Uncovering the Pre-Deterioration State during Disease Progression Based on Sample-Specific Causality Network Entropy (SCNE).

Complex diseases do not always follow gradual progressions. Instead, they may experience sudden shifts known as critical states or tipping points, where a marked qualitative change occurs. Detecting such a pivotal transition or pre-deterioration state holds paramount importance due to its association with severe disease deterioration. Nevertheless, the task of pinpointing the pre-deterioration state for complex diseases remains an obstacle, especially in scenarios involving high-dimensional data with limited samples, where conventional statistical methods frequently prove inadequate. In this study, we introduce an innovative quantitative approach termed sample-specific causality network entropy (SCNE), which infers a sample-specific causality network for each individual and effectively quantifies the dynamic alterations in causal relations among molecules, thereby capturing critical points or pre-deterioration states of complex diseases. We substantiated the accuracy and efficacy of our approach via numerical simulations and by examining various real-world datasets, including single-cell data of epithelial cell deterioration (EPCD) in colorectal cancer, influenza infection data, and three different tumor cases from The Cancer Genome Atlas (TCGA) repositories. Compared to other existing six single-sample methods, our proposed approach exhibits superior performance in identifying critical signals or pre-deterioration states. Additionally, the efficacy of computational findings is underscored by analyzing the functionality of signaling biomarkers.

MRI-based deep learning and radiomics for prediction of occult cervical lymph node metastasis and prognosis in early-stage oral and oropharyngeal squamous cell carcinoma: a diagnostic study.

INTRODUCTION: The incidence of occult cervical lymph node metastases (OCLNM) is reported to be 20%-30% in early-stage oral cancer and oropharyngeal cancer. There is a lack of an accurate diagnostic method to predict occult lymph node metastasis and to help surgeons make precise treatment decisions. AIM: To construct and evaluate a preoperative diagnostic method to predict occult lymph node metastasis (OCLNM) in early-stage oral and oropharyngeal squamous cell carcinoma (OC and OP SCC) based on deep learning features (DLFs) and radiomics features. METHODS: A total of 319 patients diagnosed with early-stage OC or OP SCC were retrospectively enrolled and divided into training, test and external validation sets. Traditional radiomics features and DLFs were extracted from their MRI images. The least absolute shrinkage and selection operator (LASSO) analysis was employed to identify the most valuable features. Prediction models for OCLNM were developed using radiomics features and DLFs. The effectiveness of the models and their clinical applicability were evaluated using the area under the curve (AUC), decision curve analysis (DCA) and survival analysis. RESULTS: Seventeen prediction models were constructed. The Resnet50 deep learning (DL) model based on the combination of radiomics and DL features achieves the optimal performance, with AUC values of 0.928 (95% CI: 0.881-0.975), 0.878 (95% CI: 0.766-0.990), 0.796 (95% CI: 0.666-0.927) and 0.834 (95% CI: 0.721-0.947) in the training, test, external validation set1 and external validation set2, respectively. Moreover, the Resnet50 model has great prediction value of prognosis in patients with early-stage OC and OP SCC. CONCLUSION: The proposed MRI-based Resnet50 deep learning model demonstrated high capability in diagnosis of OCLNM and prognosis prediction in the early-stage OC and OP SCC. The Resnet50 model could help refine the clinical diagnosis and treatment of the early-stage OC and OP SCC.

An uncertainty-based interpretable deep learning framework for predicting breast cancer outcome.

BACKGROUND: Predicting outcome of breast cancer is important for selecting appropriate treatments and prolonging the survival periods of patients. Recently, different deep learning-based methods have been carefully designed for cancer outcome prediction. However, the application of these methods is still challenged by interpretability. In this study, we proposed a novel multitask deep neural network called UISNet to predict the outcome of breast cancer. The UISNet is able to interpret the importance of features for the prediction model via an uncertainty-based integrated gradients algorithm. UISNet improved the prediction by introducing prior biological pathway knowledge and utilizing patient heterogeneity information. RESULTS: The model was tested in seven public datasets of breast cancer, and showed better performance (average C-index = 0.691) than the state-of-the-art methods (average C-index = 0.650, ranged from 0.619 to 0.677). Importantly, the UISNet identified 20 genes as associated with breast cancer, among which 11 have been proven to be associated with breast cancer by previous studies, and others are novel findings of this study. CONCLUSIONS: Our proposed method is accurate and robust in predicting breast cancer outcomes, and it is an effective way to identify breast cancer-associated genes. The method codes are available at: https://github.com/chh171/UISNet .

