NGDE: A Niching-Based Gradient-Directed Evolution Algorithm for Nonconvex Optimization.

Nonconvex optimization issues are prevalent in machine learning and data science. While gradient-based optimization algorithms can rapidly converge and are dimension-independent, they may, unfortunately, fall into local optimal solutions or saddle points. In contrast, evolutionary algorithms (EAs) gradually adapt the population of solutions to explore global optimal solutions. However, this approach requires substantial computational resources to perform numerous fitness function evaluations, which poses challenges for high-dimensional optimization in particular. This study introduces a novel nonconvex optimization algorithm, the niching-based gradient-directed evolution (NGDE) algorithm, designed specifically for high-dimensional nonconvex optimization. The NGDE algorithm generates potential solutions and divides them into multiple niches to explore distinct areas within the feasible region. Subsequently, each individual creates candidate offspring using the gradient-directed mutation operator we designed. The convergence properties of the NGDE algorithm are investigated in two scenarios: accessing the full gradient and approximating the gradient with mini-batch samples. The experimental studies demonstrate the superior performance of the NGDE algorithm in minimizing multimodal optimization functions. Additionally, when applied to train the neural networks of LeNet-5, NGDE shows significantly improved classification accuracy, especially in smaller training sizes.

Identification and validation of a cellular senescence-related lncRNA signature for prognostic prediction in patients with multiple myeloma.

Multiple myeloma (MM) is the second most common hematologic malignancy, which primarily occurs in the elderly. Cellular senescence is considered to be closely associated with the occurrence and progression of malignant tumors including MM, and lncRNA can mediate the process of cellular senescence by regulating key signaling pathways such as p53/p21 and p16/RB. However, the role of cellular senescence related lncRNAs (CSRLs) in MM development has never been reported. Herein, we identified 11 CSRLs (AC004918.5, AC103858.1, AC245100.4, ACBD3-AS1, AL441992.2, ATP2A1-AS1, CCDC18-AS1, LINC00996, TMEM161B-AS1, RP11-706O15.1, and SMURF2P1) to build the CSRLs risk model, which was confirmed to be highly associated with overall survival (OS) of MM patients. We further demonstrated the strong prognostic value of the risk model in MM patients receiving different regimens, especially for those with three-drug combination of bortezomib, lenalidomide, and dexamethasone (VRd) as first-line therapy. Not only that, our risk model also excels in predicting the OS of MM patients at 1, 2, and 3 years. In order to verify the function of these CSRLs in MM, we selected the lncRNA ATP2A1-AS1 which presented the largest expression difference between high-risk groups and low-risk groups for subsequent analysis and validation. Finally, we found that down-regulation of ATP2A1-AS1 can promote cellular senescence in MM cell lines. In conclusion, the CSRLs risk model established in present study provides a novel and more accurate method for predicting MM patients' prognosis and identifies a new target for MM therapeutic intervention.

Research advances and treatment perspectives of pancreatic adenosquamous carcinoma.

BACKGROUND: As a malignant tumor, pancreatic cancer has an extremely low overall 5-year survival rate. Pancreatic adenosquamous carcinoma (PASC), a rare pancreatic malignancy, owns clinical presentation similar to pancreatic ductal adenocarcinoma (PDAC), which is the most prevalent pancreatic cancer subtype. PASC is generally defined as a pancreatic tumor consisting mainly of adenocarcinoma tissue and squamous carcinoma tissue. Compared with PDAC, PASC has a higher metastatic potential and worse prognosis, and lacks of effective treatment options to date. However, the pathogenesis and treatment of PASC are not yet clear and are accompanied with difficulties. CONCLUSION: The present paper systematically summarizes the possible pathogenesis, diagnosis methods, and further suggests potential new treatment directions through reviewing research results of PASC, including the clinical manifestations, pathological manifestation, the original hypothesis of squamous carcinoma and the potential regulatory mechanism. In short, the present paper provides a systematic review of the research progress and new ideas for the development mechanism and treatment of PASC.

Clinical application and prospect of immune checkpoint inhibitors for CAR-NK cell in tumor immunotherapy.

