DeepIMAGER: Deeply Analyzing Gene Regulatory Networks from scRNA-seq Data.

Understanding the dynamics of gene regulatory networks (GRNs) across diverse cell types poses a challenge yet holds immense value in unraveling the molecular mechanisms governing cellular processes. Current computational methods, which rely solely on expression changes from bulk RNA-seq and/or scRNA-seq data, often result in high rates of false positives and low precision. Here, we introduce an advanced computational tool, DeepIMAGER, for inferring cell-specific GRNs through deep learning and data integration. DeepIMAGER employs a supervised approach that transforms the co-expression patterns of gene pairs into image-like representations and leverages transcription factor (TF) binding information for model training. It is trained using comprehensive datasets that encompass scRNA-seq profiles and ChIP-seq data, capturing TF-gene pair information across various cell types. Comprehensive validations on six cell lines show DeepIMAGER exhibits superior performance in ten popular GRN inference tools and has remarkable robustness against dropout-zero events. DeepIMAGER was applied to scRNA-seq datasets of multiple myeloma (MM) and detected potential GRNs for TFs of RORC, MITF, and FOXD2 in MM dendritic cells. This technical innovation, combined with its capability to accurately decode GRNs from scRNA-seq, establishes DeepIMAGER as a valuable tool for unraveling complex regulatory networks in various cell types.

Targeting NG2 relieves the resistance of BRAF-mutant thyroid cancer cells to BRAF inhibitors.

BRAFV600E represents a constitutively active onco-kinase and stands as the most prevalent genetic alteration in thyroid cancer. However, the clinical efficacy of small-molecule inhibitors targeting BRAFV600E is often limited by acquired resistance. Here, we find that nerve/glial antigen 2 (NG2), also known as chondroitin sulfate proteoglycan 4 (CSPG4), is up-regulated in thyroid cancers, and its expression is increased with tumor progression in a BRAFV600E-driven thyroid cancer mouse model. Functional studies show that NG2 knockout almost does not affect tumor growth, but significantly improves the response of BRAF-mutant thyroid cancer cells to BRAF inhibitor PLX4720. Mechanistically, the blockade of ERK-dependent feedback by BRAF inhibitor can activate receptor tyrosine kinase (RTK) signaling, causing the resistance to this inhibitor. NG2 knockout attenuates the PLX4720-mediated feedback activation of several RTKs, improving the sensitivity of BRAF-mutant thyroid cancer cells to this inhibitor. Based on this finding, we propose and demonstrate an alternative strategy for targeting NG2 to effectively treat BRAF-mutant thyroid cancers by combining multiple kinase inhibitor (MKI) Sorafenib or Lenvatinib with PLX4720. Thus, this study uncovers a new mechanism in which NG2 contributes to the resistance of BRAF-mutant thyroid cancer cells to BRAF inhibitor, and provides a promising therapeutic option for BRAF-mutant thyroid cancers.

Deep Batch Integration and Denoise of Single-Cell RNA-Seq Data.

Numerous single-cell transcriptomic datasets from identical tissues or cell lines are generated from different laboratories or single-cell RNA sequencing (scRNA-seq) protocols. The denoising of these datasets to eliminate batch effects is crucial for data integration, ensuring accurate interpretation and comprehensive analysis of biological questions. Although many scRNA-seq data integration methods exist, most are inefficient and/or not conducive to downstream analysis. Here, DeepBID, a novel deep learning-based method for batch effect correction, non-linear dimensionality reduction, embedding, and cell clustering concurrently, is introduced. DeepBID utilizes a negative binomial-based autoencoder with dual Kullback-Leibler divergence loss functions, aligning cell points from different batches within a consistent low-dimensional latent space and progressively mitigating batch effects through iterative clustering. Extensive validation on multiple-batch scRNA-seq datasets demonstrates that DeepBID surpasses existing tools in removing batch effects and achieving superior clustering accuracy. When integrating multiple scRNA-seq datasets from patients with Alzheimer's disease, DeepBID significantly improves cell clustering, effectively annotating unidentified cells, and detecting cell-specific differentially expressed genes.

RNF115 aggravates tumor progression through regulation of CDK10 degradation in thyroid carcinoma.

