OBJECTIVES:This prospective study aimed to develop and validate integrated nomograms that combine preoperative radiomic features from dual-layer spectral detector CT (DLCT) with key postoperative pathological information for the prediction of programmed cell death ligand 1 (PD-L1) expression in invasive lung adenocarcinoma. METHODS:The study included 191 participants with invasive lung adenocarcinoma who underwent preoperative thoracic contrast-enhanced DLCT scans and PD-L1 expression testing. Radiomic features were extracted from various DLCT images. Least absolute shrinkage and selection operator was used to derive radscores for PD-L1 expression (tumor proportion score ≥1% was defined as PD-L1 positivity). Nomograms were developed by integrating radscores with clinicopathological characteristics through logistic regression analysis. Performance was assessed using receiver operator characteristic (ROC) curves, area under the curve (AUC), calibration curves, and decision curve analysis (DCA). RESULTS:Three sets of radiomic nomograms, based on iodine map (IM), virtual non-contrast (VNC), and conventional images (PCI), along with pathological stages (pTNM), were developed. The IM nomogram exhibited superior performance in both training (AUC = 0.791) and validation (AUC = 0.737) sets. Calibration and DCA confirmed the IM nomogram's consistency and clinical utility. CONCLUSIONS:The IM nomogram demonstrated potential for individualized prediction of PD-L1 expression in invasive lung adenocarcinoma and identify the candidates who may benefit from immunotherapy. ADVANCES IN KNOWLEDGE:The nomogram based on dual-layer spectral detector CT can predict the probability of PD-L1 expression in invasive lung adenocarcinoma and may help personalized immunotherapy decisions.

Wide applications of per-/poly-fluoroalkyl substances (PFASs) in manufacturing fluorinated-polymer coatings and products pose serious health and environmental threats due to their highly chemical robustness. The development of efficient sensors for detecting persistent organic pollutants is thus essential. In this study, we present the supramolecular assemblies of multinuclear platinum(ii) complexes, which demonstrate the capability to detect and extract perfluorohexanoic acid (PFHxA) in deionized water. These platinum(ii) complexes have exhibited remarkable photophysical properties and supramolecular self-assembly in mixed organic solvents, driven by Pt(ii)⋯Pt(ii) and π-π stacking interactions, to form two-dimensional supramolecular polymers with hexagonal and rectangular packing, manifested as sheet-like nanostructures. Specifically, the dynamic association of PFASs with the resulting nanosheets formed by platinum(ii) complexes induces significant changes in the UV-vis absorption and emission spectra. Mechanistic studies, supported by molecular dynamics simulations, suggest that these pronounced spectroscopic changes are induced by the strong interactions of PFASs with platinum(ii) complexes and the surrounding solvent molecules. By means of liquid-phase extraction, we have demonstrated the innovative use of platinum(ii) supramolecular polymer networks for proof-of-concept detection and extraction of PFHxA in deionized water. This work provides important insights into the fundamental molecular design principles that are crucial for the future development of novel supramolecular functional materials.

IntroductionLysine crotonylation (Kcr) is an important post-translational modification (PTM) of proteins, playing a key role in regulating various biological processes in pathogenic fungi. However, the experimental identification of Kcr sites remains challenging due to the high cost and time-consuming nature of mass spectrometry-based techniques.MethodsTo address this limitation, we developed Fungi-Kcr, a deep learning-based model designed to predict Kcr modification sites in fungal proteins. The model integrates convolutional neural networks (CNN), gated recurrent units (GRU), and word embedding to effectively capture both local and long-range sequence dependencies.ResultsComprehensive evaluations, including ten-fold cross-validation and independent testing, demonstrate that Fungi-Kcr achieves superior predictive performance compared to conventional machine learning models. Moreover, our results indicate that a general predictive model performs better than species-specific models. DiscussionThe proposed model provides a valuable computational tool for the large-scale identification of Kcr sites, contributing to a deeper understanding of fungal pathogenesis and potential therapeutic targets. The source code and dataset for Fungi-Kcr are available at https://github.com/zayra77/Fungi-Kcr.

