Gasdermin D deficiency aborts myeloid calcium influx to drive granulopoiesis in lupus nephritis.

Gasdermin D (GSDMD) is emerging as an important player in autoimmune diseases, but its exact role in lupus nephritis (LN) remains controversial. Here, we identified markedly elevated GSDMD in human and mouse LN kidneys, predominantly in CD11b+ myeloid cells. Global or myeloid-conditional deletion of GSDMD was shown to exacerbate systemic autoimmunity and renal injury in lupus mice with both chronic graft-versus-host (cGVH) disease and nephrotoxic serum (NTS) nephritis. Interestingly, RNA sequencing and flow cytometry revealed that myeloid GSDMD deficiency enhanced granulopoiesis at the hematopoietic sites in LN mice, exhibiting remarkable enrichment of neutrophil-related genes, significant increases in total and immature neutrophils as well as granulocyte/macrophage progenitors (GMPs). GSDMD-deficient GMPs and all-trans-retinoic acid (ATRA)-stimulated human promyelocytes NB4 were further demonstrated to possess enhanced clonogenic and differentiation abilities compared with controls. Mechanistically, GSDMD knockdown promoted self-renewal and granulocyte differentiation by restricting calcium influx, contributing to granulopoiesis. Functionally, GSDMD deficiency led to increased pathogenic neutrophil extracellular traps (NETs) in lupus peripheral blood and bone marrow-derived neutrophils. Taken together, our data establish that GSDMD deletion accelerates LN development by promoting granulopoiesis in a calcium influx-regulated manner, unraveling its unrecognized critical role in LN pathogenesis.

Genome-wide association study and genomic selection of spike-related traits in bread wheat.

KEY MESSAGE: Identification of 337 stable MTAs for wheat spike-related traits improved model accuracy, and favorable alleles of MTA259 and MTA64 increased grain weight and yield per plant. Wheat (Triticum aestivum L.) is one of the three primary global, staple crops. Improving spike-related traits in wheat is crucial for optimizing spike and plant morphology, ultimately leading to increased grain yield. Here, we performed a genome-wide association study using a dataset of 24,889 high-quality unique single-nucleotide polymorphisms (SNPs) and phenotypic data from 314 wheat accessions across eight diverse environments. In total, 337 stable and significant marker-trait associations (MTAs) related to spike-related traits were identified. MTA259 and MTA64 were consistently detected in seven and six environments, respectively. The presence of favorable alleles associated with MTA259 and MTA64 significantly reduced wheat spike exsertion length and spike length, while enhancing thousand kernel weight and yield per plant. Combined gene expression and network analyses identified TraesCS6D03G0692300 and TraesCS6D03G0692700 as candidate genes for MTA259 and TraesCS2D03G0111700 and TraesCS2D03G0112500 for MTA64. The identified MTAs significantly improved the prediction accuracy of each model compared with using all the SNPs, and the random forest model was optimal for genome selection. Additionally, the eight stable and major MTAs, including MTA259, MTA64, MTA66, MTA94, MTA110, MTA165, MTA180, and MTA164, were converted into cost-effective and efficient detection markers. This study provided valuable genetic resources and reliable molecular markers for wheat breeding programs.

Intestinal CXCR6+ ILC3s migrate to the kidney and exacerbate renal fibrosis via IL-23 receptor signaling enhanced by PD-1 expression.

Group 3 innate lymphoid cells (ILC3s) regulate inflammation and tissue repair at mucosal sites, but whether these functions pertain to other tissues-like the kidneys-remains unclear. Here, we observed that renal fibrosis in humans was associated with increased ILC3s in the kidneys and blood. In mice, we showed that CXCR6+ ILC3s rapidly migrated from the intestinal mucosa and accumulated in the kidney via CXCL16 released from the injured tubules. Within the fibrotic kidney, ILC3s increased the expression of programmed cell death-1 (PD-1) and subsequent IL-17A production to directly activate myofibroblasts and fibrotic niche formation. ILC3 expression of PD-1 inhibited IL-23R endocytosis and consequently amplified the JAK2/STAT3/RORγt/IL-17A pathway that was essential for the pro-fibrogenic effect of ILC3s. Thus, we reveal a hitherto unrecognized migration pathway of ILC3s from the intestine to the kidney and the PD-1-dependent function of ILC3s in promoting renal fibrosis.

