NeighborNet: Learning Intra- and Inter-Image Pixel Neighbor Representation for Breast Lesion Segmentation.

Breast lesion segmentation from ultrasound images is essential in computer-aided breast cancer diagnosis. To alleviate the problems of blurry lesion boundaries and irregular morphologies, common practices combine CNN and attention to integrate global and local information. However, previous methods use two independent modules to extract global and local features separately, such feature-wise inflexible integration ignores the semantic gap between them, resulting in representation redundancy/insufficiency and undesirable restrictions in clinic practices. Moreover, medical images are highly similar to each other due to the imaging methods and human tissues, but the captured global information by transformer-based methods in the medical domain is limited within images, the semantic relations and common knowledge across images are largely ignored. To alleviate the above problems, in the neighbor view, this paper develops a pixel neighbor representation learning method (NeighborNet) to flexibly integrate global and local context within and across images for lesion morphology and boundary modeling. Concretely, we design two neighbor layers to investigate two properties (i.e., number and distribution) of neighbors. The neighbor number for each pixel is not fixed but determined by itself. The neighbor distribution is extended from one image to all images in the datasets. With the two properties, for each pixel at each feature level, the proposed NeighborNet can evolve into the transformer or degenerate into the CNN for adaptive context representation learning to cope with the irregular lesion morphologies and blurry boundaries. The state-of-the-art performances on three ultrasound datasets prove the effectiveness of the proposed NeighborNet. The code is available at: https://github.com/fjcaoww/NeighborNet.

STC2 Inhibits Hepatic Lipid Synthesis and Correlates with Intramuscular Fatty Acid Composition, Body Weight and Carcass Traits in Chickens.

Stanniocalcin 2 (STC2) is a secreted glycoprotein involved in multiple biological processes. To systemically study the biological role of STC2 in chickens, phylogenetic tree analysis and conservation analysis were conducted. Association analysis between variations in the STC2 gene and the economic traits of Gushi-Anka F2 was conducted. The tissue expression patterns of STC2 expression in different chicken tissues and liver at different stages were detected. The biological role of STC2 in chicken liver was investigated through overexpression and interfering methods in the LMH cell line. Correlation analyses between STC2 expression and lipid components were conducted. (1) The phylogenetic tree displayed that chicken STC2 is most closely related with Japanese quail and most distantly related with Xenopus tropicalis. STC2 has the same identical conserved motifs as other species. (2) rs9949205 (T > C) found in STC2 intron was highly significantly correlated with chicken body weight at 0, 2, 4, 6, 8, 10 and 12 weeks (p < 0.01). Extremely significant correlations of rs9949205 with semi-evisceration weight (SEW), evisceration weight (EW), breast muscle weight (BMW), leg muscle weight (LMW), liver weight and abdominal fat weight (AFW) were revealed (p < 0.01). Significant associations between rs9949205 and abdominal fat percentage, liver weight rate, breast muscle weight rate and leg muscle weight rate were also found (p < 0.05). Individuals with TT or TC genotypes had significantly lower abdominal fat percentage and liver weight rate compared to those with the CC genotype, while their body weight and other carcass traits were higher. (3) STC2 showed a high expression level in chicken liver tissue, which significantly increased with the progression of age (p < 0.05). STC2 was observed to inhibit the content of lipid droplets, triglycerides (TG) and cholesterol (TC), as well the expression level of genes related to lipid metabolism in LMH cells. (4) Correlation analysis showed that the STC2 gene was significantly correlated with 176 lipids in the breast muscle (p < 0.05) and mainly enriched in omega-3 and omega-6 unsaturated fatty acids. In conclusion, the STC2 gene in chicken might potentially play a crucial role in chicken growth and development, as well as liver lipid metabolism and muscle lipid deposition. This study provides a scientific foundation for further investigation into the regulatory mechanism of the STC2 gene on lipid metabolism and deposition in chicken liver.

miR-19b-3p regulated by estrogen controls lipid synthesis through targeting MSMO1 and ELOVL5 in LMH cells.