BACKGROUND: RING Finger Protein 115 (RNF115), a notable E3 ligase, is known to modulate tumorigenesis and metastasis. In our investigation, we endeavor to unravel the putative function and inherent mechanism through which RNF115 influences the evolution of thyroid carcinoma (THCA). METHODS: We analyzed RNF115 expression in THCA using the Cancer Genome Atlas (TCGA) database. The influence of RNF115 on the progression of THCA was evaluated using both in vitro and in vivo experimental approaches. The protein regulated by RNF115 was identified through bioinformatics analysis, and its biological significance was further explored. RESULTS: In both THCA tissues and cells, RNF115 showed elevated expression levels. Enhanced expression of RNF115 fostered cell proliferation, tumor growth, and the exacerbation of epithelial-mesenchymal transition (EMT) in THCA, while also promoting tumor lung metastasis. Bioinformatics analysis identified cyclin-dependent kinase 10 (CDK10) as a downstream target of RNF115, which was found to be ubiquitinated and degraded by RNF115 in THCA cells. Functionally, overexpression of CDK10 was found to counteract the promotion of malignant phenotype in THCA induced by RNF115. From a mechanistic perspective, RNF115 activated the Raf-1 pathway and enhanced cancer cell cycle progression by degrading CDK10 in THCA cells. CONCLUSION: RNF115 triggers cell proliferation, EMT, and tumor metastasis by ubiquitinating and degrading CDK10. The regulation of the Raf-1 pathway and cell cycle progression in THCA may be profoundly influenced by this process.

MYO5A overexpression promotes invasion and correlates with low lymphocyte infiltration in head and neck squamous carcinoma.

Head and neck squamous carcinoma (HNSC) poses a significant public health challenge due to its substantial morbidity. Nevertheless, despite advances in current treatments, the prognosis for HNSC remains unsatisfactory. To address this, single-cell RNA sequencing (RNA-seq) and bulk RNA-seq data combined with in vitro studies were conducted to examine the role of MYO5A (Myosin VA) in HNSC. Our investigation revealed an overexpression of MYO5A in HNSC that promotes HNSC migration in vitro. Remarkably, knockdown of MYO5A suppressed vimentin expression. Furthermore, analyzing the TCGA database evidenced that MYO5A is a risk factor for human papillomavirus positive (HPV+) HNSC (HR = 0.81, P < 0.001). In high MYO5A expression HNSC, there was a low count of tumor infiltrating lymphocytes (TIL), including activated CD4+ T cells, CD8+ T cells, and B cells. Of note, CD4+ T cells and B cells were positively associated with improved HPV+ HNSC outcomes. Correlation analysis demonstrated a decreased level of immunostimulators in high MYO5A-expressing HNSC. Collectively, these findings suggest that MYO5A may promote HNSC migration through vimentin and involve itself in the process of immune infiltration in HNSC, advancing the understanding of the mechanisms and treatment of HNSC.

scTIGER: A Deep-Learning Method for Inferring Gene Regulatory Networks from Case versus Control scRNA-seq Datasets.

Inferring gene regulatory networks (GRNs) from single-cell RNA-seq (scRNA-seq) data is an important computational question to find regulatory mechanisms involved in fundamental cellular processes. Although many computational methods have been designed to predict GRNs from scRNA-seq data, they usually have high false positive rates and none infer GRNs by directly using the paired datasets of case-versus-control experiments. Here we present a novel deep-learning-based method, named scTIGER, for GRN detection by using the co-differential relationships of gene expression profiles in paired scRNA-seq datasets. scTIGER employs cell-type-based pseudotiming, an attention-based convolutional neural network method and permutation-based significance testing for inferring GRNs among gene modules. As state-of-the-art applications, we first applied scTIGER to scRNA-seq datasets of prostate cancer cells, and successfully identified the dynamic regulatory networks of AR, ERG, PTEN and ATF3 for same-cell type between prostatic cancerous and normal conditions, and two-cell types within the prostatic cancerous environment. We then applied scTIGER to scRNA-seq data from neurons with and without fear memory and detected specific regulatory networks for BDNF, CREB1 and MAPK4. Additionally, scTIGER demonstrates robustness against high levels of dropout noise in scRNA-seq data.

Self-supervised deep clustering of single-cell RNA-seq data to hierarchically detect rare cell populations.

