A microbial knowledge graph-based deep learning model for predicting candidate microbes for target hosts.

Predicting interactions between microbes and hosts plays critical roles in microbiome population genetics and microbial ecology and evolution. How to systematically characterize the sophisticated mechanisms and signal interplay between microbes and hosts is a significant challenge for global health risks. Identifying microbe-host interactions (MHIs) can not only provide helpful insights into their fundamental regulatory mechanisms, but also facilitate the development of targeted therapies for microbial infections. In recent years, computational methods have become an appealing alternative due to the high risk and cost of wet-lab experiments. Therefore, in this study, we utilized rich microbial metagenomic information to construct a novel heterogeneous microbial network (HMN)-based model named KGVHI to predict candidate microbes for target hosts. Specifically, KGVHI first built a HMN by integrating human proteins, viruses and pathogenic bacteria with their biological attributes. Then KGVHI adopted a knowledge graph embedding strategy to capture the global topological structure information of the whole network. A natural language processing algorithm is used to extract the local biological attribute information from the nodes in HMN. Finally, we combined the local and global information and fed it into a blended deep neural network (DNN) for training and prediction. Compared to state-of-the-art methods, the comprehensive experimental results show that our model can obtain excellent results on the corresponding three MHI datasets. Furthermore, we also conducted two pathogenic bacteria case studies to further indicate that KGVHI has excellent predictive capabilities for potential MHI pairs.

PTBGRP: predicting phage-bacteria interactions with graph representation learning on microbial heterogeneous information network.

Identifying the potential bacteriophages (phage) candidate to treat bacterial infections plays an essential role in the research of human pathogens. Computational approaches are recognized as a valid way to predict bacteria and target phages. However, most of the current methods only utilize lower-order biological information without considering the higher-order connectivity patterns, which helps to improve the predictive accuracy. Therefore, we developed a novel microbial heterogeneous interaction network (MHIN)-based model called PTBGRP to predict new phages for bacterial hosts. Specifically, PTBGRP first constructs an MHIN by integrating phage-bacteria interaction (PBI) and six bacteria-bacteria interaction networks with their biological attributes. Then, different representation learning methods are deployed to extract higher-level biological features and lower-level topological features from MHIN. Finally, PTBGRP employs a deep neural network as the classifier to predict unknown PBI pairs based on the fused biological information. Experiment results demonstrated that PTBGRP achieves the best performance on the corresponding ESKAPE pathogens and PBI dataset when compared with state-of-art methods. In addition, case studies of Klebsiella pneumoniae and Staphylococcus aureus further indicate that the consideration of rich heterogeneous information enables PTBGRP to accurately predict PBI from a more comprehensive perspective. The webserver of the PTBGRP predictor is freely available at http://120.77.11.78/PTBGRP/.

KAT8 functions in redox homeostasis and mitochondrial dynamics during mouse oocyte meiosis progression.

As a histone acetyltransferase, lysine acetyltransferase 8 (KAT8) participates in diverse biological processes. However, the effect of KAT8 on oocyte maturation in mice remains unclear. In this study, we found that mouse oocytes overexpressing Kat8-OE induced maturation failure manifested reduced rates of GVBD and first polar body emission. In addition, immunostaining results revealed that Kat8 overexpressing oocytes showed inappropriate mitochondrial distribution patterns, overproduction of reactive oxygen species (ROS), accumulation of phosphorylated γH2AX, hyperacetylation of α-tubulin, and severely disrupted spindle/chromosome organization. Moreover, we revealed that Kat8 overexpression induced a decline in SOD1 proteins and KAT8's interaction with SOD1 in mouse ovaries via immunoprecipitation. Western blotting data confirmed that Kat8-OE induced downregulation of SOD1 expression, which is a key factor for the decline of oocyte quality in advanced maternal age. Also, the injection of Myc-Sod1 cRNA could partially rescue maternal age-induced meiotic defects in oocytes. In conclusion, our data demonstrated that high level of KAT8 inhibited SOD1 activity, which in turn induced defects of mitochondrial dynamics, imbalance of redox homeostasis, and spindle/chromosome disorganization during mouse oocyte maturation.

