MSI-DTI: predicting drug-target interaction based on multi-source information and multi-head self-attention.

Identifying drug-target interactions (DTIs) holds significant importance in drug discovery and development, playing a crucial role in various areas such as virtual screening, drug repurposing and identification of potential drug side effects. However, existing methods commonly exploit only a single type of feature from drugs and targets, suffering from miscellaneous challenges such as high sparsity and cold-start problems. We propose a novel framework called MSI-DTI (Multi-Source Information-based Drug-Target Interaction Prediction) to enhance prediction performance, which obtains feature representations from different views by integrating biometric features and knowledge graph representations from multi-source information. Our approach involves constructing a Drug-Target Knowledge Graph (DTKG), obtaining multiple feature representations from diverse information sources for SMILES sequences and amino acid sequences, incorporating network features from DTKG and performing an effective multi-source information fusion. Subsequently, we employ a multi-head self-attention mechanism coupled with residual connections to capture higher-order interaction information between sparse features while preserving lower-order information. Experimental results on DTKG and two benchmark datasets demonstrate that our MSI-DTI outperforms several state-of-the-art DTIs prediction methods, yielding more accurate and robust predictions. The source codes and datasets are publicly accessible at https://github.com/KEAML-JLU/MSI-DTI.

Human lip vermilion: Physiology and age-related changes.

BACKGROUND: The human lip vermilion, also known as the red lip, is important to the quality of life and has long attracted the attention of cosmetic researchers. However, there is limited existing literature on the physiological characteristics and age-related alterations in the human vermilion. OBJECTIVE: We aim to provide an overview of the physiological characteristics and age-related alterations in the human vermilion. METHODS: This article is a result of previous research. We conducted a literature search using various academic databases such as Google Scholar, Web of Science, and PubMed. Our findings provided a comprehensive understanding of the physiological characteristics and age-related changes of the human lip vermilion. RESULTS: The human lip vermilion has a unique structure and physiological characteristics, and during the aging process, a few changes may occur in the human lip vermilion. CONCLUSION: Understanding the human lip vermilion's physiological characteristics and age-related changes can provide key information for the future innovation of lip vermilion care products. Further investigations are necessary to reach a consensus on the physiological characteristics and age-related alterations in the human vermilion.

Lactic Acid Chemical Peeling in Skin Disorders.

Lactic acid is the most widely occurring natural organic acid in nature. It not only exhibits mild and safe properties but also possesses multiple physiological activities, such as antibacterial effects, immune regulation, and promotion of wound healing, making it one of the most popular chemical peeling agents. Chemical peels are commonly used in the field of aesthetic dermatology as a non-invasive therapeutic approach. This research aims to provide valuable references for clinical dermatologists by summarizing the characteristics of lactic acid, elucidating its mechanism of action in peeling, and investigating the clinical applications of this compound. Furthermore, it anticipates the potential for lactic acid to be the most suitable chemical peeling agent for Chinese skin.

Potential role of inflammaging mediated by the complement system in enlarged facial pores.

BACKGROUND: Enlarged facial pores are a common cosmetic concern of the skin, rather than a disease, and have not received much attention from dermatologists in recent years. Consequently, progress in understanding their pathogenesis has been limited, and current cosmetic solutions have limitations. Given that the complement system has regained interest as a key player in chronic inflammatory skin conditions, various mechanisms involving this system are being investigated. OBJECTIVE: We aimed to shed light on the mechanism underlying enlarged facial pores by examining the role of the complement system in skin. METHODS: We conducted a comprehensive literature search utilizing various academic databases including PubMed, Web of Science, and Google Scholar. Employing keywords such as "complement system," "inflammation," "facial pores," "enlarged," and "mechanisms," we compiled a selection of relevant studies. These studies provided a comprehensive understanding of the intricate mechanisms underlying the relationship between the "complement system" and "inflammation" within the context of facial pore enlargement. RESULTS: Our findings suggest that inflammaging mediated by complement activation may be a critical player in the formation of enlarged facial pores. Specifically, overactivation of the complement system leading to the accumulation of complement fragments could be a major contributor to this process. Notably, the complement system in skin may be involved in a range of skin issues, including aging. CONCLUSION: Modulating the complement system presents a promising avenue for future research in improving skin health. Further basic and clinical research is necessary to validate these findings, but we hope that this study can serve as a theoretical foundation for the development of targeted cosmetics.