Single-cell RNA sequencing (scRNA-seq) is a widely used technique for characterizing individual cells and studying gene expression at the single-cell level. Clustering plays a vital role in grouping similar cells together for various downstream analyses. However, the high sparsity and dimensionality of large scRNA-seq data pose challenges to clustering performance. Although several deep learning-based clustering algorithms have been proposed, most existing clustering methods have limitations in capturing the precise distribution types of the data or fully utilizing the relationships between cells, leaving a considerable scope for improving the clustering performance, particularly in detecting rare cell populations from large scRNA-seq data. We introduce DeepScena, a novel single-cell hierarchical clustering tool that fully incorporates nonlinear dimension reduction, negative binomial-based convolutional autoencoder for data fitting, and a self-supervision model for cell similarity enhancement. In comprehensive evaluation using multiple large-scale scRNA-seq datasets, DeepScena consistently outperformed seven popular clustering tools in terms of accuracy. Notably, DeepScena exhibits high proficiency in identifying rare cell populations within large datasets that contain large numbers of clusters. When applied to scRNA-seq data of multiple myeloma cells, DeepScena successfully identified not only previously labeled large cell types but also subpopulations in CD14 monocytes, T cells and natural killer cells, respectively.

A Feedback Loop Involving Exosomal miR-146a and NG2 to Propel the Development and Progression of Hypothyroidism.

Background: Thyrotropin receptor (TSHR) plays a central role in maintaining thyroid function and TSHR impairment causes hypothyroidism, which is often associated with metabolic disarrangement. The most common type of hypothyroidism is autoimmune disease-related and the mechanism, particularly with respect to the role of microRNAs (miRNAs), has not been delineated. Methods: Serum from 30 patients with subclinical hypothyroidism (SCH) and 30 healthy individuals were collected and exosomal miR-146a (exo-miR-146a) was examined, followed by extensive mechanistic investigation using various molecular and cellular experimental approaches and genetic-knockout mouse models. Results: Our clinical investigation showed that exo-miR-146a was systemically elevated in the serum of patients with SCH (p = 0.04) compared with healthy individuals, prompting us to investigate the biological effects of miR-146a in cells. We found that miR-146a could target and down-regulate neuron-glial antigen 2 (Ng2), with consequent down-regulation of TSHR. We next generated a thyroid-specific Ng2 knockout (Thy-Ng2-/-) mouse model and found a significant down-regulation of TSHR in Thy-Ng2-/- mice, accompanied by the development of hypothyroidism and metabolic disorders. We further found that a decrease in NG2 resulted in decreased receptor tyrosine kinase-linked downstream signaling and down-regulation of c-Myc, consequently resulting in up-regulation of miR-142 and miR-146a in thyroid cells. Up-regulated miR-142 targeted the 3'-untranslated region (UTR) of TSHR messenger RNA (mRNA) and post-transcriptionally down-regulated TSHR, explaining the development of hypothyroidism above. Local up-regulation of miR-146a in thyroid cells augments the earlier cited processes initiated by systemically elevated miR-146a, thereby forming a feedback loop to propel the development and progression of hypothyroidism. Conclusions: This study has uncovered a self-augmenting molecular loop initiated by elevated exo-miR-146a to suppress TSHR through targeting and down-regulating NG2, thereby initiating and propelling the development and progression of hypothyroidism.

Efficacy and safety of Artemisia annua sublingual immunotherapy in patients with seasonal allergic rhinoconjunctivitis over two pollen seasons.

OBJECTIVE: This study investigates the efficacy and safety of sublingual immunotherapy (SLIT) with A. annua allergens in patients with seasonal allergic rhinoconjunctivitis over two pollen seasons. METHODS: Seventy patients with moderate-severe seasonal allergic rhinoconjunctivitis were divided evenly into the SLIT and control groups. The SLIT last from 3 months before the summer-autumn pollen season in 2021 till the end of the summer-autumn pollen season in 2022. The daily individual symptom score, total rhinoconjunctivitis symptom score (dTRSS), total medication score (dTMS), combined score of medication and rhinoconjunctivitis symptom (dCSMRS), visual analog scale (VAS) score, and adverse events (AEs) were evaluated. RESULTS: The average pollen concentration in 2022 was twice that previous two-year during the pollen season. Fifty-six patients completed treatments (SLIT group: 29, control group: 27). Compared with baseline, the individual symptoms, dTRSS, dTMS, dCSMRS, and VAS scores of SLIT group declined in 2021. After 16 months of SLIT, all efficacy indexes in 2022 were still lower than baseline and equivalent to those in 2021. In control group, the efficacy indexes in 2022 were higher than that in 2020 and 2021. The efficacy indexes of SLIT group were lower than those of control group in 2021 and 2022. SLIT is effective for both mono- and poly-sensitized patients. AEs incidence in SLIT group was 82.7% without severe AEs. CONCLUSIONS: The A. annua-SLIT can obtain efficacy and safety over two pollen seasons for patients with moderate-severe seasonal allergic rhinoconjunctivitis.

