Integration of clinical phenoms and metabolomics facilitates precision medicine for lung cancer.

Lung cancer is a common malignancy that is frequently associated with systemic metabolic disorders. Early detection is pivotal to survival improvement. Although blood biomarkers have been used in its early diagnosis, missed diagnosis and misdiagnosis still exist due to the heterogeneity of lung cancer. Integration of multiple biomarkers or trans-omics results can improve the accuracy and reliability for lung cancer diagnosis. As metabolic reprogramming is a hallmark of lung cancer, metabolites, specifically lipids might be useful for lung cancer detection, yet systematic characterizations of metabolites in lung cancer are still incipient. The present study profiled the polar metabolome and lipidome in the plasma of lung cancer patients to construct an inclusive metabolomic atlas of lung cancer. A comprehensive analysis of lung cancer was also conducted combining metabolomics with clinical phenotypes. Furthermore, the differences in plasma lipid metabolites were compared and analyzed among different lung cancer subtypes. Alcohols, amides, and peptide metabolites were significantly increased in lung cancer, while carboxylic acids, hydrocarbons, and fatty acids were remarkably decreased. Lipid profiling revealed a significant increase in plasma levels of CER, PE, SM, and TAG in individuals with lung cancer as compared to those in healthy controls. Correlation analysis confirmed the association between a panel of metabolites and TAGs. Clinical trans-omics studies elucidated the complex correlations between lipidomic data and clinical phenotypes. The present study emphasized the clinical importance of lipidomics in lung cancer, which involves the correlation between metabolites and the expressions of other omics, ultimately influencing clinical phenotypes. This novel trans-omics network approach would facilitate the development of precision therapy for lung cancer.

A lysing polysaccharide monooxygenase from Aspergillus niger effectively facilitated rumen microbial fermentation of rice straw.

Objective: This study investigated the impact of Aspergillus niger lysing polysaccharide monooxygenase (AnLPMO) on in vitro rumen microbial fermentation of rice straw. Methods: AnLPMO was heterologously expressed in Escherichia coli. Fourier transform infrared spectrometry and X-ray photoelectron spectroscopy analyzed the surface structure of rice straw after AnLPMO treatment. Two in vitro experiments, coupled with 16S high-throughput sequencing and qRT-PCR techniques, assessed the influence of AnLPMO on rumen microbial fermentation of rice straw. Results: AnLPMO exhibited peak activity at 40 °C and pH 6.5, with a preference for rice straw xylan hydrolysis, followed by Avicel. AnLPMO application led to the fractional removal of cellulose and hemicelluloses and a notable reduction in the levels of carbon elements and C-C groups present on the surface of rice straw. Compared to the control (no AnLPMO), supplementing AnLPMO at 1.1 U-2.0 U significantly enhanced in vitro digestibility of dry matter (IVDMD, P < 0.01), total gas production (P < 0.01), and concentrations of total volatile fatty acids (VFA, P < 0.01), acetate (P < 0.01), and ammonia-N (P < 0.01). Particularly, the 1.4 U AnLPMO group showed a 14.8% increase in IVDMD. In the second experiment, compared to deactivated AnLPMO (1.4 U), supplementing bioactive AnLPMO at 1.4 U increased IVDMD (P = 0.01), total gas production (P = 0.04), and concentrations of total VFA (P < 0.01), propionate (P < 0.01), and ammonia-N (P < 0.01), with a limited 9.6% increase in IVDMD. Supplementing AnLPMO stimulated the growth of ruminal bacterial taxa facilitating fiber degradation, including Proteobacteria, Spirochaetes, Succinivibrio, Rikenellaceae_RC9_Gut_Group, Prevotelaceae_UCG-003, Desulfovibrio, Fibrobacter succinogenes, Ruminococcus albus, R. flavefaciens, Prevotella bryantii, P. ruminicola, and Treponema bryantii. Conclusion: These findings highlight AnLPMO's potential as a feed additive for improving rice straw utilization in ruminant production.

Boosting Reinforcement Learning via Hierarchical Game Playing With State Relay.

