Identification of potential drug targets for insomnia by Mendelian randomization analysis based on plasma proteomics.

Introduction: Insomnia, a common clinical disorder, significantly impacts the physical and mental well-being of patients. Currently, available hypnotic medications are unsatisfactory due to adverse reactions and dependency, necessitating the identification of new drug targets for the treatment of insomnia. Methods: In this study, we utilized 734 plasma proteins as genetic instruments obtained from genome-wide association studies to conduct a Mendelian randomization analysis, with insomnia as the outcome variable, to identify potential drug targets for insomnia. Additionally, we validated our results externally using other datasets. Sensitivity analyses entailed reverse Mendelian randomization analysis, Bayesian co-localization analysis, and phenotype scanning. Furthermore, we constructed a protein-protein interaction network to elucidate potential correlations between the identified proteins and existing targets. Results: Mendelian randomization analysis indicated that elevated levels of TGFBI (OR = 1.01; 95% CI, 1.01-1.02) and PAM ((OR = 1.01; 95% CI, 1.01-1.02) in plasma are associated with an increased risk of insomnia, with external validation supporting these findings. Moreover, there was no evidence of reverse causality for these two proteins. Co-localization analysis confirmed that PAM (coloc.abf-PPH4 = 0.823) shared the same variant with insomnia, further substantiating its potential role as a therapeutic target. There are interactive relationships between the potential proteins and existing targets of insomnia. Conclusion: Overall, our findings suggested that elevated plasma levels of TGFBI and PAM are connected with an increased risk of insomnia and might be promising therapeutic targets, particularly PAM. However, further exploration is necessary to fully understand the underlying mechanisms involved.

NTSM: a non-salient target segmentation model for oral mucosal diseases.

BACKGROUND: Oral mucosal diseases are similar to the surrounding normal tissues, i.e., their many non-salient features, which poses a challenge for accurate segmentation lesions. Additionally, high-precision large models generate too many parameters, which puts pressure on storage and makes it difficult to deploy on portable devices. METHODS: To address these issues, we design a non-salient target segmentation model (NTSM) to improve segmentation performance while reducing the number of parameters. The NTSM includes a difference association (DA) module and multiple feature hierarchy pyramid attention (FHPA) modules. The DA module enhances feature differences at different levels to learn local context information and extend the segmentation mask to potentially similar areas. It also learns logical semantic relationship information through different receptive fields to determine the actual lesions and further elevates the segmentation performance of non-salient lesions. The FHPA module extracts pathological information from different views by performing the hadamard product attention (HPA) operation on input features, which reduces the number of parameters. RESULTS: The experimental results on the oral mucosal diseases (OMD) dataset and international skin imaging collaboration (ISIC) dataset demonstrate that our model outperforms existing state-of-the-art methods. Compared with the nnU-Net backbone, our model has 43.20% fewer parameters while still achieving a 3.14% increase in the Dice score. CONCLUSIONS: Our model has high segmentation accuracy on non-salient areas of oral mucosal diseases and can effectively reduce resource consumption.

All-Inorganic Halide Perovskites Boost High-Ranged Figure-of-Merit in Bi0.4Sb1.6Te3 for Thermoelectric Cooling and Low-Grade Heat Recovery.

The all-inorganic halide perovskite CsPbX3 (X = Cl, Br, or I) offers various advantages, such as tunable electronic structure and high carrier mobility. However, its potential application in thermoelectric materials remains underexplored. In this study, we propose a simple yet effective method to synthesize a CsPbX3/Bi0.4Sb1.6Te3 (BST) nanocomposite by sintering a uniformly mixed raw powder. The intrinsic excitation of the BST system is suppressed by exploiting the rich phase structure and tunable electrical transport properties of CsPbX3, and the thermoelectric properties were synergistically optimized. Notably, for CsPbI3, its phase-transition-induced dislocation arrays together with low group velocities drastically reduce thermal conductivity. As a result, the composite achieves an ultrahigh average figure-of-merit (ZT) of 1.4 from 298 to 523 K. The two-pair TE module demonstrates a superior conversion efficiency of 7.3%. This study expands the potential applications of inorganic halide perovskites, into thermoelectrics.

Transfer learning enables identification of multiple types of RNA modifications using nanopore direct RNA sequencing.