miR-19b-3p is reported to undertake various biological role, while its function and action mechanism in chicken hepatic lipid metabolism is unclear. Conservation analysis and tissue expression pattern of miR-19b-3p and its target gene were evaluated, respectively. Dual luciferase reporter system and Western blot technologies were adopted to validate miR-19b-3p target gene. Overexpression and knockdown assays were done to explore the biological functions of miR-19b-3p and target gene in Leghorn Male Hepatoma cell line (LMH). Regulatory approaches of estrogen on miR-19b-3p and target gene expressions are analyzed through site-directed mutation combined with estrogen receptors antagonist treatment assays. The results showed that chicken miR-19b-3p mature sequences are highly conserved among Capra hircus, Columba livia, Rattus norvegicus, Mus musculus, Cricetulus griseus, Danio rerio, Danio novaehollandiae, Orycodylus porosus, Crocodylus porosus, Gadus morhua, and widely expressed in lung, ovary, spleen, duodenum, kidney, heart, liver, leg muscle, and pectoral muscle tissues. miR-19b-3p could significantly increase intracellular triglyceride (TG) content and decrease intracellular cholesterol (TC) content via targeting methylsterol monooxygenase 1 (MSMO1) and elongase of very long chain fatty acids 5 (ELOVL5), which are highly conserved among species, in both mRNA and protein levels. Estrogen could inhibit miR-19b-3p expression, but directly promoted MSMO1 transcription via estrogen receptor α (ERα) and indirectly regulated ELOVL5 expression at the transcription level. Meanwhile, estrogen could also upregulate MSMO1 and ELOVL5 expression through inhibiting miR-19b-3p expression at the post-transcription level. Taken together, these results highlight the role and regulatory mechanism of miR-19b-3p in hepatic lipid metabolism in chicken, and might produce useful comparative information for human obesity studies and biomedical research.

Genomic Insights into Molecular Regulation Mechanisms of Intramuscular Fat Deposition in Chicken.

Intramuscular fat (IMF) plays an important role in the tenderness, water-holding capacity, and flavor of chicken meat, which directly affect meat quality. In recent years, regulatory mechanisms underlying IMF deposition and the development of effective molecular markers have been hot topics in poultry genetic breeding. Therefore, this review focuses on the current understanding of regulatory mechanisms underlying IMF deposition in chickens, which were identified by multiple genomic approaches, including genome-wide association studies, whole transcriptome sequencing, proteome sequencing, single-cell RNA sequencing (scRNA-seq), high-throughput chromosome conformation capture (HiC), DNA methylation sequencing, and m6A methylation sequencing. This review comprehensively and systematically describes genetic and epigenetic factors associated with IMF deposition, which provides a fundamental resource for biomarkers of IMF deposition and provides promising applications for genetic improvement of meat quality in chicken.

The arginine methyltransferase Prmt1 coordinates the germline arginine methylome essential for spermatogonial homeostasis and male fertility.

Arginine methylation, catalyzed by the protein arginine methyltransferases (PRMTs), is a common post-translational protein modification (PTM) that is engaged in a plethora of biological events. However, little is known about how the methylarginine-directed signaling functions in germline development. In this study, we discover that Prmt1 is predominantly distributed in the nuclei of spermatogonia but weakly in the spermatocytes throughout mouse spermatogenesis. By exploiting a combination of three Cre-mediated Prmt1 knockout mouse lines, we unravel that Prmt1 is essential for spermatogonial establishment and maintenance, and that Prmt1-catalyzed asymmetric methylarginine coordinates inherent transcriptional homeostasis within spermatogonial cells. In conjunction with high-throughput CUT&Tag profiling and modified mini-bulk Smart-seq2 analyses, we unveil that the Prmt1-deposited H4R3me2a mark is permissively enriched at promoter and exon/intron regions, and sculpts a distinctive transcriptomic landscape as well as the alternative splicing pattern, in the mouse spermatogonia. Collectively, our study provides the genetic and mechanistic evidence that connects the Prmt1-deposited methylarginine signaling to the establishment and maintenance of a high-fidelity transcriptomic identity in orchestrating spermatogonial development in the mammalian germline.