DFT Guided Design and Preparation of Quasi-Nanocrystalline Hf-La2O3 Cathode with Unprecedented Thermal Emission Performance.

With the guidance of density functional theory (DFT), a high-performance hafnium (Hf) cathode for an air/water vapor plasma torch is designed and the concepts and principles for high performance are elucidated. A quasi-nanocrystalline hexagonal close-packed (HCP) Hf-La2O3 cathode based on these design principles is successfully fabricated via a powder metallurgy route. Under identical voltage and temperature conditions, the thermal emission current density of this quasi-nanocrystalline Hf-La2O3 cathode is ≈20 times greater than that of conventional Hf cathodes. Additionally, its cathodic lifespan is significantly extended. Quasi-nanocrystalline Hf-La2O3 products are manufactured into cathode devices with standard dimensions. This fabrication process is straightforward, requires minimal doped oxides, and is cost-effective. Consequently, the approach offers substantial performance enhancements over traditional Hf melting methods without incurring significantly additional costs.

Initial [18F]DCFPyL PET/CT in treatment-naïve prostate cancer: correlation with post-ADT PSA outcomes and recurrence.

PURPOSE: Positron emission tomography (PET) with prostate-specific membrane antigen (PSMA) targeting tracers has emerged as a valuable diagnostic tool for prostate cancer (PCa), androgen deprivation therapy (ADT) stands as the cornerstone treatment for advanced PCa, yet forecasting the response to hormonal therapy poses a significant clinical hurdle. METHODS: In a prospective cohort of 86 PCa patients undergoing short-term ADT, this study evaluated the prognostic potential of [18F]DCFPyL PET/CT scans. Comprehensive data encompassing clinical profiles, baseline prostate-specific antigen (PSA) levels, and imaging metrics were assessed. We developed predictive models for assessing decreases in PSA levels (PSA50 and PSA70) based on a combination of PET-related parameters and clinical factors. Kaplan-Meier survival analysis was utilized to ascertain the prognostic value of PET-based metrics. RESULTS: In this study, elevated [18F]DCFPyL uptake within the primary tumor, as indicated by a SUV ≥ 6.78 (p = 0.0024), and a reduction in the tumor volume (TV) of primary PSMA-avid tumor with PSMA-TV < 41.96 cm3 (p = 0.038), as well as an increased burden of metastatic PSMA-avid tumor, with PSMA-TV (PSMA-TV ≥ 71.39 cm3) (p = 0.012) were identified in association with diminished progression-free survival (PFS). PET and clinical parameters demonstrated constrained predictive capacity for PSA50 response as indicated by an area under the curve (AUC) of 0.442. CONCLUSION: Our study revealed that pretreatment [18F]DCFPyL uptake in primary or metastatic tumor sites is prognostically relevant in high-risk PCa patients undergoing ADT. Further research is needed to develop robust predictive models in this multifaceted landscape of PCa management.

NET1 is a critical regulator of spindle assembly and actin dynamics in mouse oocytes.

BACKGROUND: Neuroepithelial transforming gene 1 (NET1) is a RhoA subfamily guanine nucleotide exchange factor that governs a wide array of biological processes. However, its roles in meiotic oocyte remain unclear. We herein demonstrated that the NET1-HACE1-RAC1 pathway mediates meiotic defects in the progression of oocyte maturation. METHODS: NET1 was reduced using a specific small interfering RNA in mouse oocytes. Spindle assembly, chromosomal alignment, the actin cap, and chromosomal spreads were visualized by immunostaining and analyzed under confocal microscopy. We also applied mass spectroscopy, and western blot analysis for this investigation. RESULTS: Our results revealed that NET1 was localized to the nucleus at the GV stage, and that after GVBD, NET1 was localized to the cytoplasm and predominantly distributed around the chromosomes, commensurate with meiotic progression. NET1 resided in the cytoplasm and significantly accumulated on the spindle at the MI and MII stages. Mouse oocytes depleted of Net1 exhibited aberrant first polar body extrusion and asymmetric division defects. We also determined that Net1 depletion resulted in reduced RAC1 protein expression in mouse oocytes, and that NET1 protected RAC1 from degradation by HACE1, and it was essential for actin dynamics and meiotic spindle formation. Importantly, exogenous RAC1 expression in Net1-depleted oocytes significantly rescued these defects. CONCLUSIONS: Our results suggest that NET1 exhibits multiple roles in spindle stability and actin dynamics during mouse oocyte meiosis.