The Feasibility Study of Intraoperative RLN Monitoring Using Cricothyroid Membrane-Inserted Needle Electrodes During Thyroid Surgery.

OBJECTIVE: This study evaluated the feasibility, stability, safety, and economy of cricothyroid membrane (CM)-inserted needle electrodes for recurrent laryngeal nerve monitoring. STUDY DESIGN: Parallel and controlled study. SETTING: Clinical research center for thyroid diseases of Shaanxi province. METHODS: A total of 64 patients in the needle electrodes group (104 recurrent laryngeal nerves [RLNs]) and 44 patients in the endotracheal tube (ETT)-based electrodes group (80 RLNs) underwent monitored thyroidectomy. The evoked electromyography (EMG) signals detected by the 2 electrodes were recorded and analyzed. The changes in EMG during Berry's ligament traction and tracheal displacement were compared. All patients underwent preoperative and postoperative laryngoscopy within 1 week. RESULTS: Both electrodes successfully recorded typical evoked laryngeal EMG waveforms from RLNs. The needle electrodes recorded relatively higher amplitudes and similar latencies compared to ETT-based electrodes. The evoked EMG signals attributed to needle electrodes could accurately predict the function of RLNs with 100% sensitivity and specificity. The reduction in the recorded amplitudes attributed to needle electrodes was higher than that observed with ETT-based electrodes during Berry's ligament traction or trachea displacement, whereas a similar increase in the latencies was recorded in the 2 groups. Particularly, Berry's ligament traction was more likely to lead to EMG amplitude reduction and latency prolongation. The needle electrodes group recorded 2 cases of minor bleeding on the CM. The needle electrodes were more cost-effective than ETT-based electrodes. CONCLUSION: The CM-inserted needle electrodes are feasible, stable, safe, and economical for RLN monitoring, and they provide an alternative novel intraoperative neural monitoring format for thyroid surgeons.

[The influence of additional roll test on the repositioning procedure by SRM-vertigo diagnosis system for horizontal canal benign paroxysmal positional vertigo].

Objective:To evaluate the influence of an additional roll test on the repositioning procedure by SRM-vertigo diagnosis system for horizontal canal benign paroxysmal positional vertigo（HC-BPPV）. Methods:A total of 713 patients diagnosed with HC-BPPV in Department of Otolaryngology Head and Neck Surgery，the First Affiliated Hospital of Xi'an Jiaotong University from Jan 2020 to Feb 2022 were enrolled. The patients were divided into two groups by hospital card numbers, in which the number is odd were considered as group A, and the number is even were considered as group B. The group A underwent two circles of Barbecue repositioning procedure by SRM-vertigo diagnosis system, while the group B first performed an additional roll test and then underwent two circles of Barbecue repositioning procedure by SRM-vertigo diagnosis system, to observe the cure rate and compare influence of HC-BPPV by an additional roll test. The quality of life and sleep of patients before and one-month after the treatment were assessed by the dizziness handicap inventory（DHI） and the pittsburgh sleep quality（PSQI）. Results:The cure rate of group A was 63.21%, and the cure rate of group B was 87.68%，the difference between the two groups was statistically significant（P<0.05）; The DHI score of patients after the repositioning was significantly lower than that before the repositioning（P<0.05）. The PSQI score after the repositioning was significantly lower than that before the repositioning（P<0.05）. The DHI and the PSQI scores after the repositioning were significantly lower than that before the repositioning, with a statistically significant difference （P< 0.05）. The total score of DHI in group B after treatment was lower than that in group A, with a statistically significant difference（P<0.05）. The total score of PSQI in group B after treatment was lower than that in group A, with non-statistically significant difference （P< 0.05）. Conclusion:An additional roll test before the repositioning procedure by SRM-vertigo diagnosis system can significantly improve the cure rate of HC-BPPV, relieve anxiety, and improve the quality of life.