DB-DCAFN: dual-branch deformable cross-attention fusion network for bacterial segmentation.

Sputum smear tests are critical for the diagnosis of respiratory diseases. Automatic segmentation of bacteria from sputum smear images is important for improving diagnostic efficiency. However, this remains a challenging task owing to the high interclass similarity among different categories of bacteria and the low contrast of the bacterial edges. To explore more levels of global pattern features to promote the distinguishing ability of bacterial categories and maintain sufficient local fine-grained features to ensure accurate localization of ambiguous bacteria simultaneously, we propose a novel dual-branch deformable cross-attention fusion network (DB-DCAFN) for accurate bacterial segmentation. Specifically, we first designed a dual-branch encoder consisting of multiple convolution and transformer blocks in parallel to simultaneously extract multilevel local and global features. We then designed a sparse and deformable cross-attention module to capture the semantic dependencies between local and global features, which can bridge the semantic gap and fuse features effectively. Furthermore, we designed a feature assignment fusion module to enhance meaningful features using an adaptive feature weighting strategy to obtain more accurate segmentation. We conducted extensive experiments to evaluate the effectiveness of DB-DCAFN on a clinical dataset comprising three bacterial categories: Acinetobacter baumannii, Klebsiella pneumoniae, and Pseudomonas aeruginosa. The experimental results demonstrate that the proposed DB-DCAFN outperforms other state-of-the-art methods and is effective at segmenting bacteria from sputum smear images.

Maternal NAT10 orchestrates oocyte meiotic cell-cycle progression and maturation in mice.

In mammals, the production of mature oocytes necessitates rigorous regulation of the discontinuous meiotic cell-cycle progression at both the transcriptional and post-transcriptional levels. However, the factors underlying this sophisticated but explicit process remain largely unclear. Here we characterize the function of N-acetyltransferase 10 (Nat10), a writer for N4-acetylcytidine (ac4C) on RNA molecules, in mouse oocyte development. We provide genetic evidence that Nat10 is essential for oocyte meiotic prophase I progression, oocyte growth and maturation by sculpting the maternal transcriptome through timely degradation of poly(A) tail mRNAs. This is achieved through the ac4C deposition on the key CCR4-NOT complex transcripts. Importantly, we devise a method for examining the poly(A) tail length (PAT), termed Hairpin Adaptor-poly(A) tail length (HA-PAT), which outperforms conventional methods in terms of cost, sensitivity, and efficiency. In summary, these findings provide genetic evidence that unveils the indispensable role of maternal Nat10 in oocyte development.

Multimodal cross enhanced fusion network for diagnosis of Alzheimer's disease and subjective memory complaints.

Deep learning methods using multimodal imagings have been proposed for the diagnosis of Alzheimer's disease (AD) and its early stages (SMC, subjective memory complaints), which may help to slow the progression of the disease through early intervention. However, current fusion methods for multimodal imagings are generally coarse and may lead to suboptimal results through the use of shared extractors or simple downscaling stitching. Another issue with diagnosing brain diseases is that they often affect multiple areas of the brain, making it important to consider potential connections throughout the brain. However, traditional convolutional neural networks (CNNs) may struggle with this issue due to their limited local receptive fields. To address this, many researchers have turned to transformer networks, which can provide global information about the brain but can be computationally intensive and perform poorly on small datasets. In this work, we propose a novel lightweight network called MENet that adaptively recalibrates the multiscale long-range receptive field to localize discriminative brain regions in a computationally efficient manner. Based on this, the network extracts the intensity and location responses between structural magnetic resonance imagings (sMRI) and 18-Fluoro-Deoxy-Glucose Positron Emission computed Tomography (FDG-PET) as an enhancement fusion for AD and SMC diagnosis. Our method is evaluated on the publicly available ADNI datasets and achieves 97.67% accuracy in AD diagnosis tasks and 81.63% accuracy in SMC diagnosis tasks using sMRI and FDG-PET. These results achieve state-of-the-art (SOTA) performance in both tasks. To the best of our knowledge, this is one of the first deep learning research methods for SMC diagnosis with FDG-PET.