Muc2 mucin O-glycosylation interacts with enteropathogenic Escherichia coli to influence the development of ulcerative colitis based on the NF-kB signaling pathway.

BACKGROUND: Ulcerative colitis (UC) is a chronic inflammatory disease of the intestine characterized by a compromised intestinal epithelial barrier. Mucin glycans are crucial in preserving barrier function during bacterial infections, although the underlying mechanisms remain largely unexplored. METHODS: A cohort comprising 15 patients diagnosed with UC and 15 healthy individuals was recruited. Stool samples were collected to perform 16S rRNA gene sequencing, while biopsy samples were subjected to nanocapillary liquid chromatography-tandem mass spectrometry (nanoLC-MS/MS) to assess O-glycosylation. Gene expression was evaluated through qPCR analysis and Western blotting. Furthermore, animal experiments were conducted to investigate the effects of Escherichia coli and/or O-glycan inhibitor benzyl-α-GalNAc on the development of colitis in mice. RESULTS: Our findings revealed that the mucus barrier was disrupted during the early stages of UC, while the MUC2 protein content remained unaltered. Additionally, a noteworthy reduction in the O-glycosylation of MUC2 was observed, along with significant changes in the intestinal microbiota during the early stages of UC. These changes included a decrease in intestinal species richness and an increase in the abundance of Escherichia coli (E. coli). Moreover, subsequent to the administration of galactose or O-glycan inhibitor to intestinal epithelial cells, it was observed that the cell culture supernatant had the ability to modify the proliferation and adhesive capacity of E. coli. Furthermore, when pathogenic E. coli or commensal E. coli were cocultured with intestinal epithelium, both strains elicited activation of the NF-KB signaling pathway in epithelial cells and facilitated the expression of serine protease in comparison to the untreated control. Consistently, the inhibition of O-glycans has been observed to enhance the pathogenicity of E. coli in vivo. Furthermore, a correlation has been established between the level of O-glycans and the development of ulcerative colitis. Specifically, a reduction in the O-glycan content of MUC2 cells has been found to increase the virulence of E. coli, thereby compromising the integrity of the intestinal epithelial barrier. CONCLUSIONS: Together, there exist complex interactions between the intestinal epithelium, O-glycans, and the intestinal microbiota, which may inform the development of novel therapeutic strategies for the treatment of ulcerative colitis.

Efficacy and safety of allogeneic umbilical cord-derived mesenchymal stem cells for the treatment of complex perianal fistula in Crohn's disease: a pilot study.

OBJECTIVES: The aim of the study was to evaluate the efficacy and safety of allogeneic umbilical cord-derived mesenchymal stem cells (TH-SC01) for complex perianal fistula in patients with Crohn's disease (CD). METHODS: This was an open-label, single-arm clinical trial conducted at Jinling Hospital. Adult patients with complex treatment-refractory CD perianal fistulas (pfCD) were enrolled and received a single intralesional injection of 120 million TH-SC01 cells. Combined remission was defined as an absence of suppuration through an external orifice, complete re-epithelization, and absence of collections larger than 2 cm measured by magnetic resonance imaging (MRI) at 24 weeks after cell administration. RESULTS: A total of 10 patients were enrolled. Six patients (60.0%) achieved combined remission at 24 weeks. The number of draining fistulas decreased in 9 (90.0%) and 7 (70.0%) patients at weeks 12 and 24, respectively. Significant improvement in Perianal Crohn Disease Activity Index, Pelvic MRI-Based Score, Crohn Disease Activity Index, and quality of life score were observed at 24 weeks. No serious adverse events occurred. The probability of remaining recurrence-free was 70% at week 52. CONCLUSION: The study demonstrated that local injection of TH-SC01 cells might be an effective and safe treatment for complex treatment-refractory pfCD after conventional and/or biological treatments fail (ClinicalTrials.gov ID, NCT04939337). TRIAL REGISTRATION: The study was retrospectively registered on www. CLINICALTRIALS: gov (NCT04939337) on June 25, 2021.