The management of non-culprit vessel(s) in patients with unstable angina/non-ST elevation myocardial infarction and chronic kidney dysfunction.

BACKGROUND AND AIMS: The clinical effects of multivessel interventions in patients with unstable angina/non-ST-segment elevation myocardial infarction (UA/NSTEMI), multivessel disease (MVD) and chronic kidney disease (CKD) remain uncertain. This study aimed to investigate the safety and effectiveness of intervention in non-culprit lession(s) among this cohort. METHODS: We consecutively included patients diagnosed with UA/NSTEMI, MVD and CKD between January 2008 and December 2018 at our centre. After successful percutaneous coronary intervention (PCI), we compared 48-month overall mortality between those undergoing multivessel PCI (MV-PCI) through a single-procedure or staged-procedure approach and culprit vessel-only PCI (CV-PCI) after 1:1 propensity score matching. We conducted stratified analyses and tests for interaction to investigate the modifying effects of critical covariates. Additionally, we recorded the incidence of contrast-induced nephropathy (CIN) to assess the perioperative safety of the two treatment strategies. RESULTS: Of the 749 eligible patients, 271 pairs were successfully matched. Those undergoing MV-PCI had reduced all-cause mortality (hazard ratio (HR): 0.67, 95% confidence interval (CI): 0.48-0.67). Subgroup analysis showed that those with advanced CKD (estimated glomerular filtration rate (eGFR) ≤ 30 mL/min/1.73 m2 ) could not benefit from MV-PCI (P = 0.250), and the survival advantage also tended to diminish in diabetes (P interaction < 0.01; HR = 0.95, 95% CI = 0.65-1.45). Although the staged-procedure approach (N = 157) failed to bring additional survival benefits compared to single-procedure MV-PCI (N = 290) (P = 0.460), it showed a tendency to decrease the death risk. CIN risks in MV-PCI and CV-PCI groups were not significantly different (risk ratio = 1.60, 95% CI = 0.94-2.73). CONCLUSION: Among patients with UA/NSTEMI and non-diabetic CKD and an eGFR > 30 mL/min/1.73 m2 , MV-PCI was associated with a reduced risk of long-term death but did not increase the incidence of CIN during the management of MVD compared to CV-PCI. And staged procedures might be a preferable option over single-procedure MV-PCI.

AFEI: adaptive optimized vertical federated learning for heterogeneous multi-omics data integration.

Vertical federated learning has gained popularity as a means of enabling collaboration and information sharing between different entities while maintaining data privacy and security. This approach has potential applications in disease healthcare, cancer prognosis prediction, and other industries where data privacy is a major concern. Although using multi-omics data for cancer prognosis prediction provides more information for treatment selection, collecting different types of omics data can be challenging due to their production in various medical institutions. Data owners must comply with strict data protection regulations such as European Union (EU) General Data Protection Regulation. To share patient data across multiple institutions, privacy and security issues must be addressed. Therefore, we propose an adaptive optimized vertical federated-learning-based framework adaptive optimized vertical federated learning for heterogeneous multi-omics data integration (AFEI) to integrate multi-omics data collected from multiple institutions for cancer prognosis prediction. AFEI enables participating parties to build an accurate joint evaluation model for learning more information related to cancer patients from different perspectives, based on the distributed and encrypted multi-omics features shared by multiple institutions. The experimental results demonstrate that AFEI achieves higher prediction accuracy (6.5% on average) than using single omics data by utilizing the encrypted multi-omics data from different institutions, and it performs almost as well as prognosis prediction by directly integrating multi-omics data. Overall, AFEI can be seen as an efficient solution for breaking down barriers to multi-institutional collaboration and promoting the development of cancer prognosis prediction.

Assembly of DNA triangular pyramid frustum for ultrasensitive quantification of exosomal miRNA.

