Search

Abstract Flood forecasting using traditional physical hydrology models requires consideration of multiple complex physical processes including the spatio-temporal distribution of rainfall, the spatial heterogeneity of watershed sub-surface characteristics, and runoff generation and routing behaviours. Data-driven models offer novel solutions to these challenges, though they are hindered by difficulties in hyperparameter selection and a decline in prediction stability as the lead time extends. This study introduces a hybrid model, the RS-LSTM-Transformer, which combines Random Search (RS), Long Short-Term Memory networks (LSTM), and the Transformer architecture. Applied to the typical Jingle watershed in the middle reaches of the Yellow River, this model utilises rainfall and runoff data from basin sites to simulate flood processes, and its outcomes are compared against those from RS-LSTM, RS-Transformer, RS-BP, and RS-MLP models. It was evaluated against RS-LSTM, RS-Transformer, RS-BP, and RS-MLP models using the Nash–Sutcliffe Efficiency Coefficient (NSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Bias percentage as metrics. At a 1-h lead time during calibration and validation, the RS-LSTM-Transformer model achieved NSE, RMSE, MAE, and Bias values of 0.970, 14.001m3/s, 5.304m3/s, 0.501% and 0.953, 14.124m3/s, 6.365m3/s, 0.523%, respectively. These results demonstrate the model's superior simulation capabilities and robustness, providing more accurate peak flow forecasts as the lead time increases. The study highlights the RS-LSTM-Transformer model's potential in flood forecasting and the advantages of integrating various data-driven approaches for innovative modelling.

Abstract The glymphatic system plays an important role in the transportation of cerebrospinal fluid (CSF) and the clearance of metabolite waste in brain. However, current imaging modalities for studying the glymphatic system are limited. Herein, we apply NIR-II nanoprobes with non-invasive and high-contrast advantages to comprehensively explore the function of glymphatic system in mice under anesthesia and cerebral ischemia–reperfusion injury conditions. Our results show that the supplement drug dexmedetomidine (Dex) enhances CSF influx in the brain, decreases its outflow to mandibular lymph nodes, and leads to significant differences in CSF accumulation pattern in the spine compared to isoflurane (ISO) alone, while both ISO and Dex do not affect the clearance of tracer-filled CSF into blood circulation. Notably, we confirm the compromised glymphatic function after cerebral ischemia–reperfusion injury, leading to impaired glymphatic influx and reduced glymphatic efflux. This technique has great potential to elucidate the underlying mechanisms between the glymphatic system and central nervous system diseases. Graphical Abstract

Study region: Flood formation involves complex nonlinear processes and numerous variables, with data-driven models becoming a key non-engineering approach to flood prevention and mitigation. Yet, single machine learning models are insufficient to fully capture the complex dynamics of the flood process. Study focus: We propose an interpretable flood forecasting hybrid model based on Transformer, LSTM, and Adaptive Random Search Algorithm (AGRS), termed as AGRS-LSTM-Transformer. Investigating the predictive performance of the hybrid model, this study compares it against AGRS-LSTM, AGRS-Transformer, AGRS-BP, and AGRS-MLP models, utilizing flood data from 1971 to 2013 years in the Jingle watershed. New hydrological insights for the region: The AGRS-LSTM-Transformer model demonstrates superior performance over benchmark models, achieving accurate runoff forecasts with a lead time of 1–6 hours. It achieves a Nash-Sutcliffe Efficiency (NSE) greater than 0.905, and root mean squared error (RMSE), mean absolute error (MAE), Bias, and relative error (RE) values for the runoff process below 34.891 m³/s, 25.125 m³/s, 9.537 %, and 8.025 %, respectively. The coupling between the LSTM layers and the Transformer input components plays a crucial role in the architecture of the AGRS-LSTM-Transformer model. Rainfall from stations situated near the main river channel and downstream flow sections exerted a positive influence. Runoff from the preceding moment significantly impacts the predicted flow, with the contribution of runoff inputs exceeding that of rainfall. Inputs nearer to the forecast moment do not invariably improve forecasting accuracy, and historical rainfall and runoff volumes with extended lag times may detrimentally impact the model prediction. The study highlights the potential of hybrid data-driven models in enhancing the accuracy of flood forecasting, offering insights for reducing uncertainty in flood prediction and interpreting machine learning flood forecasting models. This provides a scientific basis for flood early warning systems and water resource management.