Role of FBXL5 in redox homeostasis and spindle assembly during oocyte maturation in mice.

As an E3 ubiquitin ligase, F-box and leucine-rich repeat protein 5 (FBXL5) participates in diverse biologic processes. However, the role of Fbxl5 in mouse oocyte meiotic maturation has not yet been fully elucidated. The present study revealed that mouse oocytes depleted of Fbxl5 were unable to complete meiosis, as Fbxl5 silencing led to oocyte meiotic failure with reduced rates of GVBD and polar body extrusion. In addition, Fbxl5 depletion induced aberrant mitochondrial dynamics as we noted the overproduction of reactive oxygen species (ROS) and the accumulation of phosphorylated γH2AX with Fbxl5 knockdown. We also found that Fbxl5-KD led to the abnormal accumulation of CITED2 proteins in mouse oocytes. Our in vitro ubiquitination assay showed that FBXL5 interacted with CITED2 and that it mediated the degradation of CITED2 protein through the ubiquitination-proteasome pathway. Collectively, our data revealed critical functions of FBXL5 in redox hemostasis and spindle assembly during mouse oocyte maturation.

Strength Enhancement of Polyurethane Film by Solution Annealing.

Recently, polyurethane elastomer (TPU) has attracted more and more attention depending on its excellent optical, mechanical, and retreatment properties. The high strength of polyurethane has always been pursued, which can enable its application in more fields. In this work, an aliphatic polyurethane elastomer membrane (HRPU6) was successfully synthesized, and its strength was obviously improved by solvent annealing technology. The tensile strength and adhesion strength can reach 64.56 and 2.58 MPa, but 36.55 and 1.57 MPa only before solvent annealing, respectively. The impact strength of laminated glass based on HRPU has also been significantly improved after solvent annealing, confirmed through drop ball impact testing. It has been confirmed that the increase in strength of HRPU6 is attributed to the enhancement of hydrogen bonding and the improvement of the phase separation degree.

Loss of FAM172A gene prompts cell proliferation in liver regeneration.

The present study was designed to explore the function of FAM172A in liver regeneration and HCC. Mice were sacrificed after 70% partial hepatectomy (PH). RNA sequencing was performed on primary hepatocytes of WT and FAM172A-/- mice. We used HepG2 cells to construct cell lines with stably knockdown and overexpression of FAM172A. The expression of FAM172A in liver tissues was investigated by immunohistochemical staining, and we also used public database to perform survival analysis and prognostic model in HCC. Compared with WT mice after PH, normalized liver weight/body weight (LW/BW) ratio and the proliferating cell nuclear antigen (PCNA) protein level of FAM172A-/- mice elevated. The DEGs were mainly enriched in inflammatory response, tumor necrosis factor production, and wound healing. FAM172A knockdown enhanced the NFκB-TNFα and pERK-YAP1-Cyclin D1 axis. FAM172A peptide inhibited proliferation of primary hepatocytes. Moreover, the low expression of FAM172A in human HCC tissues implies a lower likelihood of survival and a valid diagnostic marker for HCC. Loss of FAM172A gene promotes cell proliferation by pERK-YAP1-Cyclin D1 and pNFκB-TNFα pathways during liver regeneration after PH. FAM172A may be a favorable diagnosis marker of HCC.

Identifying genes for regulating osteogenic differentiation of human periodontal ligament stem cells in inflammatory environments by bioinformatics analysis.