DNA-encoded libraries are invaluable tools for high-throughput screening and functional genomics studies. However, constructing high-abundance libraries in mammalian cells remains challenging. Here, we present dsDNA-assembly-PCR (dsDAP), a novel Gibson-assembly-PCR strategy for creating DNA-encoded libraries, offering improved flexibility and efficiency over previous methods. We demonstrated this approach by investigating the impact of translation initiation sequences (TIS) on protein expression in HEK293T cells. Both CRISPR-Cas9 and piggyBac systems were employed for genomic integration, allowing comparison of different integration methods. Our results confirmed the importance of specific nucleotides in the TIS region, particularly the preference for adenine at the -3 position in high-expression sequences. We also explored the effects of library dilution on genotype-phenotype correlations. This Gibson-assembly-PCR strategy overcomes limitations of existing methods, such as restriction enzyme dependencies, and provides a versatile tool for constructing high-abundance libraries in mammalian cells. Our approach has broad applications in functional genomics, drug discovery, and the study of gene regulation.

Sialylation, a key post-translational modification essential for protein function, is regulated by steroid hormones, along with other glycosylations like fucosylation. These modifications influence tumor growth and metastasis by modulating immune activation. MicroRNAs (miRNAs), crucial in gene expression, affect sialylation and are emerging as promising biomarkers in breast cancer, though their prognostic value remains unclear. Sialylation-related miRNAs were identified through Pearson correlation analysis, and an eight-miRNA risk signature was developed using univariate and Least Absolute Shrinkage and Selection Operator (LASSO) regression in the TCGA dataset. The prognostic value was validated in two independent GEO datasets. Multivariate analysis confirmed that the miRNA risk score is an independent predictor of overall survival (OS). A nomogram integrating clinical characteristics and the risk score was created to predict 1-, 3-, and 5-year OS, assessed through calibration curves, ROC curves, and area under the ROC curve (AUC). Biological pathways were explored using GSEA and GSVA, while immune infiltrates were identified through CIBERSORT and TIMER.The eight-miRNA signature effectively predicted OS, recurrence-free survival, and disease-free survival. High-risk patients exhibited increased macrophage and neutrophil levels, indicative of a poor prognosis. High-risk patients, especially those with triple-negative breast cancer, had significantly worse outcomes. This risk score could inform personalized treatment strategies in breast cancer management.

Endoscopic ultrasonography (EUS) is commonly utilized in preoperative staging of esophageal cancer, however with additional pain and cost as well as adverse events. Meanwhile, the accuracy of EUS is highly depend on the training and practice of operators and not universally available. Different operators would lead to high inter-observer variability. Therefore, it is desirable to explore an alternative way to determine preoperative T stage in esophageal cancer. Whether conventional endoscopy possess the ability to predict EUS T stage has never been investigated yet. In current study, with the assistance of Artificial intelligence, we have developed a deep learning model to predict EUS T stage based on 9,714 images collected from 3,333 patients. ResNet-152 pre-trained on the ImageNet dataset was trained with the appropriate transfer learning and fine-tuning strategies on the conventional endoscopic images and their corresponding labels (e.g., T1, T2, T3, T4 and Normal). Meanwhile, augmentation strategies including rotation and flipping were performed to increase the number of images to improve the prediction accuracy. Finally, 4,382 T1, 243 T2, 3,985 T3, 1,102 T4, 14,302 controls images were obtained and split into training dataset, validation dataset and independent testing dataset with the ratio of 4:1:1. Our model could achieve a satisfied performance with an area under the receiver-operating curve (AUC) were 0.9767, 0.9637, 0.9597 and 0.9442 for T1, T2, T3 and T4, respectively in independent testing dataset. In conclusion, conventional gastroscopy combined with artificial intelligence have the great potential to predict EUS T stage.

Post-translational modifications (PTMs) diversify protein functions by adding/removing chemical groups to certain amino acid. As a newly-reported PTM, lysine β-hydroxybutyrylation (Kbhb) presents a new avenue to functional proteomics. Therefore, accurate and efficient prediction of Kbhb sites is imperative. However, the current experimental methods for identifying PTM sites are often expensive and time-consuming. Up to now, there is no computational method proposed for Kbhb sites detection. To this end, we present the first deep learning-based method, termed SLAM, to in silico identify lysine β-hydroxybutyrylation sites. The performance of SLAM is evaluated on both 5-fold cross-validation and independent test, achieving 0.890, 0.899, 0.907 and 0.923 in terms of AUROC values, on the general and species-specific independent test sets, respectively. As one example, we predicted the potential Kbhb sites in human S-adenosyl-L-homocysteine hydrolase, which is in agreement with experimentally-verified Kbhb sites. In summary, our method could enable accurate and efficient characterization of novel Kbhb sites that are crucial for the function and stability of proteins and could be applied in the structure-guided identification of other important PTM sites. The SLAM online service and source code is available at https://ai4bio.online/SLAM and https://github.com/Gabriel-QIN/SLAM, respectively.