Due to its wide application, deep reinforcement learning (DRL) has been extensively studied in the motion planning community in recent years. However, in the current DRL research, regardless of task completion, the state information of the agent will be reset afterward. This leads to a low sample utilization rate and hinders further explorations of the environment. Moreover, in the initial training stage, the agent has a weak learning ability in general, which affects the training efficiency in complex tasks. In this study, a new hierarchical reinforcement learning (HRL) framework dubbed hierarchical learning based on game playing with state relay (HGR) is proposed. In particular, we introduce an auxiliary penalty to regulate task difficulty, and one training mechanism, the state relay mechanism, is designed. The relay mechanism can make full use of the intermediate states of the agent and expand the environment exploration of low-level policy. Our algorithm can improve the sample utilization rate, reduce the sparse reward problem, and thereby enhance the training performance in complex environments. Simulation tests are carried out on two public experiment platforms, i.e., MazeBase and MuJoCo, to verify the effectiveness of the proposed method. The results show that HGR significantly benefits the reinforcement learning (RL) area.

TSVM: Transfer Support Vector Machine for Predicting MPRA Validated Regulatory Variants.

Genome-wide association studies have shown that common genetic variants associated with complex diseases are mostly located in non-coding regions, which may not be causal. In addition, the limited number of validated non-coding functional variants makes it difficult to develop an effective supervised learning model. Therefore, improving the accuracy of predicting non-coding causal variants has become critical. This study aims to build a transfer learning-based machine learning method for predicting regulatory variants to overcome the problem of limited sample size. This paper presents a supervised learning method transfer support vector machine (TSVM) for massively parallel reporter assays (MPRA) validated regulatory variants prediction. First, uses a convolutional neural network to extract features with transfer learning. Second, the extracted features are selected by random forest method. Third, the selected features are used to train support vector machine for classification. We performed scale sensitivity experiments on the MPRA dataset and validated the effectiveness of transfer learning. The model achieves the Mcc of 0.326 and the AUC of 0.720, which are higher than the state-of-the-art method.

Progress of researches on mechanisms of acupuncture underlying improvement of cerebral ischemia-reperfusion injury by regulating calcium overload. / é’ˆåˆºè°ƒæŽ§è„‘ç¼ºè¡€å†çŒæ³¨æŸä¼¤åŽé’™è¶ è½½æœºåˆ¶ç ”ç©¶è¿›å±•.

Ischemic stroke is currently the most common type of stroke, and the key pathological link is cerebral ischemia-reperfusion injury (CIRI), while the key factor leading to apoptosis and necrosis of ischemic nerve cells is calcium overload. Current studies have confirmed that acupuncture therapy has a good modulating effect on calcium homeostasis and can reduce cerebral ischemia-reperfusion induced damage of neuronal cells by inhibiting calcium overload. After reviewing the relevant literature published in the past 15 years, we find that acupuncture plays a role in regulating the pathological mechanism of calcium overload after CIRI by inhibiting the opening of connexin 43 hemichannels, regulating the intracellular free calcium ion concentration, suppressing the expression of calmodulin, and blocking the function of L-type voltage-gated calcium channels, thereby inhibiting calcium overload, regulating calcium homeostasis and antagonizing neuronal damage resulted from cerebral ischemia-reperfusion, which may provide ideas for future research.

Playing Extensive Games with Learning of Opponent's Cognition.

Decision-making is a basic component of agents' (e.g., intelligent sensors) behaviors, in which one's cognition plays a crucial role in the process and outcome. Extensive games, a class of interactive decision-making scenarios, have been studied in diverse fields. Recently, a model of extensive games was proposed in which agent cognition of the structure of the underlying game and the quality of the game situations are encoded by artificial neural networks. This model refines the classic model of extensive games, and the corresponding equilibrium concept-cognitive perfect equilibrium (CPE)-differs from the classic subgame perfect equilibrium, since CPE takes agent cognition into consideration. However, this model neglects the consideration that game-playing processes are greatly affected by agents' cognition of their opponents. To this end, in this work, we go one step further by proposing a framework in which agents' cognition of their opponents is incorporated. A method is presented for evaluating opponents' cognition about the game being played, and thus, an algorithm designed for playing such games is analyzed. The resulting equilibrium concept is defined as adversarial cognition equilibrium (ACE). By means of a running example, we demonstrate that the ACE is more realistic than the CPE, since it involves learning about opponents' cognition. Further results are presented regarding the computational complexity, soundness, and completeness of the game-solving algorithm and the existence of the equilibrium solution. This model suggests the possibility of enhancing an agent's strategic ability by evaluating opponents' cognition.