BACKGROUND AND OBJECTIVES: Periodontitis is an immuno-inflammatory disease caused by dental plaque biofilms and inflammations. The regeneration of bone tissue in inflammatory environment is of great significance for the treatment of periodontal disease, but the specific molecular mechanism of bone formation in periodontitis still needs further exploration. The objective of this study was to identify key osteogenesis-related genes (ORGs) in periodontitis. METHODS: We used two datasets from the Gene Expression Omnibus (GEO) database to find differentially expressed mRNAs and miRNAs, further performed Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. Then we predicted the downstream genes of the differentially expressed miRNAs (DEMs) by the TargetScan database and established a miRNA-mRNA regulatory network. Finally, the osteogenic mechanism of periodontitis was explored through quantitative real-time PCR (qRT-PCR) by inducing inflammatory environment and osteogenic differentiation of hPDLSCs. RESULTS: Through differential expression analysis and prediction of downstream target genes of DEMs, we created a miRNA-mRNA regulatory network consisting of 29 DEMs and 11 differentially expressed osteogenesis-related genes (DEORGs). In addition, the qRT-PCR results demonstrated that BTBD3, PLAT, AKAP12, SGK1, and GLCE expression levels were significantly upregulated, while those of TIMP3, ZCCHC14, LIN7A, DNAH6, NNT, and ITGA6 were downregulated under the dual effects of inflammatory stimulation and osteogenic induction. CONCLUSION: DEORGs might be important factors in the osteogenic phase of periodontitis, and the miRNA-mRNA network may shed light on the clarification of the role and mechanism of osteogenesis in periodontitis and contribute to the development of novel therapeutic strategies.

A study on the application of radiomics based on cardiac MR non-enhanced cine sequence in the early diagnosis of hypertensive heart disease.

BACKGROUND: The prevalence of hypertensive heart disease (HHD) is high and there is currently no easy way to detect early HHD. Explore the application of radiomics using cardiac magnetic resonance (CMR) non-enhanced cine sequences in diagnosing HHD and latent cardiac changes caused by hypertension. METHODS: 132 patients who underwent CMR scanning were divided into groups: HHD (42), hypertension with normal cardiac structure and function (HWN) group (46), and normal control (NOR) group (44). Myocardial regions of the end-diastolic (ED) and end-systolic (ES) phases of the CMR short-axis cine sequence images were segmented into regions of interest (ROI). Three feature subsets (ED, ES, and ED combined with ES) were established after radiomic least absolute shrinkage and selection operator feature selection. Nine radiomic models were built using random forest (RF), support vector machine (SVM), and naive Bayes. Model performance was analyzed using receiver operating characteristic curves, and metrics like accuracy, area under the curve (AUC), precision, recall, and specificity. RESULTS: The feature subsets included first-order, shape, and texture features. SVM of ED combined with ES achieved the highest accuracy (0.833), with a macro-average AUC of 0.941. AUCs for HHD, HWN, and NOR identification were 0.967, 0.876, and 0.963, respectively. Precisions were 0.972, 0.740, and 0.826; recalls were 0.833, 0.804, and 0.863, respectively; and specificities were 0.989, 0.863, and 0.909, respectively. CONCLUSIONS: Radiomics technology using CMR non-enhanced cine sequences can detect early cardiac changes due to hypertension. It holds promise for future use in screening for latent cardiac damage in early HHD.

Reversed-engineered human alveolar lung-on-a-chip model.

Here, we present a physiologically relevant model of the human pulmonary alveoli. This alveolar lung-on-a-chip platform is composed of a three-dimensional porous hydrogel made of gelatin methacryloyl with an inverse opal structure, bonded to a compartmentalized polydimethylsiloxane chip. The inverse opal hydrogel structure features well-defined, interconnected pores with high similarity to human alveolar sacs. By populating the sacs with primary human alveolar epithelial cells, functional epithelial monolayers are readily formed. Cyclic strain is integrated into the device to allow biomimetic breathing events of the alveolar lung, which, in addition, makes it possible to investigate pathological effects such as those incurred by cigarette smoking and severe acute respiratory syndrome coronavirus 2 pseudoviral infection. Our study demonstrates a unique method for reconstitution of the functional human pulmonary alveoli in vitro, which is anticipated to pave the way for investigating relevant physiological and pathological events in the human distal lung.

A New Brominated Isocoumarin from the Marine Starfish-Associated Fungus Aspergillus sp. WXF1904.