Deep Learning Model for Coronary Angiography.

The visual inspection of coronary artery stenosis is known to be significantly affected by variation, due to the presence of other tissues, camera movements, and uneven illumination. More accurate and intelligent coronary angiography diagnostic models are necessary for improving the above problems. In this study, 2980 medical images from 949 patients are collected and a novel deep learning-based coronary angiography (DLCAG) diagnose system is proposed. Firstly, we design a module of coronary classification. Then, we introduce RetinaNet to balance positive and negative samples and improve the recognition accuracy. Additionally, DLCAG adopts instance segmentation to segment the stenosis of vessels and depict the degree of the stenosis vessels. Our DLCAG is available at http://101.132.120.184:8077/ . When doctors use our system, all they need to do is login to the system, upload the coronary angiography videos. Then, a diagnose report is automatically generated.

Ropivacaine with Dexmedetomidine or Dexamethasone in a Thoracic Paravertebral Nerve Block Combined with an Erector Spinae Plane Block for Thoracoscopic Lobectomy Analgesia: A Randomized Controlled Trial.

Objective: This study aimed to investigate the effect of ropivacaine with dexmedetomidine or dexamethasone in a thoracic paravertebral nerve block (TPVB) combined with an erector spinae plane block (ESPB) for thoracoscopic lobectomy analgesia. Methods: A total of 97 patients undergoing thoracoscopic lobectomy under general anesthesia were enrolled in this study and randomly divided into three groups, ie, a ropivacaine group (Group R), a ropivacaine + dexmedetomidine group (Group R1), and a ropivacaine + dexamethasone group (Group R2). Ultrasound-guided TPVB combined with an erector spinae plane block was given after anesthesia induction. The following were applied to each group: Group R received 30 mL of 0.5% ropivacaine + 5 mL of a normal saline mixture; Group R1 received 30 mL of 0.5% ropivacaine + 5 mL of a 1 µg/kg dexmedetomidine mixture; Group R2 received 30 mL of 0.5% ropivacaine + 5 mL of an 8 mg dexamethasone mixture. The primary observation index was the time to the first postoperative remedial analgesia. The secondary observation indexes were the intraoperative consumption of propofol and sufentanil, time to waking from anesthesia, time to extubation, postoperative numerical rating scaltpe (NRS) score, postoperative sufentanil consumption, remedial analgesic dosage, and adverse reactions. Results: When compared with Group R, the time to first postoperative remedial analgesia was longer, the intraoperative and postoperative sufentanil consumption and flurbiprofen axetil remedial analgesic dose were lower, and the time to waking from anesthesia and time to extubation were shorter in groups R1 and R2 (P < 0.05). The NRS scores at 1, 6, 12, and 24 h postoperatively in groups R1 and R2 were lower than in Group R at the same time points (P < 0.05). Conclusion: Ropivacaine with dexmedetomidine or dexamethasone in TPVB combined with ESPB could prolong the time to first postoperative remedial analgesia, reduce perioperative sufentanil and postoperative remedial analgesic drug consumption, and decrease the postoperative NRS score in patients undergoing thoracoscopic lobectomy.

Discovering trends and hotspots of biosafety and biosecurity research via machine learning.