Effects of tomato-Sedum alfredii Hance intercropping on crop production and Cd remediation as affected by soil types.

Intercropping crops with hyperaccumulators is a proven model for coupling crop safety production and soil heavy metal remediation. And both crop genotypes and soil properties might have great impacts on the effect of intercropping. Therefore, a greenhouse pot experiment was designed to investigate the effects of intercropping different tomato varieties with the cadmium (Cd) hyperaccumulator Sedum alfredii Hance (S. alfredii Hance) on different soils. The results showed that intercropping promoted Cd uptake by S. alfredii Hance and reduced soil total Cd concentration. There was no significant effect of intercropping on tomato yield and Cd concentration. Different tomato varieties had different effects on tomato yield and Cd concentration. The yield of cherry tomato was 1.04 times higher than that of common large fruit tomato, while the Cd concentration in all parts was lower than that of common large fruit tomato. Different typical zonal soils had different effects on tomato production and soil remediation. And among the four studied soils, tomatoes grown on ZJ soil had the highest yields and lowest fruit Cd concentration, making them more suitable for remediation coupled with safety production. This study provided a comprehensive analysis of tomato production benefits and soil remediation effects, which could be useful as a guide in vegetable safety production coupled with soil remediation practices in the Cd-contaminated greenhouse.

Photothermal synergistic nitric oxide controlled release injectable self-healing adhesive hydrogel for biofilm eradication and wound healing.

The development of injectable self-healing adhesive hydrogel dressings with excellent bactericidal activity and wound healing ability is urgently in demand for combating biofilm infections. Herein, a multifunctional hydrogel (QP/QT-MB) with near-infrared (NIR) light-activated mild photothermal/gaseous antimicrobial activity was developed based on the dynamic reversible borate bonds and hydrogen bonds crosslinking between quaternization chitosan (QCS) derivatives alternatively containing phenylboronic acid and catechol-like moieties in conjunction with the in situ encapsulation of BNN6-loaded mesoporous polydopamine (MPDA@BNN6 NPs). Given the dynamic reversible cross-linking feature, the versatile hybrid hydrogel exhibited injectability, flexibility, and rapid self-healing ability. The numerous phenylboronic acid and catechol-like moieties on the QCS backbone confer the hydrogel with specific bacterial affinity, desirable tissue adhesion, and antioxidant stress ability that enhance bactericidal activity and facilitate the regeneration of infection wounds. Under NIR irradiation, the QP/QT-MB hydrogels exhibited a desirable mild photothermal effect and NIR-activity controllable NO delivery, combined with the endogenous contact antimicrobial activity of hydrogel, contributing jointly to induce dispersal of biofilms and disruption of the bacterial plasma membranes, ultimately leading to bacteria inactivation and biofilm elimination. In vivo experiments demonstrated that the fabricated QP/QT-MB hydrogel platform was capable of inducing efficient eradication of the S. aureus biofilm in a severely infected wound model and accelerating infected wound repair by promoting collagen deposition, angiogenesis, and suppressing inflammatory responses. Additionally, the QP/QT-MB hydrogel demonstrated excellent biocompatibility in vitro and in vivo. Collectively, the hydrogel (QP/QT-MB) reveals great potential application prospects as a promising alternative in the field of biofilm-associated infection treatment.

TIVC: An Efficient Local Search Algorithm for Minimum Vertex Cover in Large Graphs.

The minimum vertex cover (MVC) problem is a canonical NP-hard combinatorial optimization problem aiming to find the smallest set of vertices such that every edge has at least one endpoint in the set. This problem has extensive applications in cybersecurity, scheduling, and monitoring link failures in wireless sensor networks (WSNs). Numerous local search algorithms have been proposed to obtain "good" vertex coverage. However, due to the NP-hard nature, it is challenging to efficiently solve the MVC problem, especially on large graphs. In this paper, we propose an efficient local search algorithm for MVC called TIVC, which is based on two main ideas: a 3-improvements (TI) framework with a tiny perturbation and edge selection strategy. We conducted experiments on real-world large instances of a massive graph benchmark. Compared with three state-of-the-art MVC algorithms, TIVC shows superior performance in accuracy and possesses a remarkable ability to identify significantly smaller vertex covers on many graphs.

Analysis of factors influencing the organizational capacity of Institutional Review Boards In China: a crisp-set Qualitative Comparative Analysis based on 107 cases.