MgIG Attenuates Oxaliplatin-induced Hepatotoxicity through Suppression of Connexin 43 in Hepatic Stellate Cells.

Background and Aims: Oxaliplatin is widely used in cancer chemotherapy with adverse effects such as liver toxicity. Magnesium isoglycyrrhizinate (MgIG) has hepatoprotective effects, but the underlying mechanism remains elusive. The study's aim was to investigate the mechanism underlying the hepatoprotective effects of MgIG against oxaliplatin-induced liver injury. Methods: A xenografted colorectal cancer mouse model was established with MC38 cells. Mice were given oxaliplatin (6 mg/kg/week) for 5 weeks to mimic oxaliplatin-induced liver injury in vivo. LX-2 human hepatic stellate cell s(HSCs) were employed for in vitro studies. Serological tests, hematoxylin and eosin staining, oil red O staining and transmission electron microscopy were used for histopathological examinations. Real-time PCR, western blotting, immunofluorescence and immunohistochemical staining were used to determine Cx43 mRNA or protein levels. Flow cytometry was used to assay reactive oxygen species (ROS) and mitochondrial membrane. Short hairpin RNA targeting Cx43 was lentivirally transduced in LX-2 cells. Ultra-high performance liquid chromatography-tandem mass spectrometry was used to determine MgIG and metabolite concentration. Results: MgIG (40 mg/kg/day) treatment significantly reduced serum aspartate transaminase (AST) and alanine transaminase (ALT) levels in the mouse model, and alleviated liver pathological changes, including necrosis, sinusoidal expansion, mitochondrial damage, and fibrosis. MgIG reduced the abnormal expression of Cx43 in the mitochondria and nuclei of HSCs. MgIG inhibited the activation of HSCs via reducing ROS generation, mitochondrial dysfunction, and N-cadherin transcription. MgIG's inhibition of HSCs activation was abolished after knockdown of Cx43 in LX-2 cells. Conclusions: Cx43 mediated MgIG's hepatoprotective effects against oxaliplatin-induced toxicity.

ELOVL gene family plays a virtual role in response to breeding selection and lipid deposition in different tissues in chicken (Gallus gallus).

BACKGROUND: Elongases of very long chain fatty acids (ELOVLs), a family of first rate-limiting enzymes in the synthesis of long-chain fatty acids, play an essential role in the biosynthesis of complex lipids. Disrupting any of ELOVLs affects normal growth and development in mammals. Genetic variations in ELOVLs are associated with backfat or intramuscular fatty acid composition in livestock. However, the effects of ELOVL gene family on breeding selection and lipid deposition in different tissues are still unknown in chickens. RESULTS: Genetic variation patterns and genetic associations analysis showed that the genetic variations of ELOVL genes were contributed to breeding selection of commercial varieties in chicken, and 14 SNPs in ELOVL2-6 were associated with body weight, carcass or fat deposition traits. Especially, one SNP rs17631638T > C in the promoter of ELOVL3 was associated with intramuscular fat content (IMF), and its allele frequency was significantly higher in native and layer breeds compared to that in commercial broiler breeds. Quantitative real-time PCR (qRT-PCR) determined that the ELOVL3 expressions in pectoralis were affected by the genotypes of rs17631638T > C. In addition, the transcription levels of ELOVL genes except ELOVL5 were regulated by estrogen in chicken liver and hypothalamus with different regulatory pathways. The expression levels of ELOVL1-6 in hypothalamus, liver, abdominal fat and pectoralis were correlated with abdominal fat weight, abdominal fat percentage, liver lipid content and IMF. Noteworthily, expression of ELOVL3 in pectoralis was highly positively correlated with IMF and glycerophospholipid molecules, including phosphatidyl choline, phosphatidyl ethanolamine, phosphatidyl glycerol and phospholipids inositol, rich in ω-3 and ω-6 long-chain unsaturated fatty acids, suggesting ELOVL3 could contribute to intramuscular fat deposition by increasing the proportion of long-chain unsaturated glycerophospholipid molecules in pectoralis. CONCLUSIONS: In summary, we demonstrated the genetic contribution of ELOVL gene family to breeding selection for specialized varieties, and revealed the expression regulation of ELOVL genes and their potential roles in regulating lipid deposition in different tissues. This study provides new insights into understanding the functions of ELOVL family on avian growth and lipid deposition in different tissues and the genetic variation in ELOVL3 may aid the marker-assisted selection of meat quality in chicken.