Abstract Fluorescent dyes that emit in the second near‐infrared (NIR‐II, 1000–3000 nm) region have provided significant advances toward real‐time and high‐resolution imaging of vessel and lymphatic system. However, in vivo NIR‐II tracking of the fate of labeled cells still remains challenging. Here, we develop a shielding unit–donor–acceptor–donor–shielding unit (S‐D‐A‐D‐S) NIR‐II fluorophore (FE‐4ZW) with zwitterionic terminal groups for high‐efficiency cell labeling without using cell‐penetrating peptides, which provides for enhanced non‐invasive in vivo determination of the location of cell migration. The tethering terminal sulfoammonium inner salts are featured with its high affinity for cell membranes, thereby enabling the stable labeling even for fixed cells. The fate of transplanted stem cell and the tumor cell migration along lymphatic system in brain or periphery tissues are clearly monitored by the cell‐internalized FE‐4ZW. We also confirmed that a clinically used surfactant, D‐α‐tocopheryl polyethylene glycol‐1000 succinate, can reduce the liver and spleen uptake of FE‐4ZW. The fluorophore design strategy and cell‐labeling technology reported here open a new realm in the visualization of cell migration and insight into the relocation process, thereby ultimately providing an opportunity to investigate in greater detail of the underlying mechanisms of stem cell therapy and tumor metastasis.

Aiming at the difficulty of sample collection, high cost, and small sample problems caused by many types of defects and difficulty in covering all small samples in the process of industrial steel surface defect detection, a steel surface defect image generation method improved by cycle-consistent generation adversarial network (CycleGAN) was proposed. Firstly, channel attention map (CAM) and spatial attention map (SAM) were embedded in the CycleGAN model to enhance the feature extraction ability of the model. Secondly, the Weight Demodulation (WD) mechanism was introduced to fix feature artifacts and white spots, further improving the quality of the generated images. Thirdly, shape consistency loss was introduced to supervise the generator training process to solve the problem of inherent ambiguity in the process of image geometric transformation. Finally,the model before and after the improvement were experimented on the NEU-DET dataset. The results show that the improved model has more diversity and accuracy in the effect of defect image generation. PSNR and SSIM increase by 13.0% and 7.8% respectively, and FID values are reduced by 33.1%. This method can stably generate high-quality images of various steel surface defects, which can achieve the purpose of increasing training data, and also has reference value for the amplification of other defect datasets.

Research and development in the field of micro/nano-robots have made significant progress in the past, especially in the field of clinical medicine, where further research may lead to many revolutionary achievements. Through the research and experiment of microrobots, a controllable drug delivery system will be realized, which will solve many problems in drug treatment. In this work, we design and study the ability of magnetic-driven hydrogel microrobots to carry Lycorine hydrochloride (LH) to inhibit colorectal cancer (CRC) cells. We have successfully designed a magnetic field driven, biocompatible drug carrying hydrogel microsphere robot with Fe3O4 particles inside, which can achieve magnetic field response, and confirmed that it can transport drug through fluorescence microscope. We have successfully demonstrated the motion mode of hydrogel microrobots driven by a rotating external magnetic field. This driving method allows the microrobots to move in a precise and controllable manner, providing tremendous potential for their use in various applications. Finally, we selected drug LH and loaded it into the hydrogel microrobot for a series of experiments. LH significantly inhibited CRC cells proliferation in a dose- and time-dependent manner. LH inhibited the proliferation, mobility of CRC cells and induced apoptosis. This delivery system can significantly improve the therapeutic effect of drugs on tumors.

Study region: Typical basin in humid areas in the Huaihe River Study focus: Accurate flood forecasting is essential for making timely decisions regarding flood control and disaster reduction. The theory of watershed runoff generation and convergence serves as a crucial foundation for flood forecasting, while the calculation of runoff is necessary to simulate flood discharge. Identifying watershed runoff generation mechanisms has been a challenging task, particularly under complex underlying surface conditions. To improve the accuracy of flood simulation, this study examines the underlying surface information in the watershed, such as particle composition and content, soil bulk density, geological slope, land use, and other spatial attributes, aiming to analyze the mechanisms of runoff generation. In the study of sub-watersheds, various combinations of runoff generation mechanisms are identified to determine the patterns of runoff. Subsequently, a semi-distributed hydrological model is developed, which incorporates both saturation-excess and infiltration-excess runoff, utilizing the information obtained from the underlying surface. The model is validated using rainfall-runoff data from 14 events at the Xiagushan watershed. New hydrological insights for the region: The analysis of the fundamental physical conditions of the underlying surface of the watershed revealed that 69.70% of the area is prone to saturation-excess runoff, with an additional 30.30% of the area being susceptible to infiltration-excess runoff. The model considers the spatial distribution of runoff patterns by incorporating complex underlying surface information and demonstrates high accuracy in simulating flood events (NSE= 0.87, Epeak = 12.08%, Wpeak = 13.16%, Tpeak = 0.14 h, R2 = 0.90). The model is straightforward, practical, and exhibits promising potential in terms of timeliness and applicability, thus lending itself well to further application in other watersheds, contributing to the scientific foundation of flood warning and forecasting efforts.