Coronavirus disease 2019 (COVID-19) has infected hundreds of millions of people and killed millions of them. As an RNA virus, COVID-19 is more susceptible to variation than other viruses. Many problems involved in this epidemic have made biosafety and biosecurity (hereafter collectively referred to as 'biosafety') a popular and timely topic globally. Biosafety research covers a broad and diverse range of topics, and it is important to quickly identify hotspots and trends in biosafety research through big data analysis. However, the data-driven literature on biosafety research discovery is quite scant. We developed a novel topic model based on latent Dirichlet allocation, affinity propagation clustering and the PageRank algorithm (LDAPR) to extract knowledge from biosafety research publications from 2011 to 2020. Then, we conducted hotspot and trend analysis with LDAPR and carried out further studies, including annual hot topic extraction, a 10-year keyword evolution trend analysis, topic map construction, hot region discovery and fine-grained correlation analysis of interdisciplinary research topic trends. These analyses revealed valuable information that can guide epidemic prevention work: (1) the research enthusiasm over a certain infectious disease not only is related to its epidemic characteristics but also is affected by the progress of research on other diseases, and (2) infectious diseases are not only strongly related to their corresponding microorganisms but also potentially related to other specific microorganisms. The detailed experimental results and our code are available at https://github.com/KEAML-JLU/Biosafety-analysis.

Aberrant HO-1/NQO1-Reactive Oxygen Species-ERK Signaling Pathway Contributes to Aggravation of TPA-Induced Irritant Contact Dermatitis in Nrf2-Deficient Mice.

NF-erythroid 2-related factor 2 (Nrf2) is a major transcription factor to protect cells against reactive oxygen species (ROS) and reactive toxicants. Meanwhile, Nrf2 can inhibit contact dermatitis through redox-dependent and -independent pathways. However, the underlying mechanisms of how Nrf2 mediates irritant contact dermatitis (ICD) are still unclear. In this article, we elucidated the role of Nrf2 in 12-O-tetradecanoylphorbol-13-acetate (TPA)-induced acute ICD. Our study demonstrated that the ear thickness, redness, swelling, and neutrophil infiltration were significantly increased, accompanied by increased expression of inflammatory cytokines (IL-1α, IL-1ß, IL-6, etc.) and decreased expression of antioxidant genes (HO-1 and NQO1) in Nrf2 knockout mice. Moreover, ERK phosphorylation was elevated in mouse embryonic fibroblasts (MEFs) from Nrf2 knockout mouse. Inhibition of ERK significantly alleviated TPA-induced cutaneous inflammation and ROS accumulation in MEFs derived from mouse. Conversely, ROS scavenging inhibited the ERK activation and TPA-induced inflammation in MEFs. Taken together, the findings illustrate the key role of the Nrf2/ROS/ERK signaling pathway in TPA-induced acute ICD.

Ultrasound-guided single thoracic paravertebral nerve block and erector spinae plane block for perioperative analgesia in thoracoscopic pulmonary lobectomy: a randomized controlled trial.

OBJECTIVE: To explore the effect of a single preoperative ultrasound-guided thoracic paravertebral nerve block (TPVB) and erector spinae plane block (ESPB) for perioperative analgesia in thoracoscopic pulmonary lobectomy. METHODS: Seventy-two patients aged 40-70 years who underwent thoracoscopic pulmonary lobectomy under general anesthesia were enrolled and randomly divided into the control group (Group C), the TPVB group (Group T) and the ESPB group (Group E). The primary observation indicators included the visual analogue scale (VAS) at 1, 6, 12, 24, and 48 h postoperatively at rest and with a cough. The secondary observation indicators included the intraoperative sufentanil consumption, anesthesia awakening time and extubation time, the sufentanil consumption in the analgesic pump, and flurbiprofen ester consumption for remedial analgesia within 48 h after surgery and the incidence of postoperative adverse events. RESULTS: The intraoperative sufentanil consumption, anesthesia awakening time, and extubation time were lower in groups T and E than those in group C (p < 0.05). Patients in group T had lower VAS scores at rest and with a cough at 1, 6, and 12 h postoperatively than in group C at the same time points (p < 0.05). The VAS scores at rest at 1 and 6 h postoperatively and coughing status at 1, 6, and 12 h postoperatively were lower in group E than in group C at the same time points (p < 0.05). CONCLUSION: The ultrasound-guided preoperative single TPVB and ESPB for thoracoscopic pulmonary lobectomy could both reduce the postoperative pain VAS score and reduce the dose of perioperative sufentanil and postoperative remedial analgesics.