[Progress in active surveillance for adult low-risk papillary thyroid microcarcinoma].

The incidence of papillary thyroid microcarcinoma （PTMC） increases rapidly. However, epidemiological and autopsy studies show that the prevalence of low-risk papillary thyroid microcarcinoma （LR-PTMC） is very high, but the mortality is very low. There is over-diagnosis and over-treatment for LR-PTMC. Active surveillance （AS） was adopted for LR-PTMCs instead of immediate surgery, and more than 70% of the lesions remained stable or shrank in clinical observation. Therefore, AS is recommended for LR-PTMCs in clinical guidelines of several academic organizations around the world. However, PTMC is not equal to low-risk cancer. The implementation of AS strategy requires a strict grasp of indications and full consideration of population characteristics to ensure the maximum benefit of patients. This paper summarizes the present clinical progress of active surveillance for adult LR-PTMC.

Frizzled-3 suppression overcomes multidrug chemoresistance by Wnt/ß-catenin signaling pathway inhibition in hepatocellular carcinoma cells.

Multidrug resistance (MDR) is a major obstacle to the efficacy of hepatocellular carcinoma (HCC) chemotherapy. Previous studies have identified that low FZD3 predicted decreased survival after intraperitoneal versus intravenous-only chemotherapy in ovarian cancer. This study aimed to identify a potential target in HCC chemotherapy. The FZD3 expression variant in HCC cell lines was detected by RT-qPCR and western blotting. The FZD3 expression in the early recurrent HCC group (RE group) and the non-early recurrent HCC group (non-RE group) was measured by RT-qPCR. Then, the 50% inhibitory concentrations (IC50) in HCC cell lines were studied by MTT assay. TOP/FOP FLASH luciferase assay was performed to measure TCF-binding activities. We found that FZD3 was upregulated in three HCC cell lines, and the FZD3 expression was significantly higher in the RE group than in the non-RE group (P = 0.0344). A positive correlation between FZD3 and MDR1 was observed in HCC tissues (R2 = 0.6368, P = 0.0001). Then, we found that FZD3 knockdown significantly altered Huh-7 cell chemotherapeutic sensitivity to cisplatin [50.43 µM in the FZD3 siRNA (siFZD3) group vs 98.59 µM in the siRNA negative control (siNC) group; P = 0.007] or doxorubicin (7.43 µM in the siFZD3 group vs 14.93 µM in the siNC group; P = 0.017). TOP/FOP FLASH luciferase assay showed FZD3 could inhibit Wnt/ß-catenin signaling in HCC cells. Moreover, FZD3 expression knockdown in SNU-449 and Huh-7 cells markedly reduced ß-catenin and phosho-ß-catenin (S37) protein expression, and Cyclin D1, c-myc and MDR1 were significantly decreased. This is the first study to describe the significantly increased FZD3 expression in patients with early recurrent HCC. FZD3 knockdown led to increased sensitivity to chemotherapy by Wnt/ß-catenin signaling inhibition in HCC cell lines. Our study suggests FZD3 as a potential target for reversing chemoresistance in HCC.

SKP2 Contributes to AKT Activation by Ubiquitination Degradation of PHLPP1, Impedes Autophagy, and Facilitates the Survival of Thyroid Carcinoma.

Papillary thyroid carcinoma (PTC) is the most common subtype of thyroid carcinoma. Despite a good prognosis, approximately a quarter of PTC patients are likely to relapse. Previous reports suggest an association between S-phase kinase-associated protein 2 (SKP2) and the prognosis of thyroid cancer. SKP1 is related to apoptosis of PTC cells; however, its role in PTC remains largely elusive. This study aimed to understand the expression and molecular mechanism of SKP2 in PTC. SKP2 expression was upregulated in PTC tissues and closely associated with clinical diagnosis. In vitro and in vivo knockdown of SKP2 expression in PTC cells suppressed cell growth and proliferation and induced apoptosis. SKP2 depletion promoted cell autophagy under glucose deprivation. SKP2 interacted with PH domain leucine-rich repeat protein phosphatase-1 (PHLPP1), triggering its degradation by ubiquitination. Furthermore, SKP2 activates the AKT-related pathways via PHLPP1, which leads to the cytoplasmic translocation of SKP2, indicating a reciprocal regulation between SKP2 and AKT. In conclusion, the upregulation of SKP2 leads to PTC proliferation and survival, and the regulatory network among SKP2, PHLPP1, and AKT provides novel insight into the molecular basis of SKP2 in tumor progression.