GPNMB promotes abdominal fat deposition in chickens: genetic variation, expressional profile, biological function, and transcriptional regulation.

Glycoprotein nonmetastatic melanoma protein B (GPNMB) is a vital secreted factor that promotes the occurrence of obesity in mammals. However, the effects of GPNMB on abdominal fat deposition is still unknown in chickens. In this study, we looked into the genetic and expression association of GPNMB gene with abdominal fat traits in chicken, and found that a genetic variation rs31126482 in GPNMB promoter was significantly associated with abdominal fat weight (AFW, P < 0.05) and abdominal fat percentage (AFP, P < 0.01). Express profile analysis of the GPNMB indicated that the gene was mainly expressed in abdominal fat tissue, and its expression level was strongly positively correlated with AFW (R2 = 0.6356, P = 4.10E-05) and AFP (R2 = 0.6450, P = 2.90E-05). We then investigated biological function of GPNMB on adipogenesis in chicken, and found that GPNMB could inhibit abdominal preadipocyte proliferation, but promote abdominal preadipocyte differentiation and lipid deposition. Furthermore, we explored regulatory mechanism of GPNMB gene in chicken, and detected one nonclassical estrogen regulatory element (AP1) and one peroxisome proliferator-activated receptor α (PPARα) responsive element in the 2 kb promoter region of GPNMB gene, and demonstrated that estrogen could up-regulate GPNMB mRNA expression in adipose tissue and primary abdominal preadipocytes, while PPARα could down-regulate GPNMB expression in primary preadipocytes. Taken together, this study brings new insights into understanding the function and transcriptional control of GPNMB gene, and provides genetic markers for breeding selection to improve abdominal fat traits in chicken.

CDFRegNet: A cross-domain fusion registration network for CT-to-CBCT image registration.

BACKGROUND AND OBJECTIVE: Computer tomography (CT) to cone-beam computed tomography (CBCT) image registration plays an important role in radiotherapy treatment placement, dose verification, and anatomic changes monitoring during radiotherapy. However, fast and accurate CT-to-CBCT image registration is still very challenging due to the intensity differences, the poor image quality of CBCT images, and inconsistent structure information. METHODS: To address these problems, a novel unsupervised network named cross-domain fusion registration network (CDFRegNet) is proposed. First, a novel edge-guided attention module (EGAM) is designed, aiming at capturing edge information based on the gradient prior images and guiding the network to model the spatial correspondence between two image domains. Moreover, a novel cross-domain attention module (CDAM) is proposed to improve the network's ability to guide the network to effectively map and fuse the domain-specific features. RESULTS: Extensive experiments on a real clinical dataset were carried out, and the experimental results verify that the proposed CDFRegNet can register CT to CBCT images effectively and obtain the best performance, while compared with other representative methods, with a mean DSC of 80.01±7.16%, a mean TRE of 2.27±0.62 mm, and a mean MHD of 1.50±0.32 mm. The ablation experiments also proved that our EGAM and CDAM can further improve the accuracy of the registration network and they can generalize well to other registration networks. CONCLUSION: This paper proposed a novel CT-to-CBCT registration method based on EGAM and CDAM, which has the potential to improve the accuracy of multi-domain image registration.

LMA-Net: A lesion morphology aware network for medical image segmentation towards breast tumors.