Chimeric antigen receptor (CAR) engineering of natural killer (NK) cells is an attractive research field in tumor immunotherapy. While CAR is genetically engineered to express certain molecules, it retains the intrinsic ability to recognize tumor cells through its own receptors. Additionally, NK cells do not depend on T cell receptors for cytotoxic killing. CAR-NK cells exhibit some differences to CAR-T cells in terms of more precise killing, numerous cell sources, and increased effectiveness in solid tumors. However, some problems still exist with CAR-NK cell therapy, such as cytotoxicity, low transfection efficiency, and storage issues. Immune checkpoints inhibit immune cells from performing their normal killing function, and the clinical application of immune checkpoint inhibitors for cancer treatment has become a key therapeutic strategy. The application of CAR-T cells and immune checkpoint inhibitors is being evaluated in numerous ongoing basic research and clinical studies. Immune checkpoints may affect the function of CAR-NK cell therapy. In this review, we describe the combination of existing CAR-NK cell technology with immune checkpoint therapy and discuss the research of CAR-NK cell technology and future clinical treatments. We also summarize the progress of clinical trials of CAR-NK cells and immune checkpoint therapy.

m6A RNA methylation-mediated HNF3γ reduction renders hepatocellular carcinoma dedifferentiation and sorafenib resistance.

Hepatocyte nuclear factor 3γ (HNF3γ) is a hepatocyte nuclear factor, but its role and clinical significance in hepatocellular carcinoma (HCC) remain unclear. Herein, we report that HNF3γ expression is downregulated in patient HCC and inversely correlated with HCC malignancy and patient survival. Moreover, our data suggested that the HNF3γ reduction in HCC could be mediated by METTL14-dependent m6A methylation of HNF3γ mRNA. HNF3γ expression was increased during hepatic differentiation and decreased in dedifferentiated HCC cells. Interestingly, HNF3γ delivery promoted differentiation of not only HCC cells but also liver CSCs, which led to suppression of HCC growth. Mechanistic analysis suggested an HNF3γ-centered regulatory network that includes essential liver differentiation-associated transcription factors and functional molecules, which could synergistically facilitate HCC cell differentiation. More importantly, enforced HNF3γ expression sensitized HCC cells to sorafenib-induced growth inhibition and cell apoptosis through transactivation of OATP1B1 and OATP1B3 expression, which are major membrane transporters for sorafenib uptake. Clinical investigation showed that patient-derived HCC xenografts with high HNF3γ expression exhibited a sorafenib response and patients with high HCC HNF3γ levels benefited from sorafenib therapy. Together, these results suggest that HNF3γ plays an essential role in HCC differentiation and may serve as a therapeutic target and predictor of sorafenib benefit in patients.

A cost-sensitive online learning method for peptide identification.

BACKGROUND: Post-database search is a key procedure in peptide identification with tandem mass spectrometry (MS/MS) strategies for refining peptide-spectrum matches (PSMs) generated by database search engines. Although many statistical and machine learning-based methods have been developed to improve the accuracy of peptide identification, the challenge remains on large-scale datasets and datasets with a distribution of unbalanced PSMs. A more efficient learning strategy is required for improving the accuracy of peptide identification on challenging datasets. While complex learning models have larger power of classification, they may cause overfitting problems and introduce computational complexity on large-scale datasets. Kernel methods map data from the sample space to high dimensional spaces where data relationships can be simplified for modeling. RESULTS: In order to tackle the computational challenge of using the kernel-based learning model for practical peptide identification problems, we present an online learning algorithm, OLCS-Ranker, which iteratively feeds only one training sample into the learning model at each round, and, as a result, the memory requirement for computation is significantly reduced. Meanwhile, we propose a cost-sensitive learning model for OLCS-Ranker by using a larger loss of decoy PSMs than that of target PSMs in the loss function. CONCLUSIONS: The new model can reduce its false discovery rate on datasets with a distribution of unbalanced PSMs. Experimental studies show that OLCS-Ranker outperforms other methods in terms of accuracy and stability, especially on datasets with a distribution of unbalanced PSMs. Furthermore, OLCS-Ranker is 15-85 times faster than CRanker.