IGFL2-AS1 facilitates tongue squamous cell carcinoma progression via Wnt/ß-catenin signaling pathway.

OBJECTIVES: Tongue squamous cell carcinoma (TSCC) is the most common malignancy in oral cancer. Long noncoding RNAs (lncRNAs) are important regulators in cancer biology. In our present study, we investigated a novel lncRNA IGF-like family member 2 antisense RNA 1 (IGFL2-AS1) in TSCC. METHODS: RT-qPCR analyzed IGFL2-AS1 expression in TSCC cells. Functional assays assessed the impact of IGFL2-AS1 on TSCC cell proliferation, migration, and invasion. Western blot analyzed the protein levels of EMT-related markers. Mechanism assays analyzed the regulatory mechanism of IGFL2-AS1 in TSCC cells. In-vivo experiments were conducted to prove the role of IGFL2-AS1 in TSCC progression. RESULTS: IGFL2-AS1 was significantly up-regulated in TSCC cells and tissues, and IGFL2-AS1 knockdown inhibited cell proliferation, migration, invasion and EMT in TSCC. Moreover, IGFL2-AS1 functioned as a competing endogenous RNA (ceRNA) to sponge miR-1224-5p and thereby modulated SATB homeobox 1 (SATB1) expression. Additionally, SATB1 activated the Wnt/ß-catenin signaling pathway in TSCC cells and IGFL2-AS1 regulated the Wnt/ß-catenin signaling pathway and TSCC progression via elevating SATB1 expression. CONCLUSIONS: The data revealed that IGFL2-AS1 played a cancer promoting role in TSCC and may aid in exploring a brand new biomarker that might contribute to TSCC treatment.

Investigation of Volumetric Block Proportion (VBP) Effect on Excavation-Induced Ground Response of Talus-like Rock Mass Based on DEM Simulations.

Due to the complexity of the talus-like rock mass with different values of volumetric block proportion (VPB), it is thus crucial to explore the VBP effect on the excavation-induced ground responses. We conduct a series of 2D DEM (discrete element method) simulations on a common circular tunnel excavation in the talus-like rock mass with different VBPs (0%, 15%, 50%, 85% and 100%). For each VBP, two support scenarios, i.e., unsupported and supported by a rigid lining, are considered. The micro characteristics of the excavation-induced ground responses, including the contact force, force chain, coordination number and shear-slip contact, and the stress distribution and ground settlement are elaborated in detail. Accordingly, three types of talus-like rock masses are identified as soil-, hybrid- and rock-types, corresponding to VBP = 0-15%, 50%, and 85-100%, respectively. It is found that the lining support is essential for maintaining the ground stability of a tunnel excavation in the soil- and hybrid-type talus-like rock masses while the backbones formed by rock blocks in the rock-type talus-like rock mass can provide a certain support for the surrounding ground. Our findings have important implications for optimizing the construction scheme of tunnel excavation in different types of talus-like rock masses.

Combining Phenotypes of Nucleotide Excision Repair Pathway to Predict the Risk of Head and Neck Squamous Cell Carcinomas in a Chinese Population.

Background: Nucleotide excision repair (NER) is pivotal in the development of smoking-related malignancies. Nine core genes (XPA, XPB, XPC, XPD, XPF, XPG, ERCC1, DDB1, and DDB2) are highly involved in the NER process. We combined two phenotypes of NER pathway (NER protein and NER gene mRNA expression) and evaluated their associations with the risks of the head and neck squamous cell carcinomas (HNSCCs) in a Chinese population. Methods: We conducted a case-control study of 337 HNSCC patients and 285 cancer-free controls by measuring the expression levels of nine core NER proteins and NER gene mRNA in cultured peripheral lymphocytes. Results: Compared with the controls, cases had statistically significantly lower protein expression levels of XPA (P < 0.001) and lower mRNA expression levels of XPA and XPB (P = 0.005 and 0.001, respectively). After dividing the subjects by controls' medians of expression levels, we found an association between increased risks of HNSCCs and low XPA protein level (P trend = 0.031), as well as low mRNA levels of XPA and XPB (P trend = 0.024 and 0.001, respectively). Subsequently, we correlated the two phenotypes and found associations between the NER mRNA and protein levels. Finally, the sensitivity of the expanded model with protein and mRNA expression levels, in addition to demographic variables, on HNSCCs risk was significantly improved. Conclusions: Combining two phenotypes of NER pathway may be more effective than the model only including one single phenotype for the assessment of risks of HNSCCs.