Breast tumor segmentation plays a critical role in the diagnosis and treatment of breast diseases. Current breast tumor segmentation methods are mainly deep learning (DL) based methods, which exacted the contrast information between tumors and backgrounds, and produced tumor candidates. However, all these methods were developed based on traditional standard convolutions, which may not be able to model various tumor shapes and extract pure information of tumors (the extracted information usually contain non-tumor information). Besides, the loss functions used in these methods mainly aimed to minimize the intra-class distances, while ignoring the influence of inter-class distances upon segmentation. In this paper, we propose a novel lesion morphology aware network to segment breast tumors in 2D magnetic resonance images (MRI). The proposed network employs a hierarchical structure that contains two stages: breast segmentation stage and tumor segmentation stage. In the tumor segmentation stage, we devise a tumor morphology aware network to incorporate pure tumor characteristics, which facilitates contrastive information extraction. Further, we propose a hybrid intra- and inter-class distance optimization loss to supervise the network, which can minimize intra-class distances meanwhile maximizing inter-class distances, hence reducing the potential false positive/negative pixels in segmentation results. Verified on a clinical 2D MRI breast tumor dataset, our proposed method achieves eminent segmentation results and outperforms state-of-the-art methods, implying that the proposed method has a good prospect for clinical use.

Optimized protocol for isolation of germ cells from mouse testis by centrifugal elutriation.

Germline development is challenging to study due to the diversity of cell types in mammalian testis. Here, we present an optimized protocol, namely centrifugal elutriation, that allows the simultaneous isolation of mouse germ cells at different stages with high purity within ∼2 h. This approach exploits the JE-5.0 centrifugal elutriation system that fractionates cells based on differential sedimentation gravity. We herein provide the optimized parameters and procedures for isolation of elongating spermatids, round spermatids, and pachytene spermatocytes from adult mouse testes. For complete details on the use and execution of this protocol, please refer to Bao et al. (2018).

Danshensu Attenuated Epithelial-Mesenchymal Transformation and Chemoresistance of Colon Cancer Cells Induced by Platelets.

BACKGROUND: The interactions between platelets and tumor cells are well-known to play important roles in the progression of malignant tumors. Danshensu, a main water-soluble component of Salvia miltiorrhiza, can resist platelet aggregation and exert significant anti-tumor effects on various types of tumors. However, whether Danshensu could inhibit the progression of malignant tumors by suppressing the activities of platelets had not been reported. METHODS: The effects of Danshensu on the platelet activity and epithelial-mesenchymal transformation (EMT)-like invasive phenotype of SW620 colon cancer cells were assessed by stimulating with the supernatants from co-cultured platelets and SW620 cells with direct contact (SCP). The expression and secretion of proteins were determined by western blot and enzyme-linked immunosorbent assay (ELISA), respectively. Hematoxylin and eosin (H&E) staining was performed to analyzed the histopathology of tumor tissues and immunohistochemical staining was conducted to examine the protein expression in tumors. RESULTS: Co-incubation of SW620 cells with platelets directly or SCP both generated long spindle-shaped invasive phenotype. Pretreatment of platelets with Danshensu (25 µM) inhibited the morphological changes of SW620 cells induced by SCP, which was associated with the inhibitory effects of Danshensu on platelet secretion. Danshensu diminished the secretion of a list of biological factors in SCP, including interleukin (IL)-6, tumor necrosis factor alpha (TNF-α), IL-1ß and vascular endothelial growth factor (VEGF) that are all involved in tumor cell EMT and chemoresistance. Moreover, Danshensu up-regulated the expression of E-cadherin but down-regulated the levels of N-cadherin and Vimentin, resulting in the repression of SW620 cell migration. It was also shown that Danshensu enhanced the sensitivity of SW620 cells to oxaliplatin by suppressing the expression of MDR1. Furthermore, Danshensu could not only reduced the growth of subcutaneous tumors and liver metastasis that induced by SCP, but also down-regulated the expression of MDR1 in vivo. Mechanistic studies revealed that Danshensu suppressed the activation of the TGF-ß/Smad signaling pathway. CONCLUSIONS: Danshensu attenuated EMT-like characteristics and chemoresistance by inhibiting secretion capability of platelets and activation of the TGF-ß/Smad signaling pathway, suggesting that it may be optimized to be a therapeutic agent for fighting against colon cancer.