Nanopore direct RNA sequencing (DRS) has emerged as a powerful tool for RNA modification identification. However, concurrently detecting multiple types of modifications in a single DRS sample remains a challenge. Here, we develop TandemMod, a transferable deep learning framework capable of detecting multiple types of RNA modifications in single DRS data. To train high-performance TandemMod models, we generate in vitro epitranscriptome datasets from cDNA libraries, containing thousands of transcripts labeled with various types of RNA modifications. We validate the performance of TandemMod on both in vitro transcripts and in vivo human cell lines, confirming its high accuracy for profiling m6A and m5C modification sites. Furthermore, we perform transfer learning for identifying other modifications such as m7G, Ψ, and inosine, significantly reducing training data size and running time without compromising performance. Finally, we apply TandemMod to identify 3 types of RNA modifications in rice grown in different environments, demonstrating its applicability across species and conditions. In summary, we provide a resource with ground-truth labels that can serve as benchmark datasets for nanopore-based modification identification methods, and TandemMod for identifying diverse RNA modifications using a single DRS sample.

Carbon-nitrogen transmutation in polycyclic arenol skeletons to access N-heteroarenes.

Developing skeletal editing tools is not a trivial task, and realizing the corresponding single-atom transmutation in a ring system without altering the ring size is even more challenging. Here, we introduce a skeletal editing strategy that enables polycyclic arenols, a highly prevalent motif in bioactive molecules, to be readily converted into N-heteroarenes through carbon-nitrogen transmutation. The reaction features selective nitrogen insertion into the C-C bond of the arenol frameworks by azidative dearomatization and aryl migration, followed by ring-opening, and ring-closing (ANRORC) to achieve carbon-to-nitrogen transmutation in the aromatic framework of the arenol. Using widely available arenols as N-heteroarene precursors, this alternative approach allows the streamlined assembly of complex polycyclic heteroaromatics with broad functional group tolerance. Finally, pertinent transformations of the products, including synthesis complex biheteroarene skeletons, were conducted and exhibited significant potential in materials chemistry.

Decreased brain iron deposition based on quantitative susceptibility mapping correlates with reduced neurodevelopmental status in children with autism spectrum disorder.

To investigate potential correlations between the susceptibility values of certain brain regions and the severity of disease or neurodevelopmental status in children with autism spectrum disorder (ASD), 18 ASD children and 15 healthy controls (HCs) were recruited. The neurodevelopmental status was assessed by the Gesell Developmental Schedules (GDS) and the severity of the disease was evaluated by the Autism Behavior Checklist (ABC). Eleven brain regions were selected as regions of interest and the susceptibility values were measured by quantitative susceptibility mapping. To evaluate the diagnostic capacity of susceptibility values in distinguishing ASD and HC, the receiver operating characteristic (ROC) curve was computed. Pearson and Spearman partial correlation analysis were used to depict the correlations between the susceptibility values, the ABC scores, and the GDS scores in the ASD group. ROC curves showed that the susceptibility values of the left and right frontal white matter had a larger area under the curve in the ASD group. The susceptibility value of the right globus pallidus was positively correlated with the GDS-fine motor scale score. These findings indicated that the susceptibility value of the right globus pallidus might be a viable imaging biomarker for evaluating the neurodevelopmental status of ASD children.

Plexin D1 emerges as a novel target in the development of neural lineage plasticity in treatment-resistant prostate cancer.

Treatment-induced neuroendocrine prostate cancer (t-NEPC) often arises from adenocarcinoma via lineage plasticity in response to androgen receptor signaling inhibitors, such as enzalutamide. However, the specific regulators and targets involved in the transition to NEPC are not well understood. Plexin D1 (PLXND1) is a cellular receptor of the semaphorin (SEMA) family that plays important roles in modulating the cytoskeleton and cell adhesion. Here, we found that PLXND1 is highly expressed and positively correlated with neuroendocrine markers in patients with NEPC. High PLXND1 expression is associated with poorer prognosis in prostate cancer patients. Additionally, PLXND1 was upregulated and negatively regulated by androgen receptor signaling in enzalutamide-resistant cells. Knockdown or knockout of PLXND1 inhibit neural lineage pathways, suppressing NEPC cell proliferation, PDX tumor organoid viability, and xenograft tumor growth. Mechanistically, the chaperone protein HSP70 regulates PLXND1 protein stability through degradation, and inhibition of HSP70 decreases PLXND1 expression and NEPC organoid growth. In summary, our findings suggest that PLXND1 could be a new therapeutic target and molecular indicator for NEPC.

Visible-Light-Induced Alkoxypyridylation of Alkenes Using N-Alkoxypyridinium Salts as Bifunctional Reagents.