Early screening of biomarkers benefits therapy and prognosis of cancers. MiRNAs encapsulated in tumor-derived exosomes are emerging biomarkers for early diagnosis of cancers. Nevertheless, traditional methods suffer certain drawbacks, which hamper their wide applications. In this contribution, we have developed a convenient electrochemical approach for quantification of exosomal miRNA based on the assembly of DNA triangular pyramid frustum (TPF) and strand displacement amplification. Four single-stranded DNA helps the formation of primary DNA triangle with three thiols for gold electrode immobilization at the bottom and three amino groups on overhangs for the capture of silver nanoparticles. On the other hand, target miRNA induced strand displacement reaction produces abundant specific DNA strands, which help the DNA structural transition from triangle to TPF. Amino groups are thus hidden and the declined silver stripping current can be used for the evaluation of target miRNA concentration. This biosensor exhibits excellent analytical performances and successfully achieves analysis of exosomal miRNAs from cells and clinical serum samples.

Current Progress on Neuroinflammation-mediated Postoperative Cognitive Dysfunction: An Update.

Postoperative cognitive dysfunction (POCD) is a common complication of the central nervous system (CNS) in elderly patients after surgery, showing cognitive changes such as decreased learning and memory ability, impaired concentration, and even personality changes and decreased social behavior ability in severe cases. POCD may appear days or weeks after surgery and persist or even evolve into Alzheimer's disease (AD), exerting a significant impact on patients' health. There are many risk factors for the occurrence of POCD, including age, surgical trauma, anesthesia, neurological diseases, etc. The level of circulating inflammatory markers increases with age, and elderly patients often have more risk factors for cardiovascular diseases, resulting in an increase in POCD incidence in elderly patients after stress responses such as surgical trauma and anesthesia. The current diagnostic rate of POCD is relatively low, which affects the prognosis and increases postoperative complications and mortality. The pathophysiological mechanism of POCD is still unclear, however, central nervous inflammation is thought to play a critical role in it. The current review summarizes the related studies on neuroinflammation-mediated POCD, such as the involvement of key central nervous cells such as microglia and astrocytes, proinflammatory cytokines such as TNF-α and IL-1ß, inflammatory signaling pathways such as PI3K/Akt/mTOR and NF-κB. In addition, multiple predictive and diagnostic biomarkers for POCD, the risk factors, and the positive effects of anti-inflammatory therapy in the prevention and treatment of POCD have also been reviewed. The exploration of POCD pathogenesis is helpful for its early diagnosis and long-term treatment, and the intervention strategies targeting central nervous inflammation of POCD are of great significance for the prevention and treatment of POCD.

Ultrasensitive miRNA biosensor amplified by ladder hybridization chain reaction on triangular prism structured DNA.

Accurate and sensitive analysis of biomarkers is a promising way to provide comprehensive physio-pathological information that is significant for early diagnosis of certain diseases. miRNA is a type of noncoding small RNAs which are involved in the regulation of a number of cellular processes. It has been regarded as an important tumor biomarker. Herein, we have constructed a three-dimensional DNA layer on electrode interface and achieved ladder hybridization chain reaction strategy for the enrichment of electrochemical signals. In addition, duplex-specific nuclease catalyzed amplification is previously performed on magnetic nanocomposites, which further improves the sensitivity for the detection of target miRNA initiator. This approach shows great molecular recognition efficiency as well as cascade signal amplification. The analytical performances are superior. In addition, the identification of cancer cell types according to target biomarker information is achieved and the testing results in clinical serum samples further demonstrate its great potential utility for diagnosis.

Tetrahedral DNA Supported Walking Nanomachine for Ultrasensitive miRNA Detection in Cancer Cells and Serums.

A three-dimensional DNA tetrahedral nanostructure is constructed to support a walker strand on top and multiple track strands around it via the assembly of triplex-forming oligonucleotide (TFO). This design facilitates the regeneration of the sensing interface by simply adjusting pH conditions. On the basis of the tetrahedral DNA supported walking nanomachine, ultrasensitive electrochemical analysis of miRNA (miR-141) is achieved. Target miRNA assists the formation of three-way junction nanostructure. It contains a duplex region (hybridized by track and walker strands) that could be specially recognized and digested by certain nicking endonuclease. As a result, walker strand and target miRNA are released and move around the attached tracks for continuous cleavage reactions, releasing a larger number of signal reporters. By measuring the variation of signal responses, ultrasensitive analysis of miRNA is achieved. The limit of detection (LOD) is calculated to be 4.9 aM, which is rather low. In addition, the proposed method is successfully applied for the detection of miRNA in cell and serum samples, which could distinguish pathological information from healthy controls.