Coupling of COPII vesicle trafficking to nutrient availability by the IRE1α-XBP1s axis.

The cytoplasmic coat protein complex-II (COPII) is evolutionarily conserved machinery that is essential for efficient trafficking of protein and lipid cargos. How the COPII machinery is regulated to meet the metabolic demand in response to alterations of the nutritional state remains largely unexplored, however. Here, we show that dynamic changes of COPII vesicle trafficking parallel the activation of transcription factor X-box binding protein 1 (XBP1s), a critical transcription factor in handling cellular endoplasmic reticulum (ER) stress in both live cells and mouse livers upon physiological fluctuations of nutrient availability. Using live-cell imaging approaches, we demonstrate that XBP1s is sufficient to promote COPII-dependent trafficking, mediating the nutrient stimulatory effects. Chromatin immunoprecipitation (ChIP) coupled with high-throughput DNA sequencing (ChIP-seq) and RNA-sequencing analyses reveal that nutritional signals induce dynamic XBP1s occupancy of promoters of COPII traffic-related genes, thereby driving the COPII-mediated trafficking process. Liver-specific disruption of the inositol-requiring enzyme 1α (IRE1α)-XBP1s signaling branch results in diminished COPII vesicle trafficking. Reactivation of XBP1s in mice lacking hepatic IRE1α restores COPII-mediated lipoprotein secretion and reverses the fatty liver and hypolipidemia phenotypes. Thus, our results demonstrate a previously unappreciated mechanism in the metabolic control of liver protein and lipid trafficking: The IRE1α-XBP1s axis functions as a nutrient-sensing regulatory nexus that integrates nutritional states and the COPII vesicle trafficking.

Mitophagy Directs Muscle-Adipose Crosstalk to Alleviate Dietary Obesity.

The quality of mitochondria in skeletal muscle is essential for maintaining metabolic homeostasis during adaptive stress responses. However, the precise control mechanism of muscle mitochondrial quality and its physiological impacts remain unclear. Here, we demonstrate that FUNDC1, a mediator of mitophagy, plays a critical role in controlling muscle mitochondrial quality as well as metabolic homeostasis. Skeletal-muscle-specific ablation of FUNDC1 in mice resulted in LC3-mediated mitophagy defect, leading to impaired mitochondrial energetics. This caused decreased muscle fat utilization and endurance capacity during exercise. Interestingly, mice lacking muscle FUNDC1 were protected against high-fat-diet-induced obesity with improved systemic insulin sensitivity and glucose tolerance despite reduced muscle mitochondrial energetics. Mechanistically, FUNDC1 deficiency elicited a retrograde response in muscle that upregulated FGF21 expression, thereby promoting the thermogenic remodeling of adipose tissue. Thus, these findings reveal a pivotal role of FUNDC1-dependent mitochondrial quality control in mediating the muscle-adipose dialog to regulate systemic metabolism.

Budget Online Learning Algorithm for Least Squares SVM.

Batch-mode least squares support vector machine (LSSVM) is often associated with unbounded number of support vectors (SVs'), making it unsuitable for applications involving large-scale streaming data. Limited-scale LSSVM, which allows efficient updating, seems to be a good solution to tackle this issue. In this paper, to train the limited-scale LSSVM dynamically, we present a budget online LSSVM (BOLSSVM) algorithm. Methodologically, by setting a fixed budget for SVs', we are able to update the LSSVM model according to the updated SVs' set dynamically without retraining from scratch. In particular, when a new small chunk of SVs' substitute for the old ones, the proposed algorithm employs a low rank correction technology and the Sherman-Morrison-Woodbury formula to compute the inverse of saddle point matrix derived from the LSSVM's Karush-Kuhn-Tucker (KKT) system, which, in turn, updates the LSSVM model efficiently. In this way, the proposed BOLSSVM algorithm is especially useful for online prediction tasks. Another merit of the proposed BOLSSVM is that it can be used for k -fold cross validation. Specifically, compared with batch-mode learning methods, the computational complexity of the proposed BOLSSVM method is significantly reduced from O(n4) to O(n3) for leave-one-out cross validation with n training samples. The experimental results of classification and regression on benchmark data sets and real-world applications show the validity and effectiveness of the proposed BOLSSVM algorithm.