Based on the one strain many compounds strategy, a new brominated isocoumarin, 5-bromo-6,8-dihydroxy-3,7-dimethylisocoumarin (1), along with four new natural products, methyl 3-bromo-2,4-dihydroxy-6-methylbenzoate (2), methyl 2-bromo-4,6-dihydroxybenzoate (3), (E)-3-(3-bromo-4-hydroxyphenyl) acrylic acid (4) and 4-hydroxy-3-methyl-6-phenyl-2H-pyran-2-one (5), and four known compounds, methyl orsellinate (6), 4-hydroxy-3-methyl-6-(1-methyl-1-propenyl)-2H-pyran-2-one (7), pilobolusate (8) and cis-ferulic acid (9), were isolated from the ethyl acetate extract of the fungus Aspergillus sp. WXF1904 under the condition of adding bromine salt to the production medium. The structures of the new compounds were established by analysis of NMR and MS data. Compounds (1-9) were evaluated for inhibitory activity of acetylcholinesterase and pancreatic lipase, the new compound 1, known compounds 6 and 7 displayed weak inhibitory activity against acetylcholinesterase, compounds 2, 5, 7 and 8 showed weak inhibitory activity against pancreatic lipase.

Adaptive Clustering for Point Cloud.

The point cloud segmentation method plays an important role in practical applications, such as remote sensing, mobile robots, and 3D modeling. However, there are still some limitations to the current point cloud data segmentation method when applied to large-scale scenes. Therefore, this paper proposes an adaptive clustering segmentation method. In this method, the threshold for clustering points within the point cloud is calculated using the characteristic parameters of adjacent points. After completing the preliminary segmentation of the point cloud, the segmentation results are further refined according to the standard deviation of the cluster points. Then, the cluster points whose number does not meet the conditions are further segmented, and, finally, scene point cloud data segmentation is realized. To test the superiority of this method, this study was based on point cloud data from a park in Guilin, Guangxi, China. The experimental results showed that this method is more practical and efficient than other methods, and it can effectively segment all ground objects and ground point cloud data in a scene. Compared with other segmentation methods that are easily affected by parameters, this method has strong robustness. In order to verify the universality of the method proposed in this paper, we test a public data set provided by ISPRS. The method achieves good segmentation results for multiple sample data, and it can distinguish noise points in a scene.

Point Cloud Registration Method Based on Geometric Constraint and Transformation Evaluation.

Existing point-to-point registration methods often suffer from inaccuracies caused by erroneous matches and noisy correspondences, leading to significant decreases in registration accuracy and efficiency. To address these challenges, this paper presents a new coarse registration method based on a geometric constraint and a matrix evaluation. Compared to traditional registration methods that require a minimum of three correspondences to complete the registration, the proposed method only requires two correspondences to generate a transformation matrix. Additionally, by using geometric constraints to select out high-quality correspondences and evaluating the matrix, we greatly increase the likelihood of finding the optimal result. In the proposed method, we first employ a combination of descriptors and keypoint detection techniques to generate initial correspondences. Next, we utilize the nearest neighbor similarity ratio (NNSR) to select high-quality correspondences. Subsequently, we evaluate the quality of these correspondences using rigidity constraints and salient points' distance constraints, favoring higher-scoring correspondences. For each selected correspondence pair, we compute the rotation and translation matrix based on their centroids and local reference frames. With the transformation matrices of the source and target point clouds known, we deduce the transformation matrix of the source point cloud in reverse. To identify the best-transformed point cloud, we propose an evaluation method based on the overlap ratio and inliers points. Through parameter experiments, we investigate the performance of the proposed method under various parameter settings. By conducting comparative experiments, we verified that the proposed method's geometric constraints, evaluation methods, and transformation matrix computation consistently outperformed other methods in terms of root mean square error (RMSE) values. Additionally, we validated that our chosen combination for generating initial correspondences outperforms other descriptor and keypoint detection combinations in terms of the registration result accuracy. Furthermore, we compared our method with several feature-matching registration methods, and the results demonstrate the superior accuracy of our approach. Ultimately, by testing the proposed method on various types of point cloud datasets, we convincingly established its effectiveness. Based on the evaluation and selection of correspondences and the registration result's quality, our proposed method offers a solution with fewer iterations and higher accuracy.