DNA-MnO2 Nanoconjugates Investigation and Application for Electrochemical Polymerase Chain Reaction.

Manganese dioxide (MnO2) nanosheets are emerging for biomedical applications with excellent physical and chemical properties. Adsorption of DNA on MnO2 is important for biosensing, bioimaging, and therapy. Nevertheless, current fundamental understanding about the interaction is preliminary. Herein, UV-vis absorption spectra are applied to systematically explore the biointerfacial interaction between DNA and MnO2 with the factors of salt concentration, pH value, temperature, DNA concentration, and length. The results offer important fundamental insights into the investigation of DNA-MnO2 nanocomposites. Meanwhile, the optimal parameters are applied to construct a screen-printed electrode (SPE) modified with polymerase chain reaction (PCR) primer-decorated MnO2 nanosheets. An electrochemical PCR system is then developed for ultrasensitive detection of circulating tumor DNA (ctDNA). The limit of detection is determined to be 0.1 fM, and high selectivity is demonstrated. Combining the merits of SPE, DNA-MnO2 nanosheets, and an amplified reaction, this developed strategy shows great promise in bioanalysis, clinical disease diagnosis, and biomedicine applications.

DNA Hairpins and Dumbbell-Wheel Transitions Amplified Walking Nanomachine for Ultrasensitive Nucleic Acid Detection.

Nucleic acids, including circulating tumor DNA (ctDNA), microRNA, and virus DNA/RNA, have been widely applied as potential disease biomarkers for early clinical diagnosis. In this study, we present a concept of DNA nanostructures transitions for the construction of DNA bipedal walking nanomachine, which integrates dual signal amplification for direct nucleic acid assay. DNA hairpins transition is developed to facilitate the generation of multiple target sequences; meanwhile, the subsequent DNA dumbbell-wheel transition is controlled to achieve the bipedal walker, which cleaves multiple tracks around electrode surface. Through combination of strand displacement reaction and digestion cycles, DNA monolayer at the electrode interface could be engineered and target-induced signal variation is realized. In addition, pH-assisted detachable intermolecular DNA triplex design is utilized for the regeneration of electrochemical biosensor. The high consistency between this work and standard quantitative polymerase chain reaction is validated. Moreover, the feasibilities of this biosensor to detect ctDNA and SARS-CoV-2 RNA in clinical samples are demonstrated with satisfactory accuracy and reliability. Therefore, the proposed approach has great potential applications for nucleic acid based clinical diagnostics.

Ratiometric Electrochemical Switch for Circulating Tumor DNA through Recycling Activation of Blocked DNAzymes.

Circulating tumor DNA (ctDNA) serves as a powerful noninvasive and viable biomarker for the diagnosis of cancers. The abundance of ctDNA in patients with advanced stages is significantly higher than that in patients with early stages. Herein, a ratiometric electrochemical biosensor for the detection of ctDNA is developed by smart design of DNA probes and recycles of DNAzyme activation. The conformational variation of DNA structures leads to the changes of two types of electrochemical species. This enzyme-free sensing strategy promotes excellent amplification efficiency upon target recognition. The obtained results assure good analytical performances and a limit of detection as low as 25 aM is achieved. Additionally, this method exhibits outstanding selectivity and great application prospects in biological sample analysis.

Cancer survival prognosis with Deep Bayesian Perturbation Cox Network.