Considering the ubiquitous presence of pyridine moieties in pharmaceutical compounds, it holds immense value to develop practical and straightforward methodologies for accessing heterocyclic aromatic hydrocarbons. In recent years, N-alkoxypyridinium salts have emerged as convenient radical precursors, enabling the generation of the corresponding alkoxy radicals and pyridine through single-electron transfer. Herein, we present the first report on visible-light-mediated intermolecular alkoxypyridylation of alkenes employing N-alkoxylpyridinium salts as bifunctional reagents with an exceptionally low catalyst loading (0.5 mol %).

Effects of long-term closed and socially isolating spaceflight analog environment on default mode network connectivity as indicated by fMRI.

Long-term manned spaceflight and extraterrestrial planet settlement become the focus of space powers. However, the potential influence of closed and socially isolating spaceflight on the brain function remains unclear. A 180-day controlled ecological life support system integrated experiment was conducted, establishing a spaceflight analog environment to explore the effect of long-term socially isolating living. Three crewmembers were enrolled and underwent resting-state fMRI scanning before and after the experiment. We performed both seed-based and network-based analyses to investigate the functional connectivity (FC) changes of the default mode network (DMN), considering its key role in multiple higher-order cognitive functions. Compared with normal controls, the leader of crewmembers exhibited significantly reduced within-DMN and between-DMN FC after the experiment, while two others exhibited opposite trends. Moreover, individual differences of FC changes were further supported by evidence from behavioral analyses. The findings may shed new light on the development of psychological protection for space exploration.

The anatomical structure of sex differences in trust propensity: A voxel-based morphometry study.

Trust is a key component of human relationships. Sex differences in trust behavior have been elucidated by parental investment theory and social role theory, attributing men's higher trust propensity to their increased engagement in physically and socially risky activities aimed at securing additional resources. Although sex differences in trust behavior exist and the neuropsychological signatures of trust are known, the underlying anatomical structure of sex differences is still unexplored. Our study aimed to investigate the anatomical structure of sex differences in trust behavior toward strangers (i.e., trust propensity, TP) by employing voxel-based morphometry (VBM) in a sample of healthy young adults. We collected behavioral data for TP as measured with participants in the role of trustors completing the one-shot trust game (TG) with anonymous partners as trustees. We conducted primary region of interest (ROI) and exploratory whole-brain (WB) VBM analyses of high-resolution structural images to test for the association between TP and regional gray matter volume (GMV) associated with sex differences. Confirming previous studies, our behavioral results demonstrated that men trusted more than women during the one-shot TG. Our WB analysis showed a greater GMV related to TP in men than women in the precuneus (PreC), whereas our ROI analysis in regions of the default-mode network (dorsomedial prefrontal cortex [dmPFC], PreC, superior temporal gyrus) to simulate the partner's trustworthiness, central-executive network (ventrolateral PFC) to implement a calculus-based trust strategy, and action-perception network (precentral gyrus) to performance cost-benefit calculations, as proposed by a neuropsychoeconomic model of trust. Our findings advance the neuropsychological understanding of sex differences in TP, which has implications for interpersonal partnerships, financial transactions, and societal engagements.

The EIN3 transcription factor GmEIL1 improves soybean resistance to Phytophthora sojae.

Phytophthora root and stem rot of soybean (Glycine max), caused by the oomycete Phytophthora sojae, is an extremely destructive disease worldwide. In this study, we identified GmEIL1, which encodes an ethylene-insensitive3 (EIN3) transcription factor. GmEIL1 was significantly induced following P. sojae infection of soybean plants. Compared to wild-type soybean plants, transgenic soybean plants overexpressing GmEIL1 showed enhanced resistance to P. sojae and GmEIL1-silenced RNA-interference lines showed more severe symptoms when infected with P. sojae. We screened for target genes of GmEIL1 and confirmed that GmEIL1 bound directly to the GmERF113 promoter and regulated GmERF113 expression. Moreover, GmEIL1 positively regulated the expression of the pathogenesis-related gene GmPR1. The GmEIL1-regulated defence response to P. sojae involved both ethylene biosynthesis and the ethylene signalling pathway. These findings suggest that the GmEIL1-GmERF113 module plays an important role in P. sojae resistance via the ethylene signalling pathway.

Proliferation-driven mechanical compression induces signalling centre formation during mammalian organ development.