Background Non-small cell cancer (NSCLC) patients with concomitant epidermal growth factor receptor (EGFR) and TP53 mutations have a poor prognosis with the treatment of tyrosine kinase inhibitors (TKIs), and may benefit from a combination regimen preferentially. The present study aims to compare the benefits of EGFR-TKIs and its combination with antiangiogenic drugs or chemotherapy in patients with NSCLC harboring EGFR and TP53 co-mutation in a real-life setting. Methods This retrospective analysis included 124 patients with advanced NSCLC having concomitant EGFR and TP53 mutations, who underwent next-generation sequencing prior to treatment. Patients were classified into the EGFR-TKI group and combination therapy group. The primary end point of this study was progression-free survival (PFS). The Kaplan–Meier (KM) curve was drawn to analyze PFS, and the differences between the groups were compared using the logarithmic rank test. Univariate and multivariate cox regression analysis was performed on the risk factors associated with survival. Results The combination group included 72 patients who received the regimen of EGFR-TKIs combined with antiangiogenic drugs or chemotherapy, while the EGFR-TKI monotherapy group included 52 patients treated with TKI only. The median PFS was significantly longer in the combination group than in the EGFR-TKI group (18.0 months; 95% confidence interval [CI]: 12.1–23.9 vs. 7.0 months; 95% CI: 6.1–7.9; p < 0.001) with greater PFS benefit in TP53 exon 4 or 7 mutations subgroup. Subgroup analysis showed a similar trend. The median duration of response was significantly longer in the combination group than in the EGFR-TKI group. Patients with 19 deletions or L858R mutations both achieved a significant PFS benefit with combination therapy versus EGFR-TKI alone. Conclusion Combination therapy had a higher efficacy than EGFR-TKI alone for patients with NSCLC having concomitant EGFR and TP53 mutations. Future prospective clinical trials are needed to determine the role of combination therapy for this patient population.

BACKGROUND:Upstream stimulatory factors (USFs) are members of the basic helix-loop-helix leucine zipper transcription factor family, including USF1, USF2, and USF3. The first two members have been well studied compared to the third member, USF3, which has received scarce attention in cancer research to date. Despite a recently reported association of its alteration with thyroid carcinoma, its expression has not been previously analyzed. METHODS:We comprehensively analyzed differential levels of USFs expression, genomic alteration, DNA methylation, and their prognostic value across different cancer types and the possible correlation with tumor-infiltrating immune cells and drug response by using different bioinformatics tools. RESULTS:Our findings established that USFs play an important role in cancers related to the urinary system and justify the necessity for further investigation. We implemented and offer a useful ShinyApp to facilitate researchers' efforts to inquire about any other gene of interest and to perform the analysis of drug response in a user-friendly fashion at http://zzdlab.com:3838/Drugdiscovery/.

BackgroundCytokines are involved in many inflammatory diseases and thus play an important role in tumor immune regulation. In recent years, researchers have found that breast cancer is not only related to genetic and environmental factors, but also to the chronic inflammation and immunity. However, the correlation between serum cytokines and blood tests indicators remain unclear.MethodsA total of 84 serum samples and clinicopathological data of breast cancer patients from Tianjin Cancer Institute & Hospital, Tianjin Medical University, Tianjin, P. R. China were collected. The expression levels of the 12 cytokines were detected by immunofluorescence method. Blood tests results were obtained from medical records. By stepwise Cox regression analysis, a cytokine-related gene signature was generated. Univariate and multivariate Cox regression were used to analyze the influence on the prognosis of patients. A nomogram was constructed to illustrate the cytokine-related riskscore predicting 5-year OS, which was further evaluated and validated by C-index and ROC curve. The correlation between the expression of cytokines in serum and other blood indicators was studied by using Spearman’s test.ResultsThe riskscore was calculated as IL-4×0.99069 + TNF-α×0.03683. Patients were divided into high and low risk groups according to the median riskscore, with the high-risk group has a shorter survival time by log-rank test (training set, P=0.017; validation set, P=0.013). Combined with the clinical characteristics, the riskscore was found to be an independent factor for predicting the OS of breast cancer patients in both training cohort (HR=1.2, P<0.01) and validation cohort (HR=1.6, P=0.023). The 5-year C-index and AUC of the nomogram were 0.78 and 0.68, respectively. IL-4 was further found to be negatively correlated with ALB.ConclusionIn summary, we have developed a nomogram based on two cytokines including IL-4 and TNF-α to predict OS of breast cancer and investigated their correlation with blood test indicators.