BACKGROUND: The Cox proportional hazards model with neural networks is widely used to accurately predict survival outcome for choosing cancer treatment strategies. Although this method has shown outstanding performance in many tasks, it has encountered challenges when dealing with high-dimensional datasets. In this study, we point out that the Cox network has estimation bias in processing such datasets with a large number of censored samples. The estimation bias is composed of censored estimation bias and variance estimation bias, which limit the prediction performance of the model. In order to correct this bias, this paper proposes the Deep Bayesian Perturbation Cox Network (DBP), which introduces Bayesian prior knowledge about censored samples to optimize the training process of the neural network. Specifically, the model uses a sampling module called Bayesian Perturbation to approximate the prior knowledge, which can be used as a component for other Cox-based neural networks. RESULTS: The comparison between DBP and the previous model in different kinds of genomic datasets demonstrates that our model has made significant improvements over previous state-of-the-art methods. In addition, the simulation experiments are performed to illustrate how the DBP method addresses the bias caused by Cox Network. In the case study, based on the predicted risks in BRCA data from TCGA, we identify 400 differential expressed genes and 20 KEGG pathways that are associated with breast cancer prognosis, among which 65% of the top 20 genes have been proved by literature review. CONCLUSION: Overall, these results demonstrate that our proposed method is advanced and robust in datasets with a large proportion of censored samples. Besides, it can guide to discover disease-related genes.

Lingbao Huxin Pill Alleviates Apoptosis and Inflammation at Infarct Border Zone through SIRT1-Mediated FOXO1 and NF- κ B Pathways in Rat Model of Acute Myocardial Infarction.

OBJECTIVE: To investigate whether Lingbao Huxin Pill (LBHX) protects against acute myocardial infarction (AMI) at the infarct border zone (IBZ) of myocardial tissue by regulating apoptosis and inflammation through the sirtuin 1 (SIRT1)-mediated forkhead box protein O1 (FOXO1) and nuclear factor-κ B (NF-κ B) signaling pathways. METHODS: Six-week-old Wistar rats with normal diet were randomized into the sham, the model, Betaloc (0.9 mg/kg daily), LBHX-L (0.45 mg/kg daily), LBHX-M (0.9 mg/kg daily), LBHX-H (1.8 mg/kg daily), and LBHX+EX527 (0.9 mg/kg daily) groups according to the method of random number table, 13 in each group. In this study, left anterior descending coronary artery (LADCA) ligation was performed to induce an AMI model in rats. The myocardial infarction area was examined using a 2,3,5-triphenyltetrazolium chloride solution staining assay. A TdT-mediated dUTP nick-end labeling (TUNEL) assay was conducted to assess cardiomyocyte apoptosis in the IBZ. The histopathology of myocardial tissue at the IBZ was assessed with Heidenhain, Masson and hematoxylineosin (HE) staining assays. The expression levels of tumor necrosis factor α (TNF-α), interleukin (IL)-6, IL-1 ß, and intercellular adhesion molecule-1 were measured using enzyme-linked immunosorbent assays (ELISAs). The mRNA expressions of SIRT1 and FOXO1 were detected by real-time qPCR (RT-qPCR). The protein expressions of SIRT1, FOXO1, SOD2, BAX and NF- κ B p65 were detected by Western blot analysis. RESULTS: The ligation of the LADCA successfully induced an AMI model. The LBHX pretreatment reduced the infarct size in the AMI rats (P<0.01). The TUNEL assay revealed that LBHX inhibited cardiomyocyte apoptosis at the IBZ. Further, the histological examination showed that the LBHX pretreatment decreased the ischemic area of myocardial tissue (P<0.05), myocardial interstitial collagen deposition (P<0.05) and inflammation at the IBZ. The ELISA results indicated that LBHX decreased the serum levels of inflammatory cytokines in the AMI rats (P<0.05 or P<0.01). Furthermore, Western blot analysis revealed that the LBHX pretreatment upregulated the protein levels of SIRT1, FOXO1 and SOD2 (P<0.05) and downregulated NF- κ B p65 and BAX expressions (P<0.05). The RT-qPCR results showed that LBHX increased the SIRT1 mRNA and FOXO1 mRNA levels (P<0.05). These protective effects, including inhibiting apoptosis and alleviating inflammation in the IBZ, were partially abolished by EX527, an inhibitor of SIRT1. CONCLUSION: LBHX could protect against AMI by suppressing apoptosis and inflammation in AMI rats and the SIRT1-mediated FOXO1 and NF- κ B signaling pathways were involved in the cardioprotection effect of LBHX.

Therapeutic Mechanism and Clinical Observation of Traditional Chinese Medicine Combined with Interventional Recanalization for Tubal Infertility.