Abstract Background Lymphomatoid granulomatosis (LyG) is a rare extralymphatic lymphoproliferative disease characterized by lymphocytic invasion into vascular walls and damage to blood vessels. The lungs are affected in 90% of LyG cases, followed by the skin, central nervous system (CNS), kidneys and liver. Case presentation Here we report a case of a young woman with LyG, with CNS involvement as the initial clinical manifestation. Computer tomography (CT) scans showed multiple nodular, patchy and flocculent high-density shadows in both lungs without mediastinal lymph node enlargement. Magnetic resonance imaging (MRI) scans showed multiple abnormal signal intensities in the right cerebellar hemisphere, frontal, parietal and temporal lobes, and dorsal brainstem, which became patchy and annular after enhancement. The post-operative pathological analysis of lesion samples confirmed the diagnosis of grade II LyG. Conclusions LyG should be concerned in young adults showing multiple radiological brain and lung lesions. Resection and postoperative medication of steroid hormones and IFN-α may be effective in the treatment of LyG.

In many applications, data streams are indispensable to describe the relationships between nodes in networks, such as social networks, computer networks, and hyperlink networks. Fundamentally, a graph stream is a dynamic representation of a graph, which is usually composed of a sequence of edges, where each edge is represented by two endpoints and a weight. As a result of its large volume and highly dynamic nature, several graph sketches were proposed for the purposes of summarizing large-scale graph streams and enabling fast query processing. By using a compact data structure with hash functions, the graph sketches sequentially store the edges. Nevertheless, the existing graph sketches suffer from low performance on graph query tasks as a result of unpredictable collisions between heavy edges and light edges. To store heavy edges and light edges, this paper introduces a novel learning-based Dichotomy Graph Sketch (DGS) mechanism that uses two separate graph sketches, a heavy sketch and a light sketch. During a graph stream session, DGS obtains heavy edges and light edges, and uses these edges as training samples for a deep neural network (DNN) based binary classifier. The DNN-based classifier is then used to determine whether the upcoming edges are heavy or not. We will store the edges that are classified as heavy edges in the heavy sketch, and those that are classified as light edges in the light sketch. By combining the learnable classifier and Dichotomy Graph Sketches, the proposed mechanism resolves the hashing collision problem in conventional graph sketches and significantly improves graph query accuracy. The DGS algorithm outperforms the state-of-the-art graph sketches in a variety of graph query tasks based on extensive experiments that were conducted on four real-world graph stream datasets.

In order to study the effects of a low protein diet on the production performance and intestinal microbiota composition of Hexi pigs, twenty-seven Hexi pigs with an initial body weight of 60.50 ± 2.50 kg were randomly divided into three groups (control group (CG), group 1 (G1), and group 2 (G2)) and participated in a 60-day finishing trial. The CG was fed a normal protein level diet with a protein level of 16.0%, and G1 and G2 were fed a low protein level diet with protein levels of 14.0% and 12.0%, respectively. The results showed that the low protein level diet had no significant effect on the production performance of Hexi pigs, compared with the CG, the slaughter rate of G1 and G2 increased by 2.49% (p > 0.05) and 6.18% (p > 0.05), the shear force decreased by 2.43% (p > 0.05) and 15.57% (p > 0.05), the cooking loss decreased by 24.02% (p < 0.05) and 21.09% (p > 0.05), and the cooking percentage increased by 13.20% (p > 0.05) and 11.59% (p > 0.05). From 45 min to 24 h and 48 h after slaughter, each group of pH decreased by 1.02, 0.66, and 0.42. For muscle flesh color, the lightness (L) increased by 13.31% (p > 0.05) and 18.01% (p > 0.05) in G1 and G2 and the yellowness (b) increased by 7.72% (p > 0.05) and 13.06% (p > 0.05). A low protein level diet can improve the intestinal flora richness and diversity of growing and finishing pigs. In the jejunum, the ACE index (899.95), Simpson index (0.90), and Shannon (4.75) index were higher in G1 than in the other groups, but the Chao1 index (949.92) was higher in G2 than in the remaining two groups. Proteobacteria, Actinobacteria, Euryarchaeota, and Verrucomicrobia were significantly higher in G1 than in the CG. The relative abundances of Lactobacillus, Terrisporobacter, and Megasphaera in G1 was significantly higher than in the CG (p < 0.05). In the cecum, the ACE index (900.93), Chao1 index (879.10), Simpson index (0.94), and Shannon (5.70) index were higher in G1 than in the remaining groups. The Spirochaetes in G2 were significantly higher than in the other groups, but the Verrucomicrobia was significantly lower than in the other groups. The relative abundances of Lactobacillus were higher in G1 and G2 than in the CG (p > 0.05). The relative abundances of unidentified_Clostridiales and Terrisporobacter in G2 were significantly lower than in the CG (p < 0.05). The relative abundance of Turicibacter in G1 was significantly lower than in the CG (p < 0.05). The relative abundances of other bacterial genera in G1 and G2 were increased by 30.81% (p > 0.05) and 17.98% (p > 0.05).