New insights into the role of immunity and inflammation in diabetic kidney disease in the omics era.

Diabetic kidney disease (DKD) is becoming the leading cause of chronic kidney disease, especially in the industrialized world. Despite mounting evidence has demonstrated that immunity and inflammation are highly involved in the pathogenesis and progression of DKD, the underlying mechanisms remain incompletely understood. Substantial molecules, signaling pathways, and cell types participate in DKD inflammation, by integrating into a complex regulatory network. Most of the studies have focused on individual components, without presenting their importance in the global or system-based processes, which largely hinders clinical translation. Besides, conventional technologies failed to monitor the different behaviors of resident renal cells and immune cells, making it difficult to understand their contributions to inflammation in DKD. Recently, the advancement of omics technologies including genomics, epigenomics, transcriptomics, proteomics, and metabolomics has revolutionized biomedical research, which allows an unbiased global analysis of changes in DNA, RNA, proteins, and metabolites in disease settings, even at single-cell and spatial resolutions. They help us to identify critical regulators of inflammation processes and provide an overview of cell heterogeneity in DKD. This review aims to summarize the application of multiple omics in the field of DKD and emphasize the latest evidence on the interplay of inflammation and DKD revealed by these technologies, which will provide new insights into the role of inflammation in the pathogenesis of DKD and lead to the development of novel therapeutic approaches and diagnostic biomarkers.

A Natural Small Molecule Mitigates Kidney Fibrosis by Targeting Cdc42-mediated GSK-3ß/ß-catenin Signaling.

Kidney fibrosis is a common fate of chronic kidney diseases (CKDs), eventually leading to renal dysfunction. Yet, no effective treatment for this pathological process has been achieved. During the bioassay-guided chemical investigation of the medicinal plant Wikstroemia chamaedaphne, a daphne diterpenoid, daphnepedunin A (DA), is characterized as a promising anti-renal fibrotic lead. DA shows significant anti-kidney fibrosis effects in cultured renal fibroblasts and unilateral ureteral obstructed mice, being more potent than the clinical trial drug pirfenidone. Leveraging the thermal proteome profiling strategy, cell division cycle 42 (Cdc42) is identified as the direct target of DA. Mechanistically, DA targets to reduce Cdc42 activity and down-regulates its downstream phospho-protein kinase Cζ(p-PKCζ)/phospho-glycogen synthase kinase-3ß (p-GSK-3ß), thereby promoting ß-catenin Ser33/37/Thr41 phosphorylation and ubiquitin-dependent proteolysis to block classical pro-fibrotic ß-catenin signaling. These findings suggest that Cdc42 is a promising therapeutic target for kidney fibrosis, and highlight DA as a potent Cdc42 inhibitor for combating CKDs.

Long-chain polyunsaturated fatty acids influence colorectal cancer progression via the interactions between the intestinal microflora and the macrophages.

The metabolism of long-chain polyunsaturated fatty acids (LCPUFAs) is closely associated with the risk and progression of colorectal cancer (CRC). This paper aims to investigate the role of LCPUFA in the crosstalk between intestinal microflora and macrophages, as well as the effects of these three parties on the progression of CRC. The metabolism and function of LCPUFA play important roles in regulating the composition of the human gut microflora and participating in the regulation of inflammation, ultimately affecting macrophage function and polarization, which is crucial in the tumor microenvironment. The effects of LCPUFA on cellular interactions between the two species can ultimately influence the progression of CRC. In this review, we explore the molecular mechanisms and clinical applications of LCPUFA in the interactions between intestinal microflora and intestinal macrophages, as well as its significance for CRC progression. Furthermore, we reveal the role of LCPUFA in the construction of the CRC microenvironment and explore the key nodes of the interactions between intestinal flora and intestinal macrophages in the environment. It provides potential targets for the metabolic diagnosis and treatment of CRC.