BACKGROUND: Institutional Review Boards (IRBs) play a vital role in safeguarding the rights and interests of both research participants and researchers. However, China initiated the establishment of its own IRB system relatively late in comparison to international standards. Despite commendable progress, there is a pressing need to strengthen the organizational capacity building of Chinese IRBs. Hence, this study aims to analyze the key factors driving the enhancement of organizational capacity within these committees. METHOD: The cross-sectional survey for this research was conducted from July 2020 to January 2022. Following the statistical grouping based on the "2020 China Health Statistical Yearbook", a systematic investigation of IRBs in various provinces of China was carried out. In-depth interviews and questionnaire surveys were conducted with the chairpersons and administrative executives (or secretaries) of highly cooperative IRBs. Subsequently, data were collected from 107 IRBs. Qualitative Comparative Analysis (QCA) was employed as the method to analyze the factors influencing the organizational capacity of medical ethics committees and explore the diverse combinations of these factors. RESULTS: Through a singular necessary condition analysis, the variable "protection of rights and interests" emerges as a critical factor contributing to the robust construction of Institutional Review Boards Institutional Review Boards (IRBs). Conversely, the variables of "lack of member ability, absence of review process, and deficiency in the supervision mechanism" collectively constitute a sufficient condition leading to weaker IRB construction. The state analysis uncovers three interpretation modes: member ability-oriented (M1), system process-oriented mode (M2), and resource system-oriented mode (M3). CONCLUSIONS: The results of this study are effectively explicable using the "Triangular Force" model proposed for the hypothesis of IRBs' organizational capacity, which provides a solid foundation for the development of organizational capabilities in IRBs. To enhance the organizational capacity of IRBs in China, it is imperative to elevate the competence of committee members and strengthen team development. This can be achieved by establishing a comprehensive regulatory framework and refining procedural protocols. Moreover, clarifying the organizational structure and optimizing resource allocation are essential steps in bolstering the overall organizational capabilities of these committees.

TGF-ß promotes the proliferation and osteogenic differentiation of dental pulp stem cells a systematic review and meta-analysis.

BACKGROUND: Dental pulp stem cells (DPSCs) are adult stem cells with multi-directional differentiation potential derived from ectoderm. Vitro experiments have shown that adding cytokines can help DPSCs to be transformed from multipotent stem cells to osteoblasts. TGF-ß has been proved to have an effect on the proliferation and mineralization of bone tissue, but its effect on the osteogenesis and proliferation of dental pulp stem cells is still uncertain. We aim to determine the effect of TGF-ß on the osteogenesis and proliferation of dental pulp stem cells. METHODS: We have identified studies from the Cochrane Central Register of Controlled Trials, PubMed, Embase, and China national knowledge infrastructure (CNKI) for studies interested in TGF-ß and proliferation and differentiation of dental pulp stem cells in the following indicators: A490 (an index for evaluating cell proliferation), bone sialoprotein (BSP), Col plasmid-1 (Col-1), osteocalcin (OCN), runt-related transcription factor 2 (Runx-2); and the number of mineralized nodules. Any language restrictions were rejected. Furthermore, we drew a forest plot for each outcome. We conducted a sensitivity analysis, data analysis, heterogeneity, and publication bias test. We evaluate the quality of each study under the guidance of Cochrane's tool for quality assessment. RESULTS: The pooled data showed that TGF-ß could promote the proliferation and ossification of dental pulp stem cells. All the included results support this conclusion except for the number of mineralized nodules: TGF-ß increases the A490 index (SMD 3.11, 95% CI [0.54-5.69]), promotes the production of BSP (SMD 3.11, 95% CI [0.81-6.77]), promotes the expression of Col-1 (SMD 4.71, 95% CI [1.25-8.16]) and Runx-2 (SMD 3.37, 95% CI [- 0.63 to 7.36]), increases the content of OCN (SMD 4.32, 95% CI [1.20-7.44]) in dental pulp, and has no significant effect on the number of mineralized nodules (SMD 3.87, 95% CI [- 1.76 to 9.51]) in dental pulp stem cells. CONCLUSIONS: TGF-ß promotes the proliferation and osteogenesis of dental pulp stem cells.

CLDN18.2 and 4-1BB bispecific antibody givastomig exerts antitumor activity through CLDN18.2-expressing tumor-directed T-cell activation.