Abnormally elevated ubiquilin­1 expression in breast cancer regulates metastasis and stemness via AKT signaling.

Ubiquilin­1 (UBQLN1) is an essential factor for the maintenance of proteostasis in cells. It is important for the regulation of different protein degradation mechanisms, including the ubiquitin­proteasome system, autophagy and endoplasmic reticulum­associated protein degradation pathways. However, the role of UBQLN1 in cancer progression remains largely unknown. In the present study, the expression, functions and molecular mechanisms of UBQLN1 in breast cancer tissue samples and cell lines were explored. Immunohistochemical and bioinformatics analyses revealed that UBQLN1 expression was significantly upregulated in breast cancer tissues and cell lines. UBQLN1 expression in breast cancer was significantly associated with lymph node metastasis and TNM stage. Moreover, a high UBQLN1 expression was a predictor of an unfavorable survival in patients with breast cancer. In vitro, UBQLN1 silencing markedly inhibited cell migration and invasion, epithelial­to­mesenchymal transition (EMT) and MMP expression. UBQLN1 silencing attenuated the stem cell­like properties of breast cancer cells, including their mammosphere­forming abilities. UBQLN1 knockdown also enhanced breast cancer cell chemosensitivity to paclitaxel. The expression levels of the stem cell markers. Aldehyde dehydrogenase 1 (ALDH1), Oct­4 and Sox2 were significantly decreased in the cells in which UBQLN1 was silenced, whereas breast cancer stem cells exhibited an increased expression of UBQLN1. Mechanistically, UBQLN1 knockdown inhibited the activation of AKT signaling, as revealed by the increased PTEN expression and the decreased expression of phosphorylated AKT in cells in which UBQLN1 was silenced. On the whole, the present study demonstrates that UBQLN1 is aberrantly upregulated in breast cancer and predicts a poor prognosis. The silencing of UBQLN1 inhibited the invasion, EMT and stemness of breast cancer cells, possibly via AKT signaling.

An oxidation responsive nano-radiosensitizer increases radiotherapy efficacy by remolding tumor vasculature.

As an excellent candidate material for nano-sensitizers, gold nanostructures have shown great potential in radiotherapy. Nevertheless, severe hypoxia and low accumulation of nanomedicine caused by poor perfusion at the tumor site have significantly reduced radiotherapy efficacy. Vascular normalization has gained attention owing to its ability to relieve hypoxia and increase perfusion. The synergistic therapy of tumor vascular normalization and radiotherapy has become a new option to increase anti-cancer efficacy. However, the commonly used strategy of suppressing a single growth factor to induce vascular normalization is limited by tumor compensatory effects. In this work, we developed a strategy to inhibit oxidative stress in tumors by generating chelating agents in response to hydrogen peroxide, thereby inhibiting multi-angiogenic factors simultaneously to normalize blood vessels. Concretely, sodium alginate (SA) reacted with 8-quinoline boric acid (QBA) to form SA-QBA. Then gold nanoparticles (Au NPs) were modified with SA-QBA to obtain Au@SA-QBA. The system was simple in structure and could generate 8HQ in response to H2O2in vitro to inhibit oxidative stress and reduce the expression of VEGF, bFGF, and Ang-2. In vivo, the perfusion unit (PU) increased by 78% after Au@SA-QBA treatment, and the coverage of pericytes increased by 32%, which in turn induced vascular normalization. In addition, blood routine and blood biochemical tests confirmed its good biocompatibility and 8HQ was not detected in the supernatant after homogenization of major organs. More importantly, after the synergistic treatment of vascular normalization and radiotherapy (4 Gy), the tumor growth inhibition rate was increased by 38.6% compared to the Au@SA-treated group with negligible side effects to normal tissues.