Localized sources of morphogens, called signalling centres, play a fundamental role in coordinating tissue growth and cell fate specification during organogenesis. However, how these signalling centres are established in tissues during embryonic development is still unclear. Here we show that the main signalling centre orchestrating development of rodent incisors, the enamel knot (EK), is specified by a cell proliferation-driven buildup in compressive stresses (mechanical pressure) in the tissue. Direct mechanical measurements indicate that the stresses generated by cell proliferation are resisted by the surrounding tissue, creating a circular pattern of mechanical anisotropy with a region of high compressive stress at its centre that becomes the EK. Pharmacological inhibition of proliferation reduces stresses and suppresses EK formation, and application of external pressure in proliferation-inhibited conditions rescues the formation of the EK. Mechanical information is relayed intracellularly through YAP protein localization, which is cytoplasmic in the region of compressive stress that establishes the EK and nuclear in the stretched anisotropic cells that resist the pressure buildup around the EK. Together, our data identify a new role for proliferation-driven mechanical compression in the specification of a model signalling centre during mammalian organ development.

Clinical translation of a novel FAPI dimer [68Ga]Ga-LNC1013.

Fibroblast activation protein (FAP) has emerged as a highly promising target for cancer diagnostic imaging and targeted radionuclide therapy. To exploit the therapeutic potential of suitably radiolabeled FAP inhibitors (FAPIs), this study presents the design and synthesis of a series of FAPI dimers to increase tumor uptake and retention. Preclinical evaluation and a pilot clinical PET imaging study were conducted to screen the lead compound with the potential for radionuclide therapy. METHODS: Three new FAPI dimers were synthesized by linking two quinoline-based FAPIs with different spacers. The in vitro binding affinity and preclinical small animal PET imaging of the compounds were compared with their monomeric counterparts, FAPI-04 and FAPI-46. The lead compound, [68Ga]Ga -LNC1013, was then evaluated in a pilot clinical PET imaging study involving seven patients with gastrointestinal cancer. RESULTS: The three newly synthesized FAPI homodimers had high binding affinity and specificity in vitro and in vivo. Small animal PET imaging and biodistribution studies showed that [68Ga]Ga-LNC1013 had persistent tumor retention for at least 4 h, also higher uptake than the other two dimers and the monomer counterparts, making it the lead compound to enter clinical investigation. In the pilot clinical PET imaging study, seven patients were enrolled. The effective dose of [68Ga]Ga-LNC1013 was 8.24E-03 mSv/MBq. The human biodistribution of [68Ga]Ga-LNC1013 demonstrated prominent tumor uptake and good tumor-to-background contrast. [68Ga]Ga-LNC1013 PET imaging showed potential in capturing primary and metastatic lesions and outperforming 18F-FDG PET in detecting pancreatic and esophageal cancers. The SUVmax for lesions with [68Ga]Ga-FAPI-46 decreased over time, whereas [68Ga]Ga-LNC1013 exhibited persistently high tumor uptake from 1 to 4 h post-injection. CONCLUSION: Dimerization is an effective strategy to produce FAPI derivatives with favorable tumor uptake, long tumor retention, and imaging contrast over its monomeric counterpart. We demonstrated that [68Ga]Ga-LNC1013, the lead compound without any piperazine moiety, had superior diagnostic potential over [68Ga]Ga-FAPI-46 and 18F-FDG, suggesting the future potential of LNC1013 for radioligand therapy of FAP-positive cancers.

The MYB Transcription Factor GmMYB78 Negatively Regulates Phytophthora sojae Resistance in Soybean.

Phytophthora root rot is a devastating disease of soybean caused by Phytophthora sojae. However, the resistance mechanism is not yet clear. Our previous studies have shown that GmAP2 enhances sensitivity to P. sojae in soybean, and GmMYB78 is downregulated in the transcriptome analysis of GmAP2-overexpressing transgenic hairy roots. Here, GmMYB78 was significantly induced by P. sojae in susceptible soybean, and the overexpressing of GmMYB78 enhanced sensitivity to the pathogen, while silencing GmMYB78 enhances resistance to P. sojae, indicating that GmMYB78 is a negative regulator of P. sojae. Moreover, the jasmonic acid (JA) content and JA synthesis gene GmAOS1 was highly upregulated in GmMYB78-silencing roots and highly downregulated in overexpressing ones, suggesting that GmMYB78 could respond to P. sojae through the JA signaling pathway. Furthermore, the expression of several pathogenesis-related genes was significantly lower in GmMYB78-overexpressing roots and higher in GmMYB78-silencing ones. Additionally, we screened and identified the upstream regulator GmbHLH122 and downstream target gene GmbZIP25 of GmMYB78. GmbHLH122 was highly induced by P. sojae and could inhibit GmMYB78 expression in resistant soybean, and GmMYB78 was highly expressed to activate downstream target gene GmbZIP25 transcription in susceptible soybean. In conclusion, our data reveal that GmMYB78 triggers soybean sensitivity to P. sojae by inhibiting the JA signaling pathway and the expression of pathogenesis-related genes or through the effects of the GmbHLH122-GmMYB78-GmbZIP25 cascade pathway.