Digital Grading the Color Fastness to Rubbing of Fabrics Based on Spectral Reconstruction and BP Neural Network.

To digital grade the staining color fastness of fabrics after rubbing, an automatic grading method based on spectral reconstruction technology and BP neural network was proposed. Firstly, the modeling samples are prepared by rubbing the fabrics according to the ISO standard of 105-X12. Then, to comply with visual rating standards for color fastness, the modeling samples are professionally graded to obtain the visual rating result. After that, a digital camera is used to capture digital images of the modeling samples inside a closed and uniform lighting box, and the color data values of the modeling samples are obtained through spectral reconstruction technology. Finally, the color fastness prediction model for rubbing was constructed using the modeling samples data and BP neural network. The color fastness level of the testing samples was predicted using the prediction model, and the prediction results were compared with the existing color difference conversion method and gray scale difference method based on the five-fold cross-validation strategy. Experiments show that the prediction model of fabric color fastness can be better constructed using the BP neural network. The overall performance of the method is better than the color difference conversion method and the gray scale difference method. It can be seen that the digital rating method of fabric staining color fastness to rubbing based on spectral reconstruction and BP neural network has high consistency with the visual evaluation, which will help for the automatic color fastness grading.

Role of Exosomal Non-Coding RNA in the Tumour Microenvironment of Genitourinary System Tumours.

In recent years, genitourinary system tumors are common in people of all ages, seriously affecting the quality of life of patients, the pathogenesis and treatment of these diseases are constantly being updated and improved. Exosomes, with a lipid bilayer that enable delivery of their contents into body fluids or other cells. Exosomes can regulate the tumor microenvironment, and play an important role in tumor development. In turn, cellular and non-cellular components of tumor microenvironment also affect the occurrence, progression, invasion and metastasis of tumor. Non-coding RNAs have been shown to be able to be ingested and released by exosomes, and are seen as a potential tool in cancer diagnosis and treatment. Here, we summarize the effect of non-coding RNAs of exosome contents on the tumor microenvironment of genitourinary system tumor, expound the significance of non-coding RNAs of exosome in the occurrence, development, diagnosis and treatment of cancers.

Group 3 Innate Lymphoid Cells Exacerbate Lupus Nephritis by Promoting B Cell Activation in Kidney Ectopic Lymphoid Structures.

Group 3 innate lymphoid cells (ILC3s) represent a new population in immune regulation, yet their role in lupus nephritis (LN) remains elusive. In the present work, systemic increases in ILC3s, particularly in the kidney, are observed to correlate strongly with disease severity in both human and murine LN. Using MRL/lpr lupus mice and a nephrotoxic serum-induced LN model, this study demonstrates that ILC3s accumulated in the kidney migrate predominantly from the intestine. Furthermore, intestinal ILC3s accelerate LN progression, manifested by exacerbated autoimmunity and kidney injuries. In LN kidneys, ILC3s are located adjacent to B cells within ectopic lymphoid structures (ELS), directly activating B cell differentiation into plasma cells and antibody production in a Delta-like1 (DLL1)/Notch-dependent manner. Blocking DLL1 attenuates ILC3s' effects and protects against LN. Altogether, these findings reveal a novel pathogenic role of ILC3s in B cell activation, renal ELS formation and autoimmune injuries during LN, shedding light on the therapeutic value of targeting ILC3s for LN.

Enhancing Drug Repositioning through Local Interactive Learning with Bilinear Attention Networks.