Supervised Contrastive Learning and Intra-Dataset Adversarial Adaptation for Iris Segmentation.

Precise iris segmentation is a very important part of accurate iris recognition. Traditional iris segmentation methods require complex prior knowledge and pre- and post-processing and have limited accuracy under non-ideal conditions. Deep learning approaches outperform traditional methods. However, the limitation of a small number of labeled datasets degrades their performance drastically because of the difficulty in collecting and labeling irises. Furthermore, previous approaches ignore the large distribution gap within the non-ideal iris dataset due to illumination, motion blur, squinting eyes, etc. To address these issues, we propose a three-stage training strategy. Firstly, supervised contrastive pretraining is proposed to increase intra-class compactness and inter-class separability to obtain a good pixel classifier under a limited amount of data. Secondly, the entire network is fine-tuned using cross-entropy loss. Thirdly, an intra-dataset adversarial adaptation is proposed, which reduces the intra-dataset gap in the non-ideal situation by aligning the distribution of the hard and easy samples at the pixel class level. Our experiments show that our method improved the segmentation performance and achieved the following encouraging results: 0.44%, 1.03%, 0.66%, 0.41%, and 0.37% in the Nice1 and 96.66%, 98.72%, 93.21%, 94.28%, and 97.41% in the F1 for UBIRIS.V2, IITD, MICHE-I, CASIA-D, and CASIA-T.

Decoding brain memory formation by single-cell RNA sequencing.

To understand how distinct memories are formed and stored in the brain is an important and fundamental question in neuroscience and computational biology. A population of neurons, termed engram cells, represents the physiological manifestation of a specific memory trace and is characterized by dynamic changes in gene expression, which in turn alters the synaptic connectivity and excitability of these cells. Recent applications of single-cell RNA sequencing (scRNA-seq) and single-nucleus RNA sequencing (snRNA-seq) are promising approaches for delineating the dynamic expression profiles in these subsets of neurons, and thus understanding memory-specific genes, their combinatorial patterns and regulatory networks. The aim of this article is to review and discuss the experimental and computational procedures of sc/snRNA-seq, new studies of molecular mechanisms of memory aided by sc/snRNA-seq in human brain diseases and related mouse models, and computational challenges in understanding the regulatory mechanisms underlying long-term memory formation.

Detecting Fear-Memory-Related Genes from Neuronal scRNA-seq Data by Diverse Distributions and Bhattacharyya Distance.

The detection of differentially expressed genes (DEGs) is one of most important computational challenges in the analysis of single-cell RNA sequencing (scRNA-seq) data. However, due to the high heterogeneity and dropout noise inherent in scRNAseq data, challenges in detecting DEGs exist when using a single distribution of gene expression levels, leaving much room to improve the precision and robustness of current DEG detection methods. Here, we propose the use of a new method, DEGman, which utilizes several possible diverse distributions in combination with Bhattacharyya distance. DEGman can automatically select the best-fitting distributions of gene expression levels, and then detect DEGs by permutation testing of Bhattacharyya distances of the selected distributions from two cell groups. Compared with several popular DEG analysis tools on both large-scale simulation data and real scRNA-seq data, DEGman shows an overall improvement in the balance of sensitivity and precision. We applied DEGman to scRNA-seq data of TRAP; Ai14 mouse neurons to detect fear-memory-related genes that are significantly differentially expressed in neurons with and without fear memory. DEGman detected well-known fear-memory-related genes and many novel candidates. Interestingly, we found 25 DEGs in common in five neuron clusters that are functionally enriched for synaptic vesicles, indicating that the coupled dynamics of synaptic vesicles across in neurons plays a critical role in remote memory formation. The proposed method leverages the advantage of the use of diverse distributions in DEG analysis, exhibiting better performance in analyzing composite scRNA-seq datasets in real applications.