CircRNA CircRIMS is Overexpressed in Esophageal Squamous Cell Carcinoma and Downregulate miR-613 Through Methylation to Increase Cell Proliferation.

PURPOSE: CircRNA CircRIMS has been characterized as an oncogenic circRNA in gastric cancer, while its role in other cancers is unknown. This study aimed to explore the role of CircRIMS in esophageal squamous cell carcinoma (ESCC). PATIENTS AND METHODS: Tissues collected from 60 ESCC patients were subjected to extractions of total RNA and RT-qPCRs to analyze the differential expression of CircRIMS and miR-613. The 60 ESCC patients were followed up for 5 years to analyze the prognostic value of CircRIMS for ESCC. The interaction between CircRIMS and miR-613 was showed by luciferase activity assay and fluorescence in situ hybridization. The role of CircRIMS in regulating miR-613 expression and methylation was analyzed by overexpression experiments, RT-qPCRs and Western blot assay. The role of CircRIMS and miR-613 in regulating cell proliferation was analyzed using the BrdU assay. ESCC xenograft model was used to demonstrate the role of CircRIMS and miR-613 in vivo. RESULTS: We found that CircRIMS was overexpressed in ESCC and predicted poor survival. In addition, miR-613 was under expressed in ESCC and inversely correlated with CircRIMS. In ESCC cells, CircRIMS overexpression decreased the expression of miR-613 and increased the methylation of miR-613 gene. Cell proliferation assay showed that CircRIMS overexpression reduced the inhibitory effects of miR-613 overexpression on cell proliferation. Animal experience finally illustrated that CircRNA CircRIMS downregulated miR-613 through methylation to promote tumor growth. CONCLUSION: Therefore, CircRIMS may downregulate miR-613 through methylation to increase cell proliferation in ESCC.

Construction of a pH/TGase "Dual Key"-Responsive Gold Nano-radiosensitizer with Liver Tumor-Targeting Ability.

The method of tumor microenvironment (TME)-responsive aggregation has become a promising approach to enhance treatment effect by improving the accumulation of nanoparticles in tumors. The enzymatic cross-linking strategy has widely attracted attention owing to its good aggregation stability and biocompatibility. However, the enzymes in nontumor tissue can also catalyze the cross-linking reaction and reduce accumulation of nanoparticles in tumor. In this work, a "dual key"-responsive strategy is utilized to construct a transglutaminase (TGase)/pH-responsive radiosensitizer (Au@TAcoGal) with specific aggregation behavior in hepatic tumor cells. Au@TAcoGal can retain its stability in blood circulation (pH 7.4) even in the presence of TGase in plasma. On reaching tumor sites, it can be endocytosed by hepatoma cells by the active targeting of phenylboronic acid (PBA) and aggregated under acidity and overexpression of TGase in cells. Due to its specific accumulation in hepatoma cells, radiotherapy can be operated under a lower dose of X-ray. The results show that the cellular accumulation of Au@TAcoGal increases by 30-70%, and the cell survival rate is less than 25% under X-ray irradiation. The antineoplastic results show that Au@TAcoGal exhibits a higher therapeutic effect, and the tumor inhibition rate can reach 84.21%.

Cancer Research Trend Analysis Based on Fusion Feature Representation.