Drug repositioning has emerged as a promising strategy for identifying new therapeutic applications for existing drugs. In this study, we present DRGBCN, a novel computational method that integrates heterogeneous information through a deep bilinear attention network to infer potential drugs for specific diseases. DRGBCN involves constructing a comprehensive drug-disease network by incorporating multiple similarity networks for drugs and diseases. Firstly, we introduce a layer attention mechanism to effectively learn the embeddings of graph convolutional layers from these networks. Subsequently, a bilinear attention network is constructed to capture pairwise local interactions between drugs and diseases. This combined approach enhances the accuracy and reliability of predictions. Finally, a multi-layer perceptron module is employed to evaluate potential drugs. Through extensive experiments on three publicly available datasets, DRGBCN demonstrates better performance over baseline methods in 10-fold cross-validation, achieving an average area under the receiver operating characteristic curve (AUROC) of 0.9399. Furthermore, case studies on bladder cancer and acute lymphoblastic leukemia confirm the practical application of DRGBCN in real-world drug repositioning scenarios. Importantly, our experimental results from the drug-disease network analysis reveal the successful clustering of similar drugs within the same community, providing valuable insights into drug-disease interactions. In conclusion, DRGBCN holds significant promise for uncovering new therapeutic applications of existing drugs, thereby contributing to the advancement of precision medicine.

The Impact of Photosynthetic Characteristics and Metabolomics on the Fatty Acid Biosynthesis in Tea Seeds.

The synthesis of tea fatty acids plays a crucial role in determining the oil content of tea seeds and selecting tea tree varieties suitable for harvesting both leaves and fruits. However, there is limited research on fatty acid synthesis in tea trees, and the precise mechanisms influencing tea seed oil content remain elusive. To reveal the fatty acid biosynthesis mechanism, we conducted a photosynthetic characteristic and targeted metabolomics analysis in comparison between Jincha 2 and Wuniuzao cultivars. Our findings revealed that Jincha 2 exhibited significantly higher net photosynthetic rates (Pn), stomatal conductance (Gs), and transpiration rate (Tr) compared with Wuniuzao, indicating the superior photosynthetic capabilities of Jincha 2. Totally, we identified 94 metabolites with significant changes, including key hormone regulators such as gibberellin A1 (GA1) and indole 3-acetic acid (IAA). Additionally, linolenic acid, methyl dihydrojasmonate, and methylthiobutyric acid, precursors required for fatty acid synthesis, were significantly more abundant in Jincha 2 compared with Wuniuzao. In summary, our research suggests that photosynthetic rates and metabolites contribute to the increased yield, fatty acid synthesis, and oil content observed in Jincha 2 when compared with Wuniuzao.

Measurement of the [Formula: see text]Ta([Formula: see text]) cross sections up to stellar s-process temperatures at the CSNS Back-n.

The neutron capture cross section of [Formula: see text]Ta is relevant to s-process of nuclear astrophysics, extraterrestrial samples analysis in planetary geology and new generation nuclear energy system design. The [Formula: see text]Ta([Formula: see text]) cross section had been measured between 1 eV and 800 keV at the back-streaming white neutron facility (Back-n) of China spallation neutron source(CSNS) using the time-of-flight (TOF) technique and [Formula: see text] liquid scintillator detectors. The experimental results are compared with the data of several evaluated libraries and previous experiments in the resolved and unresolved resonance region. Resonance parameters are extracted using the R-Matrix code SAMMY in the 1-700 eV region. The astrophysical Maxwell average cross section(MACS) from kT = 5 to 100 keV is calculated over a sufficiently wide range of neutron energies. For the characteristic thermal energy of an astrophysical site, at kT = 30keV the MACS value of [Formula: see text]Ta is 834 ± 75 mb, which shows an obvious discrepancy with the Karlsruhe Astrophysical Database of Nucleosynthesis in Stars (KADoNiS) recommended value 766 ± 15 mb. The new measurements strongly constrain the MACS of [Formula: see text]Ta([Formula: see text]) reaction in the stellar s-process temperatures.

Identification of shared gene signatures and molecular mechanisms between chronic kidney disease and ulcerative colitis.