BACKGROUND: Claudin18.2 (CLDN18.2) is a tight junction protein that has been identified as a clinically proven target in gastric cancer. Stimulation of 4-1BB with agonistic antibodies is also a promising strategy for immunotherapy and 4-1BB+ T cells were reported to be present within the tumor microenvironment of patients with gastric cancer. However, hepatotoxicity-mediated by 4-1BB activation was observed in clinical trials of agonistic anti-4-1BB monoclonal antibodies. METHODS: To specifically activate the 4-1BB+ T cells in tumor and avoid the on-target liver toxicity, we developed a novel CLDN18.2×4-1BB bispecific antibody (termed 'givastomig' or 'ABL111'; also known as TJ-CD4B or TJ033721) that was designed to activate 4-1BB signaling in a CLDN18.2 engagement-dependent manner. RESULTS: 4-1BB+ T cells were observed to be coexisted with CLDN18.2+ tumor cells in proximity by multiplex immunohistochemical staining of tumor tissues from patients with gastric cancer (n=60). Givastomig/ABL111 could bind to cell lines expressing various levels of CLDN18.2 with a high affinity and induce 4-1BB activation in vitro only in the context of CLDN18.2 binding. The magnitude of T-cell activation by givastomig/ABL111 treatment was closely correlated with the CLDN18.2 expression level of tumor cells from gastric cancer patient-derived xenograft model. Mechanistically, givastomig/ABL111 treatment could upregulate the expression of a panel of pro-inflammatory and interferon-γ-responsive genes in human peripheral blood mononuclear cells when co-cultured with CLDN18.2+ tumor cells. Furthermore, in humanized 4-1BB transgenic mice inoculated with human CLDN18.2-expressing tumor cells, givastomig/ABL111 induced a localized immune activation in tumor as evident by the increased ratio of CD8+/regulatory T cell, leading to the superior antitumor activity and long-lasting memory response against tumor rechallenge. Givastomig/ABL111 was well tolerated, with no systemic immune response and hepatotoxicity in monkeys. CONCLUSIONS: Givastomig/ABL111 is a novel CLDN18.2×4-1BB bispecific antibody which has the potential to treat patients with gastric cancer with a wide range of CLDN18.2 expression level through the restricted activation of 4-1BB+ T cells in tumor microenvironment to avoid the risk of liver toxicity and systemic immune response.

The Difference between Rhizosphere and Endophytic Bacteria on the Safe Cultivation of Lettuce in Cr-Contaminated Farmland.

Chromium (Cr) is a major pollutant affecting the environment and human health and microbial remediation is considered to be the most promising technology for the restoration of the heavily metal-polluted soil. However, the difference between rhizosphere and endophytic bacteria on the potential of crop safety production in Cr-contaminated farmland is not clearly elucidated. Therefore, eight Cr-tolerant endophytic strains of three species: Serratia (SR-1~2), Lysinebacillus (LB-1~5) and Pseudomonas (PA-1) were isolated from rice and maize. Additionally, one Cr-tolerant strain of Alcaligenes faecalis (AF-1) was isolated from the rhizosphere of maize. A randomized group pot experiment with heavily Cr-contaminated (a total Cr concentration of 1020.18 mg kg-1) paddy clay soil was conducted and the effects of different bacteria on plant growth, absorption and accumulation of Cr in lettuce (Lactuca sativa var. Hort) were compared. The results show that: (i) the addition of SR-2, PA-1 and LB-5 could promote the accumulation of plant fresh weight by 10.3%, 13.5% and 14.2%, respectively; (ii) most of the bacteria could significantly increase the activities of rhizosphere soil catalase and sucrase, among which LB-1 promotes catalase activity by 224.60% and PA-1 increases sucrase activity by 247%; (iii) AF-1, SR-1, LB-1, SR-2, LB-2, LB-3, LB-4 and LB-5 strains could significantly decrease shoot the Cr concentration by 19.2-83.6%. The results reveal that Cr-tolerant bacteria have good potential to reduce shoot Cr concentration at the heavily contaminated soil and endophytic bacteria have the same or even better effects than rhizosphere bacteria; this suggests that bacteria in plants are more ecological friendly than bacteria in soil, thus aiming to safely produce crops in Cr-polluted farmland and alleviate Cr contamination from the food chain.

WVDL: Weighted Voting Deep Learning Model for Predicting RNA-Protein Binding Sites.