BackgroundMetastasis of breast cancer grows from the local invasion to the distant colonization. Blocking the local invasion step would be promising for breast cancer treatment. Our present study demonstrated AQP1 was a crucial target in breast cancer local invasion.MethodsMass spectrometry combined with bioinformatics analysis was used to identify AQP1 associated proteins ANXA2 and Rab1b. Co-immunoprecipitation, immunofluorescence assays and cell functional experiments were carried out to define the relationship among AQP1, ANXA2 and Rab1b and their re-localization in breast cancer cells. The Cox proportional hazards regression model was performed toward the identification of relevant prognostic factors. Survival curves were plotted by the Kaplan-Meier method and compared by the log-rank test.ResultsHere, we show that the cytoplasmic water channel protein AQP1, a crucial target in breast cancer local invasion, recruited ANXA2 from the cellular membrane to the Golgi apparatus, promoted Golgi apparatus extension, and induced breast cancer cell migration and invasion. In addition, cytoplasmic AQP1 recruited cytosolic free Rab1b to the Golgi apparatus to form a ternary complex containing AQP1, ANXA2, and Rab1b, which induced cellular secretion of the pro-metastatic proteins ICAM1 and CTSS. Cellular secretion of ICAM1 and CTSS led to the migration and invasion of breast cancer cells. Both in vivo assay and clinical analysis data confirmed above results.ConclusionsOur findings suggested a novel mechanism for AQP1-induced breast cancer local invasion. Therefore, targeting AQP1 offers promises in breast cancer treatment.

AIMS:Gastric adenocarcinoma with enteroblastic differentiation (GAED) is a rare subset of alpha-fetoprotein (AFP)-producing carcinomas with poor prognosis. However, the molecular features associated with the malignant potential of GEAD remain partially elucidated. METHODS AND RESULTS:In this study, the relationship between clinicopathological parameters and aggressive biological behaviour was analysed in 37 patients with GAED. The results showed that GAED tended to infiltrate the deep layer of the gastric wall and possessed more frequent vascular invasion than conventional gastric adenocarcinoma (CGA) (P < 0.001). All distant metastases were observed in the GAED group, not the CGA group (P < 0.001). High HER2 expression was found in nearly 24.32% of the informative cases, and none showed EBV-encoded RNA positivity or deficient mismatch repair. The most frequently mutated gene in GAED was p53. Programmed cell death-ligand 1 (PD-L1) immunostaining revealed 13 patients with a combined positive score (CPS) ≥ 5 (65%, 13 of 20). Thus, based on these molecular markers (immunostaining, in situ hybridisation and mutation analysis), GAED may be classified as a unique subgroup of the chromosomal instability subtype with HER2+ /EBV- /MSS/TP53+ /PD-L1+ . Next-generation sequencing analyses showed that mutations in the TOPI, ELOA and NOTCH3 genes were found only in GAED, and abnormally expressed genes in GAED were significantly enriched in hepatocellular carcinoma-, gland development-, and gastric cancer-related pathways. CONCLUSION:The HER2+ /EBV- /MSS/TP53+ /PD-L1+ profile and hepatocellular carcinoma-related pathways may be significant in the malignant potential of GAED. In addition to anti-HER2 therapy, immune check-point inhibitors may be an effective treatment option for patients with GAED.