Background: There is a complex interaction between chronic kidney disease (CKD) and ulcerative colitis (UC), but the pathophysiological mechanisms underlying the coexistence of CKD and UC are unclear. This study aimed to investigate the key molecules and pathways that may mediate the co-occurrence of CKD and UC through quantitative bioinformatics analysis based on a public RNA-sequencing database. Methods: The discovery datasets of CKD (GSE66494) and UC (GSE4183), as well as validation datasets of CKD (GSE115857) and UC (GSE10616), were downloaded from the Gene Expression Omnibus (GEO) database. After identifying differentially expressed genes (DEGs) with GEO2R online tool, the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses for the DEGs were performed. Next, protein-protein interaction network was constructed with Search Tool for the Retrieval of Interacting Genes (STRING) and visualized by Cytoscape. Gene modules were identified by the plug-in MCODE and hub genes were screened using the plug-in CytoHubba. Then, correlation between immune cell infiltration and hub genes was analyzed, and the receiver operating characteristic curves were used to assess the predictive value of hub genes. Finally, immunostaining of human specimens was used to validate the relevant findings. Results: A total of 462 common DEGs were identified and selected for further analyses. GO and KEGG enrichment analyses indicated that these DEGs were primarily enriched in immune- and inflammation-related pathways. Among them, the PI3K-Akt signaling pathway ranked top in both discovery and validation cohorts, and the key signal molecule phosphorylated Akt (p-Akt) was shown to be significantly overexpressed in human CKD kidneys and UC colons, and further elevated in CKD-UC comorbidity specimens. Moreover, nine candidate hub genes, including CXCL8, CCL2, CD44, ICAM1, IL1A, CXCR2, PTPRC, ITGAX, and CSF3, were identified, of which ICAM1 was validated as a common hub gene. Besides, immune infiltration analysis revealed that neutrophils, macrophages, and CD4+ T memory cells significantly accumulated in both diseases, and ICAM1 was remarkably associated with neutrophil infiltration. Furthermore, intercellular adhesion molecule1 (ICAM1)-mediated neutrophil infiltration was validated to be upregulated in kidney and colon biopsies of CKD and UC patients, and further increased in patients diagnosed with both CKD and UC. Finally, ICAM1 had shown critical value as a diagnostic marker for the co-occurrence of CKD and UC. Conclusions: Our study elucidated that immune response, PI3K-Akt signaling pathway, and ICAM1-mediated neutrophil infiltration might be the common pathogenesis of CKD and UC, and identified ICAM1 as a key potential biomarker and therapeutic target for the comorbidity of these two diseases.

Atomically Dispersed Manganese on Carbon Substrate for Aqueous and Aprotic CO2 Electrochemical Reduction.

CO2 utilization and conversion are of great importance in alleviating the rising CO2 concentration in the atmosphere. Here, a single-atom catalyst (SAC) is reported for electrochemical CO2 utilization in both aqueous and aprotic electrolytes. Specifically, atomically dispersed Mn-N4 sites are embedded in bowl-like mesoporous carbon particles with the functionalization of epoxy groups in the second coordination spheres. Theoretical calculations suggest that the epoxy groups near the Mn-N4 site adjust the electronic structure of the catalyst with reduced reaction energy barriers for the electrocatalytic reduction of CO2 to CO. The resultant Mn-single-atom carbon with N and O doped catalyst (MCs-(N,O)) exhibits extraordinary electrocatalytic performance with a high CO faradaic efficiency of 94.5%, a high CO current density of 13.7 mA cm-2 , and a low overpotential of 0.44 V in the aqueous environment. Meanwhile, as a cathode catalyst for aprotic Li-CO2 batteries, the MCs-(N,O) with well-regulated active sites and unique mesoporous bowl-like morphology optimizes the nucleation behavior of discharge products. MCs-(N,O)-based batteries deliver a low overpotential and excellent cyclic stability of 1000 h. The findings in this work provide a new avenue to design and fabricate SACs for various electrochemical CO2 utilization systems.