Objective Noninvasive Measurement of the Volumizing Effect of a Dermal Filler: An In Vivo Study.

BACKGROUND: Information about the volumizing effects of dermal fillers is critical for physicians' understanding of product features and prudent decision-making in clinical practice. It is important for material engineers to develop and optimize new dermal fillers, especially when comparing the physiochemical properties of a new product with those of existing fillers that are used worldwide. OBJECTIVE: This study aimed to establish a reliable, noninvasive method for in vivo quantitative evaluation of the filling effect in order to predict possible effectiveness after filler injection and to evaluate the degradation trend over time. METHODS: A rabbit model of ear injection with dermal fillers was established. Hyaluronic acid (HA) filler was injected into the subcutaneous layer of rabbit ears, resulting in a stable skin bulge. Ultrasonography was used to noninvasively measure the skin bulge for volume calculation; the volume change was analyzed periodically until 38 weeks. Pathological examination, the gold standard, was performed to confirm degradation. RESULTS: The immediate volumizing effect of HA filler injection was macroscopically observed as a local skin bulge. Ultrasound was able to precisely detect the shape of the filler and calculate the length, width, and height of the skin bulge at each time point. The degree of uplift and amount of residual samples in the pathological evaluation were consistent with the results of morphological observation using ultrasound. CONCLUSION: Evaluation of the volume impact of dermal filler through the rabbit ear injection model evaluation enables material science evaluation in the early stage of material development, and has certain clinical reference value. LEVEL OF EVIDENCE I: This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .

The tandem EH domains of End3 cooperate to interact with dual XPF motifs of Sla1 for the connection of early and late stages in fungal endocytosis.

Clathrin-mediated endocytosis (CME) is imperative for physiological processes in eukaryotic cells. In fungi, the Pan1/End3/Sla1 complex controls the transition between early and late stages of CME. Although it is acknowledged that End3 uses its N-terminal to interact with the C-terminal of Sla1, detailed mechanism remains obscure. Magnaporthe oryzae, the pathogenic fungus of rice, cause blast disease that threatens rice production worldwide. Here we report the detailed interaction mechanism between End3 and Sla1 of M. oryzae, i.e. MoEnd3 and MoSla1. The two EH domains of MoEnd3 (MoEnd3-EH1 and MoEnd3-EH2) is different both in evolution and calcium binding, but are indispensable for conformational stability of each other, an unreported effect of tandem-arranged EH domains. MoEnd3-EH1 and MoEnd3-EH2 interact with peptide MoSla11145-1155 containing a NPF motif with a conserved mode, and MoEnd3-EHs (containing both EH1 and EH2 domains) binds MoSla11145-1155 with a higher affinity, supporting the synergetic effect of EH domains. In addition, MoEnd3-EHs also recognize peptide MoSla1971-981 with a new MPF motif that has not been reported before, while Sla1 of yeast contains a DPF motif that bears EH domain interaction ability. Collectively, our research shows that the two EH domains of End3 synergize to interact with dual XPF motifs of Sla1, which conforms to a bivalent receptor-bivalent ligand model to improve both affinity and specificity.

An Adhesive Bioink toward Biofabrication under Wet Conditions.

Three-dimensional (3D) bioprinting is driving significant innovations in biomedicine over recent years. Under certain scenarios such as in intraoperative bioprinting, the bioinks used should exhibit not only cyto/biocompatibility but also adhesiveness in wet conditions. Herein, an adhesive bioink composed of gelatin methacryloyl, gelatin, methacrylated hyaluronic acid, and skin secretion of Andrias davidianus is designed. The bioink exhibits favorable cohesion to allow faithful extrusion bioprinting in wet conditions, while simultaneously showing good adhesion to a variety of surfaces of different chemical properties, possibly achieved through the diverse bonds presented in the bioink formulation. As such, this bioink is able to fabricate sophisticated planar and volumetric constructs using extrusion bioprinting, where the dexterity is further enhanced using ergonomic handheld bioprinters to realize in situ bioprinting. In vitro experiments reveal that cells maintain high viability; further in vivo studies demonstrate good integration and immediate injury sealing. The characteristics of the bioink indicate its potential widespread utility in extrusion bioprinting and will likely broaden the applications of bioprinting toward situations such as in situ dressing and minimally invasive tissue regeneration.