Objective: The main reasons for the poor prognoses of pancreatic adenocarcinoma（PA） patients are rapid early-stage progression, advanced stage metastasis, and chemotherapy resistance. Identification of novel diagnostic and prognostic biomarkers of PA is therefore urgently needed.Methods: Three mRNA microarray datasets were obtained from the Gene Expression Omnibus database to select differentially expressed genes（DEGs）. Gene Ontology（GO） and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analyses for hub genes were performed using DAVID. Correlations between expression levels of hub genes and cancer-infiltrating immune cells were investigated by TIMER. Cox proportional hazard regression analyses were also performed. Serum hub genes were screened using the HPA platform and verified for diagnostic value using ELISAs.Results: We identified 59 hub genes among 752 DEGs. GO analysis indicated that these 59 hub genes were mainly involved in the defense response to viruses and the type I interferon signaling pathway. We also discovered that RSAD2 and SMC4 were associated with immune cell infiltration in the PA microenvironment. Additionally, DLGAP5 mR NA might be used as an independent risk factor for the prognoses of PA patients. Furthermore, the protein encoded by ISG15, which exists in peripheral blood, was validated as a potential diagnostic biomarker that distinguished PA patients from healthy controls（area under the curve: 0.902, 95% confidence interval: 0.819–0.961）.Conclusions: Our study suggested that RSAD2 and SMC4 were associated with immune cell infiltration in the PA microenvironment, while DLGAP5 mRNA expression might be an independent risk factor for the survival prognoses of PA patients. Moreover, ELISAs indicated that serum ISG15 could be a potential novel diagnostic biomarker for PA.

To develop a nomogram based on pretreatment dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to predict pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC) in patients with triple-negative breast cancer (TNBC). A total of 108 female patients with TNBC treated with neoadjuvant chemotherapy followed by surgery between January 2017 and October 2020 were enrolled. The patients were randomly divided into the primary cohort (n = 87) and validation cohort (n = 21) at a ratio of 4:1. The pretreatment DCE-MRI and clinicopathological features were reviewed and recorded. Univariate analysis and multivariate logistic regression analyses were used to determine the independent predictors of pCR in the primary cohort. A nomogram was developed based on the predictors, and the predictive performance of the nomogram was evaluated by the area under the receiver operating characteristic (ROC) curve (AUC). The validation cohort was used to test the predictive model. Tumor volume measured on DCE-MRI, time to peak (TTP), and androgen receptor (AR) status were identified as independent predictors of pCR. The AUCs of the nomogram were 0.84 (95% CI: 0.75–0.93) and 0.79 (95% CI: 0.59–0.99) in the primary cohort and validation cohort, respectively. Pretreatment DCE-MRI could predict pCR after NAC in patients with TNBC. The nomogram can be used to predict the probability of pCR and may help individualize treatment. • Pretreatment DCE-MRI findings can predict pathologic complete response (pCR) after neoadjuvant chemotherapy in patients with triple-negative breast cancer. • A nomogram based on the independent predictors of tumor volume measured on DCE-MRI, time to peak, and androgen receptor status could help personalized cancer treatment in TNBC patients.

Objective: To identify CT imaging biomarkers based on radiomic features for predicting brain metastases (BM) in patients with ALK-rearranged non-small cell lung cancer (NSCLC). Methods: NSCLC patients with pathologically confirmed ALK rearrangement from January 2014 to December 2020 in our hospital were enrolled retrospectively in this study. Finally, 77 patients were included according to the inclusion and exclusion criteria. Patients were divided into two groups: BM+ were those patients who were diagnosed with BM at baseline examination (n = 16) or within 1 year’s follow-up (n = 14), and BM− were those without BM followed up for at least 1 year (n = 47). Radiomic features were extracted from the pretreatment thoracic CT images. Sequential univariate logistic regression, LASSO regression, and backward stepwise logistic regression were used to select radiomic features and develop a BM-predicting model. Results: Five robust radiomic features were found to be independent predictors of BM. AUC for radiomics model was 0.828 (95% CI: 0.736–0.921), and when combined with clinical features, the AUC was increased (p = 0.017) to 0.909 (95% CI: 0.845–0.972). The individualized BM-predicting model incorporated with clinical features was visualized by the nomogram. Conclusion: Radiomic features extracted from pretreatment thoracic CT images have the potential to predict BM within 1 year after detection of the primary tumor in patients with ALK-rearranged NSCLC. The radiomics model incorporated with clinical features shows improved risk stratification for such patients.