Optimized method for segmentation of ancient mural images based on superpixel algorithm.

High-precision segmentation of ancient mural images is the foundation of their digital virtual restoration. However, the complexity of the color appearance of ancient murals makes it difficult to achieve high-precision segmentation when using traditional algorithms directly. To address the current challenges in ancient mural image segmentation, an optimized method based on a superpixel algorithm is proposed in this study. First, the simple linear iterative clustering (SLIC) algorithm is applied to the input mural images to obtain superpixels. Then, the density-based spatial clustering of applications with noise (DBSCAN) algorithm is used to cluster the superpixels to obtain the initial clustered images. Subsequently, a series of optimized strategies, including (1) merging the small noise superpixels, (2) segmenting and merging the large noise superpixels, (3) merging initial clusters based on color similarity and positional adjacency to obtain the merged regions, and (4) segmenting and merging the color-mixing noisy superpixels in each of the merged regions, are applied to the initial cluster images sequentially. Finally, the optimized segmentation results are obtained. The proposed method is tested and compared with existing methods based on simulated and real mural images. The results show that the proposed method is effective and outperforms the existing methods.

Research on the deep learning-based exposure invariant spectral reconstruction method.

The surface spectral reflectance of an object is the key factor for high-fidelity color reproduction and material analysis, and spectral acquisition is the basis of its applications. Based on the theoretical imaging model of a digital camera, the spectral reflectance of any pixels in the image can be obtained through spectral reconstruction technology. This technology can avoid the application limitations of spectral cameras in open scenarios and obtain high spatial resolution multispectral images. However, the current spectral reconstruction algorithms are sensitive to the exposure variant of the test images. That is, when the exposure of the test image is different from that of the training image, the reconstructed spectral curve of the test object will deviate from the real spectral to varying degrees, which will lead to the spectral data of the target object being accurately reconstructed. This article proposes an optimized method for spectral reconstruction based on data augmentation and attention mechanisms using the current deep learning-based spectral reconstruction framework. The proposed method is exposure invariant and will adapt to the open environment in which the light is easily changed and the illumination is non-uniform. Thus, the robustness and reconstruction accuracy of the spectral reconstruction model in practical applications are improved. The experiments show that the proposed method can accurately reconstruct the shape of the spectral reflectance curve of the test object under different test exposure levels. And the spectral reconstruction error of our method at different exposure levels is significantly lower than that of the existing methods, which verifies the proposed method's effectiveness and superiority.

Message in hand: the application of CRISPRi, RNAi, and LncRNA in adenocarcinoma.

Gene editing interference technology has been flourishing for more than 30 years. It has always been a common means to interfere with the expression of particular genes. Today it has shown a broad application prospect in clinical treatment, especially in adenocarcinoma treatment. In just a few years, the CRISPRi technology has attracted much z attention with its precise targeting and convenient operability significantly promoted the transformation from bench to bedside, and won the Nobel Prize in Chemistry 2020. In recent years, the importance of non-coding RNA has led LncRNA research to the center. At the same time, it also recalls the surprises obtained in laboratory and clinic research by RNAi technologies such as microRNA, siRNA, and shRNA at the beginning of the century. Therefore, this article focuses on CRISPRi, RNAi, and LncRNA to review their gene interference mechanisms currently expected to be translational research. Their applications and differences in adenocarcinoma research will also be described powerfully. It will provide a helpful reference for scientists to understand better and apply several RNA interference technologies.

The brassinosteroid biosynthesis gene TaD11-2A controls grain size and its elite haplotype improves wheat grain yields.