To observe the clinical effect of traditional Chinese medicine (TCM) combined with interventional recanalization therapy in the treatment of tubal obstructive infertility, first, different treatment approaches were used on rabbits, and transmission electron microscopy (TEM) indicated that interventional recanalization combined with TCM can significantly ameliorate the pathological condition of the fallopian tube after treatment. Moreover, ELISA disclosed that the treatment could significantly reduce the levels of interleukin-1ß (IL-1ß), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α) and increase the expression of interleukin-10 (IL-10), which demonstrated that TCM therapy can help against inflammation of the fallopian tubes. PCR array analysis revealed that BMP4, BMPR1A, SMAD2, SMAD3, SMAD4, and KLF10 expressions were upregulated, and SMAD7 expression was downregulated, proving that combined treatment could influence gene expression in the TGF-ß family and further regulate the secretion of proteins in SMADs. In addition, a clinical study recorded the fallopian tube patency rate of 165 patients after 12 months. The recanalization rates in the two groups were 81.9% and 53.1%, with the higher rates in the combined medicine enema group. All these findings implied that interventional recanalization combined with TCM preparation has a stronger effect. The mechanism probably involves effects on the expression of genes in the TGF-ß/SMAD and BMP/SMAD signaling pathways, with simultaneous regulation of inflammatory factors, thereby improving the ovarian environment and increasing pregnancy rates.

An Adaptive Transfer-Learning-Based Deep Cox Neural Network for Hepatocellular Carcinoma Prognosis Prediction.

BACKGROUND: Predicting hepatocellular carcinoma (HCC) prognosis is important for treatment selection, and it is increasingly interesting to predict prognosis through gene expression data. Currently, the prognosis remains of low accuracy due to the high dimension but small sample size of liver cancer omics data. In previous studies, a transfer learning strategy has been developed by pre-training models on similar cancer types and then fine-tuning the pre-trained models on the target dataset. However, transfer learning has limited performance since other cancer types are similar at different levels, and it is not trivial to balance the relations with different cancer types. METHODS: Here, we propose an adaptive transfer-learning-based deep Cox neural network (ATRCN), where cancers are represented by 12 phenotype and 10 genotype features, and suitable cancers were adaptively selected for model pre-training. In this way, the pre-trained model can learn valuable prior knowledge from other cancer types while reducing the biases. RESULTS: ATRCN chose pancreatic and stomach adenocarcinomas as the pre-training cancers, and the experiments indicated that our method improved the C-index of 3.8% by comparing with traditional transfer learning methods. The independent tests on three additional HCC datasets proved the robustness of our model. Based on the divided risk subgroups, we identified 10 HCC prognostic markers, including one new prognostic marker, TTC36. Further wet experiments indicated that TTC36 is associated with the progression of liver cancer cells. CONCLUSION: These results proved that our proposed deep-learning-based method for HCC prognosis prediction is robust, accurate, and biologically meaningful.

Propofol Affects EGF's Activity in Intestinal Cell by Down-Regulating EGFR-Mediated Intracellular Signaling.

Propofol is a commonly used anesthetic drug in clinic. In recent years, a series of non-anesthetic effects of propofol have been discovered. Studies have shown that propofol has many effects on the intestine. Epidermal growth factor (EGF) is one of the most important growth factors that could regulate intestinal growth and development. In the current study, we studied the effect of protocol on the biological activity of EGF on intestinal tissue and cell models. Through flow cytometry, indirect immunofluorescence and Western-blot and other technologies, it was found that propofol reduced the activity of EGF on intestinal cells, which inhibited EGF-induced intestinal cell proliferation and changed the cell behavior of EGF. To further explore the potential mechanism by which propofol down-regulated epidermal growth factor receptor (EGFR)-induced signaling, we carried out a series of related experiments, and found that propofol may inhibit the proliferation of intestinal cells by inhibiting the EGFR-mediated intracellular signaling pathway. The current research will lay the theoretical and experimental basis for further study of the effect of propofol on the intestine.

Detecting lncRNA-Cancer Associations by Combining miRNAs, Genes, and Prognosis With Matrix Factorization.

