), sample size (> 500 vs < 500), depression scale, and PPD assessment time did not change the results. However, we found that only poor sleep quality but not insomnia was associated with PPD. The risk was also increased only with sleep disorders measured in the 3rd trimester but not for the 1st and 2nd trimesters. Evidence suggests that sleep disorders during pregnancy may increase risk of PPD. The risk is high for sleep disorders occurring in the 3rd trimester. Based on these findings, there is a need for thorough screening and subsequent corrective measures to ensure adequate and quality sleep among pregnant females.

Research on Anomaly Network Detection Based on Self-Attention Mechanism.

Network traffic anomaly detection is a key step in identifying and preventing network security threats. This study aims to construct a new deep-learning-based traffic anomaly detection model through in-depth research on new feature-engineering methods, significantly improving the efficiency and accuracy of network traffic anomaly detection. The specific research work mainly includes the following two aspects: 1. In order to construct a more comprehensive dataset, this article first starts from the raw data of the classic traffic anomaly detection dataset UNSW-NB15 and combines the feature extraction standards and feature calculation methods of other classic detection datasets to re-extract and design a feature description set for the original traffic data in order to accurately and completely describe the network traffic status. We reconstructed the dataset DNTAD using the feature-processing method designed in this article and conducted evaluation experiments on it. Experiments have shown that by verifying classic machine learning algorithms, such as XGBoost, this method not only does not reduce the training performance of the algorithm but also improves its operational efficiency. 2. This article proposes a detection algorithm model based on LSTM and the recurrent neural network self-attention mechanism for important time-series information contained in the abnormal traffic datasets. With this model, through the memory mechanism of the LSTM, the time dependence of traffic features can be learned. On the basis of LSTM, a self-attention mechanism is introduced, which can weight the features at different positions in the sequence, enabling the model to better learn the direct relationship between traffic features. A series of ablation experiments were also used to demonstrate the effectiveness of each component of the model. The experimental results show that, compared to other comparative models, the model proposed in this article achieves better experimental results on the constructed dataset.

Ketogenic diet promotes the proliferation of oligodendrocyte precursor cells in ME region by activating fatty acid oxidation.

Hypothalamic median eminence (ME) is a potential niche for neurons and oligodendrocytes, and trophic factors may regulate hypothalamic function by inducing cellular changes in the ME region. To determine whether diet-induced plasticity exists in hypothalamic stem cells dormant under physiological conditions, we used a combination of a normal diet, a high-fat diet, and a ketogenic diet (a low-carb, high-fat diet) to compare the proliferation of tanycytes (TCs) and oligodendrocyte precursor cells (OPCs) in the ME area of mice under the different diets. The results showed that the ketogenic diet could induce and promote the proliferation of OPCs in the ME area, and blocking the fatty acid oxidation program could inhibit the proliferation of OPCs induced by a ketogenic diet. This study preliminarily revealed the diet-induced effect on OPCs in the ME region and provided enlightenment for further study on the function of OPCs in the ME region.

Tannin-Mn coordination polymer coated carbon quantum dots nanocomposite for fluorescence and magnetic resonance bimodal imaging.

The MR/FI bimodal imaging has attracted widely studied due to combining the advantages of MRI and FI can bridge gaps in sensitivity and depth between these two modalities. Herein, a novel MR/FI bimodal imaging probe is facile fabricated by coating the Mn-phenolic coordination polymer on the surface of the carbon quantum dots. The structure of the as-prepared nanocomposite probe is carefully validated via SEM, TEM, and XPS. The content of Mn2+ is calculated through the EDS and TGA. The quantum yield (QY) and emission wavelength of the probe are about 7.24% and 490 nm, respectively. The longitudinal r1 value (2.43 mM-1 s-1) with low r2/r1 (4.45) of the probe is obtained. Subsequently, fluorescence and MR imaging are performed. The metabolic pathways in vivo are inferred by studying the bio-distribution of the probe in major organs. Thus, these results indicate that probe would be an excellent dual-modal imaging probe for enhanced MR imaging and fluorescence imaging. MR/FI bimodal imaging probe is built via in-situ coated Mn-phenolic coordination polymer on the surface of the carbon quantum dots. The in vitro and vivo image property of the probe is evaluated.