A CysB regulator positively regulates cysteine synthesis, expression of type III secretion system genes, and pathogenicity in Ralstonia solanacearum.

A syringe-like type III secretion system (T3SS) plays essential roles in the pathogenicity of Ralstonia solanacearum, which is a causal agent of bacterial wilt disease on many plant species worldwide. Here, we characterized functional roles of a CysB regulator (RSc2427) in R. solanacearum OE1-1 that was demonstrated to be responsible for cysteine synthesis, expression of the T3SS genes, and pathogenicity of R. solanacearum. The cysB mutants were cysteine auxotrophs that failed to grow in minimal medium but grew slightly in host plants. Supplementary cysteine substantially restored the impaired growth of cysB mutants both in minimal medium and inside host plants. Genes of cysU and cysI regulons have been annotated to function for R. solanacearum cysteine synthesis; CysB positively regulated expression of these genes. Moreover, CysB positively regulated expression of the T3SS genes both in vitro and in planta through the PrhG to HrpB pathway, whilst impaired expression of the T3SS genes in cysB mutants was independent of growth deficiency under nutrient-limited conditions. CysB was also demonstrated to be required for exopolysaccharide production and swimming motility, which contribute jointly to the host colonization and infection process of R. solanacearum. Thus, CysB was identified here as a novel regulator on the T3SS expression in R. solanacearum. These results provide novel insights into understanding of various biological functions of CysB regulators and complex regulatory networks on the T3SS in R. solanacearum.

FMRNet: A fused network of multiple tumoral regions for breast tumor classification with ultrasound images.

PURPOSE: Recent studies have illustrated that the peritumoral regions of medical images have value for clinical diagnosis. However, the existing approaches using peritumoral regions mainly focus on the diagnostic capability of the single region and ignore the advantages of effectively fusing the intratumoral and peritumoral regions. In addition, these methods need accurate segmentation masks in the testing stage, which are tedious and inconvenient in clinical applications. To address these issues, we construct a deep convolutional neural network that can adaptively fuse the information of multiple tumoral-regions (FMRNet) for breast tumor classification using ultrasound (US) images without segmentation masks in the testing stage. METHODS: To sufficiently excavate the potential relationship, we design a fused network and two independent modules to extract and fuse features of multiple regions simultaneously. First, we introduce two enhanced combined-tumoral (EC) region modules, aiming to enhance the combined-tumoral features gradually. Then, we further design a three-branch module for extracting and fusing the features of intratumoral, peritumoral, and combined-tumoral regions, denoted as the intratumoral, peritumoral, and combined-tumoral module. Especially, we design a novel fusion module by introducing a channel attention module to adaptively fuse the features of three regions. The model is evaluated on two public datasets including UDIAT and BUSI with breast tumor ultrasound images. Two independent groups of experiments are performed on two respective datasets using the fivefold stratified cross-validation strategy. Finally, we conduct ablation experiments on two datasets, in which BUSI is used as the training set and UDIAT is used as the testing set. RESULTS: We conduct detailed ablation experiments about the proposed two modules and comparative experiments with other existing representative methods. The experimental results show that the proposed method yields state-of-the-art performance on both two datasets. Especially, in the UDIAT dataset, the proposed FMRNet achieves a high accuracy of 0.945 and a specificity of 0.945, respectively. Moreover, the precision (PRE = 0.909) even dramatically improves by 21.6% on the BUSI dataset compared with the existing method of the best result. CONCLUSION: The proposed FMRNet shows good performance in breast tumor classification with US images, and proves its capability of exploiting and fusing the information of multiple tumoral-regions. Furthermore, the FMRNet has potential value in classifying other types of cancers using multiple tumoral-regions of other kinds of medical images.