Motivation: Long non-coding RNAs (lncRNAs) play important roles in cancer development. Prediction of lncRNA-cancer association is necessary for efficiently discovering biomarkers and designing treatment for cancers. Currently, several methods have been developed to predict lncRNA-cancer associations. However, most of them do not consider the relationships between lncRNA with other molecules and with cancer prognosis, which has limited the accuracy of the prediction. Method: Here, we constructed relationship matrices between 1,679 lncRNAs, 2,759 miRNAs, and 16,410 genes and cancer prognosis on three types of cancers (breast, lung, and colorectal cancers) to predict lncRNA-cancer associations. The matrices were iteratively reconstructed by matrix factorization to optimize low-rank size. This method is called detecting lncRNA cancer association (DRACA). Results: Application of this method in the prediction of lncRNAs-breast cancer, lncRNA-lung cancer, and lncRNA-colorectal cancer associations achieved an area under curve (AUC) of 0.810, 0.796, and 0.795, respectively, by 10-fold cross-validations. The performances of DRACA in predicting associations between lncRNAs with three kinds of cancers were at least 6.6, 7.2, and 6.9% better than other methods, respectively. To our knowledge, this is the first method employing cancer prognosis in the prediction of lncRNA-cancer associations. When removing the relationships between cancer prognosis and genes, the AUCs were decreased 7.2, 0.6, and 5% for breast, lung, and colorectal cancers, respectively. Moreover, the predicted lncRNAs were found with greater numbers of somatic mutations than the lncRNAs not predicted as cancer-associated for three types of cancers. DRACA predicted many novel lncRNAs, whose expressions were found to be related to survival rates of patients. The method is available at https://github.com/Yanh35/DRACA.

Integrating multi-omics data through deep learning for accurate cancer prognosis prediction.

BACKGROUND: Genomic information is nowadays widely used for precise cancer treatments. Since the individual type of omics data only represents a single view that suffers from data noise and bias, multiple types of omics data are required for accurate cancer prognosis prediction. However, it is challenging to effectively integrate multi-omics data due to the large number of redundant variables but relatively small sample size. With the recent progress in deep learning techniques, Autoencoder was used to integrate multi-omics data for extracting representative features. Nevertheless, the generated model is fragile from data noises. Additionally, previous studies usually focused on individual cancer types without making comprehensive tests on pan-cancer. Here, we employed the denoising Autoencoder to get a robust representation of the multi-omics data, and then used the learned representative features to estimate patients' risks. RESULTS: By applying to 15 cancers from The Cancer Genome Atlas (TCGA), our method was shown to improve the C-index values over previous methods by 6.5% on average. Considering the difficulty to obtain multi-omics data in practice, we further used only mRNA data to fit the estimated risks by training XGboost models, and found the models could achieve an average C-index value of 0.627. As a case study, the breast cancer prognosis prediction model was independently tested on three datasets from the Gene Expression Omnibus (GEO), and shown able to significantly separate high-risk patients from low-risk ones (C-index>0.6, p-values<0.05). Based on the risk subgroups divided by our method, we identified nine prognostic markers highly associated with breast cancer, among which seven genes have been proved by literature review. CONCLUSION: Our comprehensive tests indicated that we have constructed an accurate and robust framework to integrate multi-omics data for cancer prognosis prediction. Moreover, it is an effective way to discover cancer prognosis-related genes.

Knockdown of CENPW Inhibits Hepatocellular Carcinoma Progression by Inactivating E2F Signaling.

AIM: This study aimed to evaluate the effects of centromere protein W (CENPW, also known as CUG2) in hepatocellular carcinoma (HCC). METHODS: CENPW expression in HCC tissues and cells was detected by RT-qPCR assay. CCK-8 and colony formation assay were used to assess cell proliferation. Wound healing and Transwell assay was used to detect cell migration and invasion, respectively. The flow cytometry was used to analyze the cell cycle distribution and apoptosis. RESULTS: CENPW expression was upregulated in HCC tissues and cells. Knockdown of CENPW inhibited cell proliferation, migration, and invasion and induced the G0/G1 phase arrest and cell apoptosis in HCC cells, which might involve the E2F signaling regulation. CONCLUSION: CENPW acted as an oncogenic role in HCC progression via activation E2F signaling. Our findings may provide new insights into the studying mechanisms of HCC.