Elabela ameliorates neuronal pyroptosis and mitochondrial fission via APJ/ZBP1 signaling in ischemic stroke.

Pyroptosis signifies a significant form of programmed neuronal demise subsequent to ischemic stroke. In our prior investigations, we demonstrated that the Elabela (ELA)-Apelin receptor (APJ) axis alleviated neuronal death by improving collateral circulation and mitigating ferroptosis in a murine model of middle cerebral artery occlusion (MCAO). However, the connection between ELA and neuronal pyroptosis remains further elucidation. Here, we observed an upregulation of ELA and APJ expression in both murine brain specimens and cultured HT-22 hippocampal neurons exposed to experimental ischemic stroke. ELA administration markedly diminished the infarct size in comparison to controls. ELA treatment ameliorated neurological deficits and anxiety-like symptoms in mice with stroke, concurrently inhibiting pyroptosis and mitochondria fission in neurons. Conversely, ELA knockdown yielded the opposite effects. Utilizing RNA-sequencing analysis, we identified a candidate for pyroptosis priming, Z-DNA-binding protein 1 (ZBP1), which was suppressed in ELA-treated HT-22 neurons during oxygen-glucose deprivation/reperfusion (OGD/R). Subsequent co-immunoprecipitation analyses demonstrated the binding between APJ and ZBP1. Specifically, APJ suppressed ZBP1 to inhibit NLRP3 inflammasome activation and dynamin-related protein 1-mediated mitochondrial fission in neurons. In summary, our findings suggest that ELA functions as a stroke-induced signal limiting neuronal pyroptosis and mitochondrial fission via APJ/ZBP1 signaling, thereby underscoring ELA as a potential therapeutic target for ischemic stroke treatment.

Bismuth Telluride Supported Sub-1 nm Polyoxometalate Cluster for High-Efficiency Thermoelectric Energy Conversion.

Size plays a crucial role in chemistry and material science. Subnanometer polyoxometalate (POM) clusters have gained attention in various fields, but their use in thermoelectrics is still limited. To address this issue, we propose the POM clusters as an effective second phase to enhance the thermoelectric properties of Bi0.4Sb1.6Te3. Thanks to their subnanometer size, POM clusters improve electrical transport behavior through the superposition of atomic orbitals and the interfacial scattering effect. Furthermore, their ultrasmall size strongly reduces thermal conductivity. Consequently, the introduction of a mere 0.1 mol % of POM into the Bi0.4Sb1.6Te3 matrix realizes a state-of-the-art zT value of 1.46 at 348 K, a 45% enhancement over Bi0.4Sb1.6Te3 (1.01), along with a maximum thermoelectric-conversion efficiency of the integrated module of 6.0%. The enhancement of carrier mobility and the suppression of thermal conduction achieved by introducing the subnanometer clusters hold promise for various applications, such as electronic devices and thermal management.

Camellia sinensis polysaccharide attenuates inflammatory responses via the ROS-mediated pathway by endocytosis.

Polysaccharide CSTPs extracted from Camellia sinensis tea-leaves possessed unique against oxidative damage by scavenging ROS. Herein, acid tea polysaccharide CSTPs-2 with tightly packed molecular structure was isolated, purified and characterized in this research. Furthermore, the effects of CSTPs-2 on ROS-involved inflammatory responses and its underlying mechanisms were investigated. The results suggest that CSTPs-2 dramatically reduced the inflammatory cytokines overexpression and LPS-stimulated cell damage. CSTPs-2 could trigger the dephosphorylation of downstream AKT/MAPK/NF-κB signaling proteins and inhibit nuclear transfer of p-NF-κB to regulate the synthesis and release of inflammatory mediators in LPS-stimulated cells by ROS scavenging. Importantly, the impact of CSTPs-2 in downregulating pro-inflammatory cytokines and mitigating ROS overproduction is associated with clathrin- or caveolae-mediated endocytosis uptake mechanisms, rather than TLR-4 receptor-mediated endocytosis. This study presents a novel perspective for investigating the cellular uptake mechanism of polysaccharides in the context of anti-inflammatory mechanisms.