A transferable deep learning approach to fast screen potential antiviral drugs against SARS-CoV-2.

The COVID-19 pandemic calls for rapid development of effective treatments. Although various drug repurpose approaches have been used to screen the FDA-approved drugs and drug candidates in clinical phases against SARS-CoV-2, the coronavirus that causes this disease, no magic bullets have been found until now. In this study, we used directed message passing neural network to first build a broad-spectrum anti-beta-coronavirus compound prediction model, which gave satisfactory predictions on newly reported active compounds against SARS-CoV-2. Then, we applied transfer learning to fine-tune the model with the recently reported anti-SARS-CoV-2 compounds and derived a SARS-CoV-2 specific prediction model COVIDVS-3. We used COVIDVS-3 to screen a large compound library with 4.9 million drug-like molecules from ZINC15 database and recommended a list of potential anti-SARS-CoV-2 compounds for further experimental testing. As a proof-of-concept, we experimentally tested seven high-scored compounds that also demonstrated good binding strength in docking studies against the 3C-like protease of SARS-CoV-2 and found one novel compound that can inhibit the enzyme. Our model is highly efficient and can be used to screen large compound databases with millions or more compounds to accelerate the drug discovery process for the treatment of COVID-19.

Evaluating Upstaging in Ductal Carcinoma In Situ Using Preoperative MRI-Based Radiomics.

BACKGROUND: About 20%-40% of patients diagnosed with ductal carcinoma in situ (DCIS) by core needle biopsy (CNB) will develop invasive cancer at the time of excision. Improving the preoperative diagnosis of DCIS is important for surgical planning. PURPOSE: To establish an MRI-based radiomics nomogram for preoperatively evaluating the upstaging of DCIS patients and help with risk stratification. STUDY TYPE: Retrospective. POPULATION: A total of 227 patients (50.5 ± 9.7 years; 67 upstaged DCIS) were divided into training (n = 109), internal (n = 47), and external (n = 71) validation cohort. FIELD STRENGTH/SEQUENCE: 1.5-T or 3-T, dynamic contrast-enhanced (DCE) imaging, and diffusion-weighted imaging (DWI). ASSESSMENT: DCIS lesions were manually segmented using ITK-SNAP software and 1304 radiomic features were extracted from DCE, DWI, and apparent diffusion coef-ficient (ADC) maps, respectively. A radscore was calculated by a random forest algo-rithm based on DCIS upstaging-related radiomic features, which selected by a coarse-to-fine method including interclass correlation coefficient, single-factor anal-ysis, and the least absolute shrinkage and selection operator (LASSO) method. Uni-variate and multivariate logistic regression was used to analyze the independent risk factors, including age, location, lesion size, estrogen receptor (ER) status, and other clinico-pathologic factors. Finally, Mann-Whitney U tests were performed to com-pare the differences in radscore between low/intermediate and high nuclear grade groups for pure DCIS patients. STATISTICAL TESTS: Student's t-tests or Mann-Whitney U tests, chi-square-tests, or Fisher's-tests, univariate and multivariate logistic regression analysis, calibration curve, Youden index, the area under the curve (AUC), Delong test, net reclassification improvement (NRI), and integrated discrimination improvement (IDI) analyses. RESULTS: Eight important radiomic features (two from ADC, three from DWI, and three from DCE) were selected for calculating radscore. Clinical model including age and ER was established with AUCs of 0.747 and 0.738 in the internal and external validation cohorts, respectively. A combined model integrating age, estrogen receptor (ER), and radscore were also constructed with AUCs of 0.887 and 0.881. Further subgroup analysis showed that pure DCIS patients with different nuclear grade have significant differences in radscore. DATA CONCLUSION: Multisequence MRI radiomics may preoperatively evaluate the upstaging of DCIS and might provide personalized image-based clinical decision support. EVIDENCE LEVEL: 4. TECHNICAL EFFICACY: Stage 2.