Lipid metabolic reprogramming is one of the hallmarks of hepatocarcinogenesis and development. Therefore, lipid-metabolism-related genes may be used as potential biomarkers for hepatocellular carcinoma (HCC). This study aimed to screen for genes with dysregulated expression related to lipid metabolism in HCC and explored the clinical value of these genes. We screened differentially expressed proteins between tumorous and adjacent nontumorous tissues of hepatitis B virus (HBV)-related HCC patients using a Nanoscale Liquid Chromatography–Tandem Mass Spectrometry platform and combined it with transcriptomic data of lipid-metabolism-related genes from the GEO and HPA databases to identify dysregulated genes that may be involved in lipid metabolic processes. The potential clinical values of these genes were explored by bioinformatics online analysis tools (GEPIA, cBioPortal, SurvivalMeth, and TIMER). The expression levels of the secreted protein (angiopoietin-like protein 6, ANGPTL6) in serum were analyzed by ELISA. The ability of serum ANGPTL6 to diagnose early HCC was assessed by ROC curves. The results showed that serum ANGPTL6 could effectively differentiate between HBV-related early HCC patients with normal serum alpha-fetoprotein (AFP) levels and the noncancer group (healthy participants and chronic hepatitis B patients) (AUC = 0.717, 95% CI: from 0.614 to 0.805). Serum ANGPTL6 can be used as a potential second-line biomarker to supplement serum AFP in the early diagnosis of HBV-related HCC.

EDITORIAL article Front. Artif. Intell., 27 October 2022Sec. Medicine and Public Health Volume 5 - 2022 | https://doi.org/10.3389/frai.2022.1051182

At present, anti-angiogenic drugs (AADs) are widely used in the systemic treatment of hepatocellular carcinoma (HCC) or other types of cancer, and have achieved good anti-cancer effect, whereas treatment-related proteinuria can affect the routine use of AADs, which in turn abates the overall efficacy. Currently, most clinicians prescribe angiotensin-converting enzyme inhibitors (ACEIs) to alleviate proteinuria according to diabetic nephropathy guidelines or expert recommendations. However, the efficacy of ACEIs in reducing AAD-related proteinuria and its effect on the anticancer effect of AADs is unknown. Our clinical data showed that some HCC patients experienced tumor progression by ACEIs administration for the treatment of proteinuria caused by AADs. Here, we confirmed that in different tumor-bearing mouse models, ACEIs did not delay the appearance of proteinuria or alleviate proteinuria caused by AADs but compromised the anticancer efficacy of AADs. This effect is unrelated to the change in the VEGF signaling pathway. Our data showed that the combination of ACEIs and AADs flared the production of kidney-derived erythropoietin (EPO). In turn, EPO compromises the anti-angiogenic effects of AADs and decreases antitumor activity. In conclusion, for the treatment of proteinuria caused by AADs, ACEIs have no efficacy while also promoting AADs resistance. This finding is of great significance to guide clinical standardized management of side effects of anti-angiogenic therapy for cancer patients.

We developed two CNNs for predicting ubiquitination sites in Arabidopsis thaliana, demonstrated their competitive performance, analyzed amino acid physicochemical properties and the CNN structures, and predicted ubiquitination sites in Arabidopsis. As an important posttranslational protein modification, ubiquitination plays critical roles in plant physiology, including plant growth and development, biotic and abiotic stress, metabolism, and so on. A lot of ubiquitination site prediction models have been developed for human, mouse and yeast. However, there are few models to predict ubiquitination sites for the plant Arabidopsis thaliana. Based on this context, we proposed two convolutional neural network (CNN) based models for predicting ubiquitination sites in A. thaliana. The two models reach AUC (area under the ROC curve) values of 0.924 and 0.913 respectively in five-fold cross-validation, and 0.921 and 0.914 respectively in independent test, which outperform other models and demonstrate the competitive edge of them. We in-depth analyze the amino acid physicochemical properties in the neighboring sequence regions of the ubiquitination sites, and study the influence of the CNN structure to the prediction performance. Potential ubiquitination sites in the global Arbidopsis proteome are predicted using the two CNN models. To facilitate the community, the source code, training and test dataset, predicted ubiquitination sites in the Arbidopsis proteome are available at GitHub ( http://github.com/nongdaxiaofeng/CNNAthUbi ) for interest users.