In vivo clonal analysis reveals spatiotemporal regulation of thalamic nucleogenesis.

The thalamus, a crucial regulator of cortical functions, is composed of many nuclei arranged in a spatially complex pattern. Thalamic neurogenesis occurs over a short period during mammalian embryonic development. These features have hampered the effort to understand how regionalization, cell divisions, and fate specification are coordinated and produce a wide array of nuclei that exhibit distinct patterns of gene expression and functions. Here, we performed in vivo clonal analysis to track the divisions of individual progenitor cells and spatial allocation of their progeny in the developing mouse thalamus. Quantitative analysis of clone compositions revealed evidence for sequential generation of distinct sets of thalamic nuclei based on the location of the founder progenitor cells. Furthermore, we identified intermediate progenitor cells that produced neurons populating more than one thalamic nuclei, indicating a prolonged specification of nuclear fate. Our study reveals an organizational principle that governs the spatial and temporal progression of cell divisions and fate specification and provides a framework for studying cellular heterogeneity and connectivity in the mammalian thalamus.

Tongue image quality assessment based on a deep convolutional neural network.

BACKGROUND: Tongue diagnosis is an important research field of TCM diagnostic technology modernization. The quality of tongue images is the basis for constructing a standard dataset in the field of tongue diagnosis. To establish a standard tongue image database in the TCM industry, we need to evaluate the quality of a massive number of tongue images and add qualified images to the database. Therefore, an automatic, efficient and accurate quality control model is of significance to the development of intelligent tongue diagnosis technology for TCM. METHODS: Machine learning methods, including Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), Adaptive Boosting Algorithm (Adaboost), Naïve Bayes, Decision Tree (DT), Residual Neural Network (ResNet), Convolution Neural Network developed by Visual Geometry Group at University of Oxford (VGG), and Densely Connected Convolutional Networks (DenseNet), were utilized to identify good-quality and poor-quality tongue images. Their performances were made comparisons by using metrics such as accuracy, precision, recall, and F1-Score. RESULTS: The experimental results showed that the accuracy of the three deep learning models was more than 96%, and the accuracy of ResNet-152 and DenseNet-169 was more than 98%. The model ResNet-152 obtained accuracy of 99.04%, precision of 99.05%, recall of 99.04%, and F1-score of 99.05%. The performances were better than performances of other eight models. The eight models are VGG-16, DenseNet-169, SVM, RF, GBDT, Adaboost, Naïve Bayes, and DT. ResNet-152 was selected as quality-screening model for tongue IQA. CONCLUSIONS: Our research findings demonstrate various CNN models in the decision-making process for the selection of tongue image quality assessment and indicate that applying deep learning methods, specifically deep CNNs, to evaluate poor-quality tongue images is feasible.

Diabetic Retinopathy Prediction by Ensemble Learning Based on Biochemical and Physical Data.

(1) Background: Diabetic retinopathy, one of the most serious complications of diabetes, is the primary cause of blindness in developed countries. Therefore, the prediction of diabetic retinopathy has a positive impact on its early detection and treatment. The prediction of diabetic retinopathy based on high-dimensional and small-sample-structured datasets (such as biochemical data and physical data) was the problem to be solved in this study. (2) Methods: This study proposed the XGB-Stacking model with the foundation of XGBoost and stacking. First, a wrapped feature selection algorithm, XGBIBS (Improved Backward Search Based on XGBoost), was used to reduce data feature redundancy and improve the effect of a single ensemble learning classifier. Second, in view of the slight limitation of a single classifier, a stacking model fusion method, Sel-Stacking (Select-Stacking), which keeps Label-Proba as the input matrix of meta-classifier and determines the optimal combination of learners by a global search, was used in the XGB-Stacking model. (3) Results: XGBIBS greatly improved the prediction accuracy and the feature reduction rate of a single classifier. Compared to a single classifier, the accuracy of the Sel-Stacking model was improved to varying degrees. Experiments proved that the prediction model of XGB-Stacking based on the XGBIBS algorithm and the Sel-Stacking method made effective predictions on diabetes retinopathy. (4) Conclusion: The XGB-Stacking prediction model of diabetic retinopathy based on biochemical and physical data had outstanding performance. This is highly significant to improve the screening efficiency of diabetes retinopathy and reduce the cost of diagnosis.