Exercise Inducible Lactate Dehydrogenase B Regulates Mitochondrial Function in Skeletal Muscle.

Lactate dehydrogenase (LDH) catalyzes the interconversion of pyruvate and lactate, which are critical fuel metabolites of skeletal muscle particularly during exercise. However, the physiological relevance of LDH remains poorly understood. Here we show that Ldhb expression is induced by exercise in human muscle and negatively correlated with changes in intramuscular pH levels, a marker of lactate production, during isometric exercise. We found that the expression of Ldhb is regulated by exercise-induced peroxisome proliferator-activated receptor γ coactivator 1α (PGC-1α). Ldhb gene promoter reporter studies demonstrated that PGC-1α activates Ldhb gene expression through multiple conserved estrogen-related receptor (ERR) and myocyte enhancer factor 2 (MEF2) binding sites. Transgenic mice overexpressing Ldhb in muscle (muscle creatine kinase (MCK)-Ldhb) exhibited increased exercise performance and enhanced oxygen consumption during exercise. MCK-Ldhb muscle was shown to have enhanced mitochondrial enzyme activity and increased mitochondrial gene expression, suggesting an adaptive oxidative muscle transformation. In addition, mitochondrial respiration capacity was increased and lactate production decreased in MCK-Ldhb skeletal myotubes in culture. Together, these results identified a previously unrecognized Ldhb-driven alteration in muscle mitochondrial function and suggested a mechanism for the adaptive metabolic response induced by exercise training.

Coupling of mitochondrial function and skeletal muscle fiber type by a miR-499/Fnip1/AMPK circuit.

Upon adaption of skeletal muscle to physiological and pathophysiological stimuli, muscle fiber type and mitochondrial function are coordinately regulated. Recent studies have identified pathways involved in control of contractile proteins of oxidative-type fibers. However, the mechanism for coupling of mitochondrial function to the muscle contractile machinery during fiber type transition remains unknown. Here, we show that the expression of the genes encoding type I myosins, Myh7/Myh7b and their intronic miR-208b/miR-499, parallels mitochondrial function during fiber type transitions. Using in vivo approaches in mice, we found that miR-499 drives a PGC-1α-dependent mitochondrial oxidative metabolism program to match shifts in slow-twitch muscle fiber composition. Mechanistically, miR-499 directly targets Fnip1, an AMP-activated protein kinase (AMPK)-interacting protein that negatively regulates AMPK, a known activator of PGC-1α. Inhibition of Fnip1 reactivated AMPK/PGC-1α signaling and mitochondrial function in myocytes. Restoration of the expression of miR-499 in the mdx mouse model of Duchenne muscular dystrophy (DMD) reduced the severity of DMD Thus, we have identified a miR-499/Fnip1/AMPK circuit that can serve as a mechanism to couple muscle fiber type and mitochondrial function.

l2 Multiple Kernel Fuzzy SVM-Based Data Fusion for Improving Peptide Identification.

SEQUEST is a database-searching engine, which calculates the correlation score between observed spectrum and theoretical spectrum deduced from protein sequences stored in a flat text file, even though it is not a relational and object-oriental repository. Nevertheless, the SEQUEST score functions fail to discriminate between true and false PSMs accurately. Some approaches, such as PeptideProphet and Percolator, have been proposed to address the task of distinguishing true and false PSMs. However, most of these methods employ time-consuming learning algorithms to validate peptide assignments [1] . In this paper, we propose a fast algorithm for validating peptide identification by incorporating heterogeneous information from SEQUEST scores and peptide digested knowledge. To automate the peptide identification process and incorporate additional information, we employ l2 multiple kernel learning (MKL) to implement the current peptide identification task. Results on experimental datasets indicate that compared with state-of-the-art methods, i.e., PeptideProphet and Percolator, our data fusing strategy has comparable performance but reduces the running time significantly.