KEY MESSAGE: TaD11-2A affects grain size and root length and its natural variations are associated with significant differences in yield-related traits in wheat. Brassinosteroids (BRs) control many important agronomic traits and therefore the manipulation of BR components could improve crop productivity and performance. However, the potential effects of BR-related genes on yield-related traits and stress tolerance in wheat (Triticum aestivum L.) remain poorly understood. Here, we identified TaD11 genes in wheat (rice D11 orthologs) that encoded enzymes involved in BR biosynthesis. TaD11 genes were highly expressed in roots (Zadoks scale: Z11) and grains (Z75), while expression was significantly suppressed by exogenous BR (24-epiBL). Ectopic expression of TaD11-2A rescued the abnormal panicle structure and plant height (PH) of the clustered primary branch 1 (cpb1) mutant, and also increased endogenous BR levels, resulting in improved grain yields and grain quality in rice. The tad11-2a-1 mutant displayed dwarfism, smaller grains, sensitivity to 24-epiBL, and reduced endogenous BR contents. Natural variations in TaD11-2A were associated with significant differences in yield-related traits, including PH, grain width, 1000-grain weight, and grain yield per plant, and its favorable haplotype, TaD11-2A-HapI was subjected to positive selection during wheat breeding. Additionally, TaD11-2A influenced root length and salt tolerance in rice and wheat at seedling stages. These results indicated the important role of BR TaD11 biosynthetic genes in controlling grain size and root length, and also highlighted their potential in the molecular biological analysis of wheat.

Recent advances in the contribution of circRNAs to cisplatin chemotherapy resistance in cancers.

Worldwide, cancer is a serious threat to the health of citizens of every country, with the incidence and mortality increasing year by year. Cisplatin is the first-line anticancer drug commonly used in clinics and is widely used for the treatment of solid tumors including lung, gastric, liver, bladder, and ovarian cancer. Although cisplatin-based chemotherapy has a high clinical response efficacy, patients will inevitably develop drug resistance after repeated use, leading to severe restrictions of its application. Circular RNAs (circRNAs) are a promising class of non-coding RNAs capable of promoting or suppressing cancer via functioning as miRNAs sponges. Recently, an increasing amount of evidence shows that circRNAs are closely related to the cisplatin resistance of cancers. Therefore, standing at the perspective of the cisplatin chemotherapy resistance, this paper reviews the research progress of circRNAs related to cisplatin resistance of various cancers.

Automated Facial Recognition for Noonan Syndrome Using Novel Deep Convolutional Neural Network With Additive Angular Margin Loss.

BACKGROUND: Noonan syndrome (NS), a genetically heterogeneous disorder, presents with hypertelorism, ptosis, dysplastic pulmonary valve stenosis, hypertrophic cardiomyopathy, and small stature. Early detection and assessment of NS are crucial to formulating an individualized treatment protocol. However, the diagnostic rate of pediatricians and pediatric cardiologists is limited. To overcome this challenge, we propose an automated facial recognition model to identify NS using a novel deep convolutional neural network (DCNN) with a loss function called additive angular margin loss (ArcFace). METHODS: The proposed automated facial recognition models were trained on dataset that included 127 NS patients, 163 healthy children, and 130 children with several other dysmorphic syndromes. The photo dataset contained only one frontal face image from each participant. A novel DCNN framework with ArcFace loss function (DCNN-Arcface model) was constructed. Two traditional machine learning models and a DCNN model with cross-entropy loss function (DCNN-CE model) were also constructed. Transfer learning and data augmentation were applied in the training process. The identification performance of facial recognition models was assessed by five-fold cross-validation. Comparison of the DCNN-Arcface model to two traditional machine learning models, the DCNN-CE model, and six physicians were performed. RESULTS: At distinguishing NS patients from healthy children, the DCNN-Arcface model achieved an accuracy of 0.9201 ± 0.0138 and an area under the receiver operator characteristic curve (AUC) of 0.9797 ± 0.0055. At distinguishing NS patients from children with several other genetic syndromes, it achieved an accuracy of 0.8171 ± 0.0074 and an AUC of 0.9274 ± 0.0062. In both cases, the DCNN-Arcface model outperformed the two traditional machine learning models, the DCNN-CE model, and six physicians. CONCLUSION: This study shows that the proposed DCNN-Arcface model is a promising way to screen NS patients and can improve the NS diagnosis rate.