Building RNA-protein germ granules: insights from the multifaceted functions of DEAD-box helicase Vasa/Ddx4 in germline development.

The segregation and maintenance of a dedicated germline in multicellular organisms is essential for species propagation in the sexually reproducing metazoan kingdom. The germline is distinct from somatic cells in that it is ultimately dedicated to acquiring the "totipotency" and to regenerating the offspring after fertilization. The most striking feature of germ cells lies in the presence of characteristic membraneless germ granules that have recently proven to behave like liquid droplets resulting from liquid-liquid phase separation (LLPS). Vasa/Ddx4, a faithful DEAD-box family germline marker highly conserved across metazoan species, harbors canonical DEAD-box motifs and typical intrinsically disordered sequences at both the N-terminus and C-terminus. This feature enables it to serve as a primary driving force behind germ granule formation and helicase-mediated RNA metabolism (e.g., piRNA biogenesis). Genetic ablation of Vasa/Ddx4 or the catalytic-dead mutations abolishing its helicase activity led to sexually dimorphic germline defects resulting in either male or female sterility among diverse species. While recent efforts have discovered pivotal functions of Vasa/Ddx4 in somatic cells, especially in multipotent stem cells, we herein summarize the helicase-dependent and -independent functions of Vasa/Ddx4 in the germline, and discuss recent findings of Vasa/Ddx4-mediated phase separation, germ granule formation and piRNA-dependent retrotransposon control essential for germline development.

Diagnostic Test Accuracy of Deep Learning Detection of COVID-19: A Systematic Review and Meta-Analysis.

RATIONALE AND OBJECTIVE: To perform a meta-analysis to compare the diagnostic test accuracy (DTA) of deep learning (DL) in detecting coronavirus disease 2019 (COVID-19), and to investigate how network architecture and type of datasets affect DL performance. MATERIALS AND METHODS: We searched PubMed, Web of Science and Inspec from January 1, 2020, to December 3, 2020, for retrospective and prospective studies on deep learning detection with at least reported sensitivity and specificity. Pooled DTA was obtained using random-effect models. Sub-group analysis between studies was also carried out for data source and network architectures. RESULTS: The pooled sensitivity and specificity were 91% (95% confidence interval [CI]: 88%, 93%; I2 = 69%) and 92% (95% CI: 88%, 94%; I2 = 88%), respectively for 19 studies. The pooled AUC and diagnostic odds ratio (DOR) were 0.95 (95% CI: 0.88, 0.92) and 112.5 (95% CI: 57.7, 219.3; I2 = 90%) respectively. The overall accuracy, recall, F1-score, LR+ and LR- are 89.5%, 89.5%, 89.7%, 23.13 and 0.13. Sub-group analysis shows that the sensitivity and DOR significantly vary with the type of network architectures and sources of data with low heterogeneity are (I2 = 0%) and (I2 = 18%) for ResNet architecture and single-source datasets, respectively. CONCLUSION: The diagnosis of COVID-19 via deep learning has achieved incredible performance, and the source of datasets, as well as network architectures, strongly affect DL performance.

The Spin1 interactor, Spindoc, is dispensable for meiotic division, but essential for haploid spermatid development in mice.

In mammals, germline development undergoes dramatic morphological and molecular changes and is epigenetically subject to intricate yet exquisite regulation. Which epigenetic players and how they participate in the germline developmental process are not fully characterized. Spin1 is a multifunctional epigenetic protein reader that has been shown to recognize H3 "K4me3-R8me2a" histone marks, and more recently the non-canonical bivalent H3 "K4me3-K9me3/2" marks as well. As a robust Spin1-interacting cofactor, Spindoc has been identified to enhance the binding of Spin1 to its substrate histone marks, thereby modulating the downstream signaling; However, the physiological role of Spindoc in germline development is unknown. We generated two Spindoc knockout mouse models through CRISPR/Cas9 strategy, which revealed that Spindoc is specifically required for haploid spermatid development, but not essential for meiotic divisions in spermatocytes. This study unveiled a new epigenetic player that participates in haploid germline development.