An adaptive classification model for peptide identification.

BACKGROUND: Peptide sequence assignment is the central task in protein identification with MS/MS-based strategies. Although a number of post-database search algorithms for filtering target peptide spectrum matches (PSMs) have been developed, the discrepancy among the output PSMs is usually significant, remaining a few disputable PSMs. Current studies show that a number of target PSMs which are close to decoy PSMs can hardly be separated from those decoys by only using the discrimination function. RESULTS: In this paper, we assign each target PSM a weight showing its possibility of being correct. We employ a SVM-based learning model to search the optimal weight for each target PSM and develop a new score system, CRanker, to rank all target PSMs. Due to the large PSM datasets generated in routine database searches, we use the Cholesky factorization technique for storing a kernel matrix to reduce the memory requirement. CONCLUSIONS: Compared with PeptideProphet and Percolator, CRanker has identified more PSMs under similar false discover rates over different datasets. CRanker has shown consistent performance on different test sets, validated the reasonability the proposed model.

High Throughput Sequencing Identifies MicroRNAs Mediating α-Synuclein Toxicity by Targeting Neuroactive-Ligand Receptor Interaction Pathway in Early Stage of Drosophila Parkinson's Disease Model.

Parkinson's disease (PD) is a prevalent neurodegenerative disorder with pathological features including death of dopaminergic neurons in the substantia nigra and intraneuronal accumulations of Lewy bodies. As the main component of Lewy bodies, α-synuclein is implicated in PD pathogenesis by aggregation into insoluble filaments. However, the detailed mechanisms underlying α-synuclein induced neurotoxicity in PD are still elusive. MicroRNAs are ~20nt small RNA molecules that fine-tune gene expression at posttranscriptional level. A plethora of miRNAs have been found to be dysregulated in the brain and blood cells of PD patients. Nevertheless, the detailed mechanisms and their in vivo functions in PD still need further investigation. By using Drosophila PD model expressing α-synuclein A30P, we examined brain miRNA expression with high-throughput small RNA sequencing technology. We found that five miRNAs (dme-miR-133-3p, dme-miR-137-3p, dme-miR-13b-3p, dme-miR-932-5p, dme-miR-1008-5p) were upregulated in PD flies. Among them, miR-13b, miR-133, miR-137 are brain enriched and highly conserved from Drosophila to humans. KEGG pathway analysis using DIANA miR-Path demonstrated that neuroactive-ligand receptor interaction pathway was most likely affected by these miRNAs. Interestingly, miR-137 was predicted to regulate most of the identified targets in this pathway, including dopamine receptor (DopR, D2R), γ-aminobutyric acid (GABA) receptor (GABA-B-R1, GABA-B-R3) and N-methyl-D-aspartate (NMDA) receptor (Nmdar2). The validation experiments showed that the expression of miR-137 and its targets was negatively correlated in PD flies. Further experiments using luciferase reporter assay confirmed that miR-137 could act on specific sites in 3' UTR region of D2R, Nmdar2 and GABA-B-R3, which downregulated significantly in PD flies. Collectively, our findings indicate that α-synuclein could induce the dysregulation of miRNAs, which target neuroactive ligand-receptor interaction pathway in vivo. We believe it will help us further understand the contribution of miRNAs to α-synuclein neurotoxicity and provide new insights into the pathogenesis driving PD.

Transcriptional regulatory circuits controlling muscle fiber type switching.

Skeletal muscle fitness plays vital roles in human health and disease and is determined by developmental as well as physiological inputs. These inputs control and coordinate muscle fiber programs, including capacity for fuel burning, mitochondrial ATP production, and contraction. Recent studies have demonstrated crucial roles for nuclear receptors and their co-activators, and microRNAs (miRNAs) in the regulation of skeletal muscle energy metabolism and fiber type determination. In this review, we present recent progress in the study of nuclear receptor signaling and miRNA networks in muscle fiber type switching. We also discuss the therapeutic potential of nuclear receptors and miRNAs in disease states that are associated with loss of muscle fitness.