Machine learning models can automatically discover biomedical research trends and promote the dissemination of information and knowledge. Text feature representation is a critical and challenging task in natural language processing. Most methods of text feature representation are based on word representation. A good representation can capture semantic and structural information. In this paper, two fusion algorithms are proposed, namely, the Tr-W2v and Ti-W2v algorithms. They are based on the classical text feature representation model and consider the importance of words. The results show that the effectiveness of the two fusion text representation models is better than the classical text representation model, and the results based on the Tr-W2v algorithm are the best. Furthermore, based on the Tr-W2v algorithm, trend analyses of cancer research are conducted, including correlation analysis, keyword trend analysis, and improved keyword trend analysis. The discovery of the research trends and the evolution of hotspots for cancers can help doctors and biological researchers collect information and provide guidance for further research.

Dimension Reduction and Clustering Models for Single-Cell RNA Sequencing Data: A Comparative Study.

With recent advances in single-cell RNA sequencing, enormous transcriptome datasets have been generated. These datasets have furthered our understanding of cellular heterogeneity and its underlying mechanisms in homogeneous populations. Single-cell RNA sequencing (scRNA-seq) data clustering can group cells belonging to the same cell type based on patterns embedded in gene expression. However, scRNA-seq data are high-dimensional, noisy, and sparse, owing to the limitation of existing scRNA-seq technologies. Traditional clustering methods are not effective and efficient for high-dimensional and sparse matrix computations. Therefore, several dimension reduction methods have been introduced. To validate a reliable and standard research routine, we conducted a comprehensive review and evaluation of four classical dimension reduction methods and five clustering models. Four experiments were progressively performed on two large scRNA-seq datasets using 20 models. Results showed that the feature selection method contributed positively to high-dimensional and sparse scRNA-seq data. Moreover, feature-extraction methods were able to promote clustering performance, although this was not eternally immutable. Independent component analysis (ICA) performed well in those small compressed feature spaces, whereas principal component analysis was steadier than all the other feature-extraction methods. In addition, ICA was not ideal for fuzzy C-means clustering in scRNA-seq data analysis. K-means clustering was combined with feature-extraction methods to achieve good results.

Broadband indium tin oxide nanowire arrays as saturable absorbers for solid-state lasers.

Indium Tin Oxide nanowire arrays (ITO-NWAs), as epsilon-near-zero (ENZ) materials, exhibit a fast response time and a low saturable absorption intensity, which make them promising photoelectric materials. In this study, ITO-NWAs were successfully fabricated using a chemical vapor deposition (CVD) method, and the saturable absorption properties of this material were characterized in the near-infrared region. Further, passively Q-switched all-solid-state lasers were realized at wavelengths of 1.0, 1.3, and 2.0 µm using the as-prepared saturable absorber (SA). To the best of our knowledge, we present the first application of ITO-NWAs in all-solid-state lasers. The results reveal that ITO-NWAs may be applied as an SA while developing Q-switched lasers and that they exhibit a broad application prospect as broadband saturable absorption materials.

Feature selection may improve deep neural networks for the bioinformatics problems.

MOTIVATION: Deep neural network (DNN) algorithms were utilized in predicting various biomedical phenotypes recently, and demonstrated very good prediction performances without selecting features. This study proposed a hypothesis that the DNN models may be further improved by feature selection algorithms. RESULTS: A comprehensive comparative study was carried out by evaluating 11 feature selection algorithms on three conventional DNN algorithms, i.e. convolution neural network (CNN), deep belief network (DBN) and recurrent neural network (RNN), and three recent DNNs, i.e. MobilenetV2, ShufflenetV2 and Squeezenet. Five binary classification methylomic datasets were chosen to calculate the prediction performances of CNN/DBN/RNN models using feature selected by the 11 feature selection algorithms. Seventeen binary classification transcriptome and two multi-class transcriptome datasets were also utilized to evaluate how the hypothesis may generalize to different data types. The experimental data supported our hypothesis that feature selection algorithms may improve DNN models, and the DBN models using features selected by SVM-RFE usually achieved the best prediction accuracies on the five methylomic datasets. AVAILABILITY AND IMPLEMENTATION: All the algorithms were implemented and tested under the programming environment Python version 3.6.6. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.