Sulfation of chondroitin and bile acids converges to antagonize Wnt/ß-catenin signaling and inhibit APC deficiency-induced gut tumorigenesis.

Sulfation is a crucial and prevalent conjugation reaction involved in cellular processes and mammalian physiology. 3'-Phosphoadenosine 5'-phosphosulfate (PAPS) synthase 2 (PAPSS2) is the primary enzyme to generate the universal sulfonate donor PAPS. The involvement of PAPSS2-mediated sulfation in adenomatous polyposis coli (APC) mutation-promoted colonic carcinogenesis has not been reported. Here, we showed that the expression of PAPSS2 was decreased in human colon tumors along with cancer stages, and the lower expression of PAPSS2 was correlated with poor prognosis in advanced colon cancer. Gut epithelial-specific heterozygous Apc deficient and Papss2-knockout (ApcΔgut-HetPapss2Δgut) mice were created, and the phenotypes were compared to the spontaneous intestinal tumorigenesis of ApcΔgut-Het mice. ApcΔgut-HetPapss2Δgut mice were more sensitive to gut tumorigenesis, which was mechanistically accounted for by the activation of Wnt/ß-catenin signaling pathway due to the suppression of chondroitin sulfation and inhibition of the farnesoid X receptor (FXR)-transducin-like enhancer of split 3 (TLE3) gene regulatory axis. Chondroitin sulfate supplementation in ApcΔgut-HetPapss2Δgut mice alleviated intestinal tumorigenesis. In summary, we have uncovered the protective role of PAPSS2-mediated chondroitin sulfation and bile acids-FXR-TLE3 activation in the prevention of gut carcinogenesis via the antagonization of Wnt/ß-catenin signaling. Chondroitin sulfate may be explored as a therapeutic agent for Papss2 deficiency-associated colonic carcinogenesis.

Classification of multi-feature fusion ultrasound images of breast tumor within category 4 using convolutional neural networks.

BACKGROUND: Breast tumor is a fatal threat to the health of women. Ultrasound (US) is a common and economical method for the diagnosis of breast cancer. Breast imaging reporting and data system (BI-RADS) category 4 has the highest false-positive value of about 30% among five categories. The classification task in BI-RADS category 4 is challenging and has not been fully studied. PURPOSE: This work aimed to use convolutional neural networks (CNNs) for breast tumor classification using B-mode images in category 4 to overcome the dependence on operator and artifacts. Additionally, this work intends to take full advantage of morphological and textural features in breast tumor US images to improve classification accuracy. METHODS: First, original US images coming directly from the hospital were cropped and resized. In 1385 B-mode US BI-RADS category 4 images, the biopsy eliminated 503 samples of benign tumor and left 882 of malignant. Then, K-means clustering algorithm and entropy of sliding windows of US images were conducted. Considering the diversity of different characteristic information of malignant and benign represented by original B-mode images, K-means clustering images and entropy images, they are fused in a three-channel form multi-feature fusion images dataset. The training, validation, and test sets are 969, 277, and 139. With transfer learning, 11 CNN models including DenseNet and ResNet were investigated. Finally, by comparing accuracy, precision, recall, F1-score, and area under curve (AUC) of the results, models which had better performance were selected. The normality of data was assessed by Shapiro-Wilk test. DeLong test and independent t-test were used to evaluate the significant difference of AUC and other values. False discovery rate was utilized to ultimately evaluate the advantages of CNN with highest evaluation metrics. In addition, the study of anti-log compression was conducted but no improvement has shown in CNNs classification results. RESULTS: With multi-feature fusion images, DenseNet121 has highest accuracy of 80.22 ± 1.45% compared to other CNNs, precision of 77.97 ± 2.89% and AUC of 0.82 ± 0.01. Multi-feature fusion improved accuracy of DenseNet121 by 1.87% from classification of original B-mode images (p < 0.05). CONCLUSION: The CNNs with multi-feature fusion show a good potential of reducing the false-positive rate within category 4. The work illustrated that CNNs and fusion images have the potential to reduce false-positive rate in breast tumor within US BI-RADS category 4, and make the diagnosis of category 4 breast tumors to be more accurate and precise.