[Three-column reconstruction via posterior approach for the treatment of unstable thoracolumbar fracture accompanied by posterior column injury].

OBJECTIVE: To investigate clinical effects of three-column reconstruction via single posterior approach for the treatment of unstable thoracolumbar fractures accompanied by posterior column injury. METHODS: From December 2008 to May 2010,three-column reconstruction via posterior approach was implemented to 21 patients with unstable thoracolumbar fractures accompanied by posterior column injuries. There were 13 males and 8 females, ranging in age from 23 to 54 years old(averaged,35.5 years old). Injured vertebrae: 1 patient had injury in T11, 4 patients had injuries in T12, 8 patients had injuries in L1, 5 patients had injuries in L2, 3 patients had injuries in L3. The Cobb angle was (25.34 +/- 3.42) degrees. The operation time,blood loss during operation, Cobb angle and the bony fusion were observed. RESULTS: Twenty-one patients were followed up, and the duration ranged from 24 to 27 years old, with an average of 25.6 months. The operation time ranged from 135 to 275 min, with a mean of 185 min. The blood loss during operation ranged from 700 to 1 650 ml (averaged, 870 ml). All the patients had complete decompression. Postoperative Cobb angle was (4.01 +/- 2.03) degrees, and (4.34 +/- 2.38) degrees at the latest follow-up. All the patients got bony fusion. CONCLUSION: To the patients with unstable thoracolumbar fractures accompanied by posterior column injuries, three-column reconstruction via single posterior approach has both anterior approach and posterior approach advantages, which can obtain excellent clinical outcomes.

Peptide identification based on fuzzy classification and clustering.

BACKGROUND: The sequence database searching has been the dominant method for peptide identification, in which a large number of peptide spectra generated from LC/MS/MS experiments are searched using a search engine against theoretical fragmentation spectra derived from a protein sequences database or a spectral library. Selecting trustworthy peptide spectrum matches (PSMs) remains a challenge. RESULTS: A novel scoring method named FC-Ranker is developed to assign a nonnegative weight to each target PSM based on the possibility of its being correct. Particularly, the scores of PSMs are updated by using a fuzzy SVM classification model and a fuzzy silhouette index iteratively. Trustworthy PSMs will be assigned high scores when the algorithm stops. CONCLUSIONS: Our experimental studies show that FC-Ranker outperforms other post-database search algorithms over a variety of datasets, and it can be extended to solve a general classification problem with uncertain labels.

Inference of gene regulatory subnetworks from time course gene expression data.

BACKGROUND: Identifying gene regulatory network (GRN) from time course gene expression data has attracted more and more attentions. Due to the computational complexity, most approaches for GRN reconstruction are limited on a small number of genes and low connectivity of the underlying networks. These approaches can only identify a single network for a given set of genes. However, for a large-scale gene network, there might exist multiple potential sub-networks, in which genes are only functionally related to others in the sub-networks. RESULTS: We propose the network and community identification (NCI) method for identifying multiple subnetworks from gene expression data by incorporating community structure information into GRN inference. The proposed algorithm iteratively solves two optimization problems, and can promisingly be applied to large-scale GRNs. Furthermore, we present the efficient Block PCA method for searching communities in GRNs. CONCLUSIONS: The NCI method is effective in identifying multiple subnetworks in a large-scale GRN. With the splitting algorithm, the Block PCA method shows a promosing attempt for exploring communities in a large-scale GRN.

Design of a multiple kernel learning algorithm for LS-SVM by convex programming.

As a kernel based method, the performance of least squares support vector machine (LS-SVM) depends on the selection of the kernel as well as the regularization parameter (Duan, Keerthi, & Poo, 2003). Cross-validation is efficient in selecting a single kernel and the regularization parameter; however, it suffers from heavy computational cost and is not flexible to deal with multiple kernels. In this paper, we address the issue of multiple kernel learning for LS-SVM by formulating it as semidefinite programming (SDP). Furthermore, we show that the regularization parameter can be optimized in a unified framework with the kernel, which leads to an automatic process for model selection. Extensive experimental validations are performed and analyzed.