RNA-binding proteins are important for the process of cell life activities. High-throughput technique experimental method to discover RNA-protein binding sites is time-consuming and expensive. Deep learning is an effective theory for predicting RNA-protein binding sites. Using weighted voting method to integrate multiple basic classifier models can improve model performance. Thus, in our study, we propose a weighted voting deep learning model (WVDL), which uses weighted voting method to combine convolutional neural network (CNN), long short term memory network (LSTM) and residual network (ResNet). First, the final forecast result of WVDL outperforms the basic classifier models and other ensemble strategies. Second, WVDL can extract more effective features by using weighted voting to find the best weighted combination. And, the CNN model also can draw the predicted motif pictures. Third, WVDL gets a competitive experiment result on public RBP-24 datasets comparing with other state-of-the-art methods. The source code of our proposed WVDL can be found in https://github.com/biomg/WVDL.

CRMSNet: A deep learning model that uses convolution and residual multi-head self-attention block to predict RBPs for RNA sequence.

RNA-binding proteins (RBPs) play significant roles in many biological life activities, many algorithms and tools are proposed to predict RBPs for researching biological mechanisms of RNA-protein binding sites. Deep learning algorithms based on traditional machine learning get better result for predicting RBPs. Recently, deep learning method fused with attention mechanism has attracted huge attention in many fields and gets competitive result. Thus, attention mechanism module may also improve model performance for predicting RNA-protein binding sites. In this study, we propose convolutional residual multi-head self-attention network (CRMSNet) that combines convolutional neural network (CNN), ResNet, and multi-head self-attention blocks to find RBPs for RNA sequence. First, CRMSNet incorporates convolutional neural networks, recurrent neural networks, and multi-head self-attention block. Second, CRMSNet can draw binding motif pictures from the convolutional layer parameters. Third, attention mechanism module combines the local and global RNA sequence information for capturing long sequence feature. CRMSNet gets competitive AUC (area under the receiver operating characteristic [ROC] curve) result in a large-scale dataset RBP-24. And CRMSNet experiment result is also compared with other state-of-the-art methods. The source code of our proposed CRMSNet method can be found in https://github.com/biomg/CRMSNet.

Effects of intercropping on safe agricultural production and phytoremediation of heavy metal-contaminated soils.

Intercropping with hyperaccumulators is believed to be an important and efficient way to achieve simultaneous safe agricultural production and phytoremediation of polluted soils. However, some studies have suggested that this technique might facilitate the uptake of heavy metals by crops. To investigate the effects of intercropping on the heavy metal contents of plants and soil, data from 135 global studies were collected and analyzed by meta-analysis. The results showed that intercropping could significantly reduce the contents of heavy metals in the main plants and soils. Plant species was the main factor that affected plant and soil metal contents in the intercropping system, and the heavy metal content could be significantly reduced when members of the Poaceae and Crassulaceae were used as main plants or when legumes were used as intercropped plants. Among all the intercropped plants, the best one for removing heavy metals from the soil was a Crassulaceae hyperaccumulator. These results not only highlight the main factors affecting intercropping systems but also provide reliable reference information for the practice of safe agricultural production coupled with phytoremediation of heavy metal-contaminated farmland.

MCNN: Multiple Convolutional Neural Networks for RNA-Protein Binding Sites Prediction.

Computational prediction of the RBP bound sites using features learned from existing annotation knowledge is an effective method because high-throughput experiments are complex, expensive and time-consuming. Many methods have been proposed to predict RNA-protein binding sites. However, the partial information of RNA sequence is not fully used. In this study, we propose multiple convolutional neural networks (MCNN) method, which predicts RNA-protein binding sites by integrating multiple convolutional neural networks constructed by RNA sequence information extracted from windows with different lengths. First, MCNN trains multiple CNNs base on RNA sequences extracted by different window lengths. Second, MCNN can extract more binding patterns of RBPs by combining these trained multiple CNNs previously. Third, MCNN only uses RNA base sequence information for RNA-protein binding sites prediction, which extracts sequence binding features and predicts the result with same architecture. This avoids the information loss of feature extraction step. Our proposed MCNN demonstrates a competitive performance comparing with other methods on a large-scale dataset derived from CLIP-seq, which is an effective method for RNA-protein binding sites prediction. The source code of our proposed MCNN method can be found in https://github.com/biomg/MCNN.