During power swing, the distance protection is easily affected by the oscillations of voltage and current which may lead to the mal-operation of the protection. Therefore, a power swing detecting unit is needed to cooperate with distance protection. In this paper, a power swing detecting method based on the difference between the structure of transmission line under fault situations and that under power swing situations is proposed. Principal component analysis is deployed to process the voltage and current data to extract the structure feature of the transmission line under power swing and fault situations. To test the effectiveness of the proposed method, a simulation system is built in PSCAD/EMTDC and different disturbance scenarios are done. The test results indicate that the proposed method can reliably distinguish between power swing and fault situations. Besides, a comparison between the proposed method and methods mentioned in other works is been done. The comparison results show that the proposed method has an advantage in operating speed and is less affected by measuring noise.

High voltage direct current (HVDC) technology is the ideal solution for long distance bulk power transmission. However, different from AC systems, differential protection of HVDC lines still faces a lot of challenges. We claim that from the perspective of state estimation, the essence of differential protection is to estimate state variables using measured data, of which the validity indicates whether there is a fault. Therefore, the performance of the differential protection is highly related to the model accuracy. Using a detailed distributed parameter line model, this paper developed a novel differential protection method for HVDC lines based on traveling waves with high reliability. The optimal differential point is also discussed to enhance protection rapidity. Experiments demonstrate that the proposed protection could soon detect internal faults with high fault resistance while maintaining reliability when external faults occur.

A novel 3T2R hybrid mechanism that consists of a 3-RRP4RR parallel mechanism and a two degree of freedom rotating mechanism is proposed according to engineering application scene of labelling for the round steel end surfaces of a bundle. A method of combining two Monte Carlo methods is proposed to solve the workspace of the hybrid mechanism. The kinematics model of the hybrid mechanism is established. According to the geometric relationship of the parallel mechanism， the inverse solution equation is obtained，and the constraints of the workspace of the parallel mechanism are determined. The set of points in the workspace of the parallel mechanism is obtained by using Monte Carlo method. The kinematics positive solution of the series mechanism and the kinematics positive solution of the hybrid mechanism are obtained according to the coordinate transformation matrix. The coordinates of the points in the workspace of the parallel mechanism is substituted into the positive solution of the position of the hybrid mechanism， and by using Monte Carlo method， the set of points in the workspace of the hybrid mechanism is obtained.

The polymer cutoff wall offers the benefits of causing little disturbance to the dam body, being convenient and economical to construct and practical to use, and possessing good resistance to seismic shocks and cracks. It has been widely used to prevent seepage and to reinforce dikes and dams. However, the polymer cutoff wall is a concealed underground structure, and nondestructive testing methods to ensure its integrity are not yet mature. This paper describes a modal analysis of polymer cutoff wall models with different damage scenarios to investigate the feasibility of nondestructive testing of the polymer cutoff wall based on vibration theory. The dynamic characteristics of the first three natural frequencies of the wall and their mode shapes show that horizontal direction damage and centrally located partial damage have a noticeable impact on the dynamic characteristics of the wall, indicating that nondestructive testing based on vibration theory is useful to test for horizontal damage and moderate damage located centrally in the wall.

A class of solutions are introduced by lexicographically minimizing the complaint of coalitions for cooperative games with transferable utility. Among them, the nucleolus is an important representative. From the perspective of measuring the satisfaction of coalitions with respect to a payoff vector, we define a family of optimal satisfaction values in this paper. The proportional division value and the proportional allocation of non-separable contribution value are then obtained by lexicographically maximizing two types of satisfaction criteria, respectively, which are defined by the lower bound and the upper bound of the core from the viewpoint of optimism and pessimism respectively. Correspondingly, we characterize these two proportional values by introducing the equal minimal satisfaction property and the associated consistency property. Furthermore, we analyze the duality of these axioms and propose more approaches to characterize these two values on basis of the dual axioms.

Stem cell transplantation is a viable strategy for regenerative medicine. However, it is inevitable to have cells undergo storage for several hours or days due to processing and transportation. Therefore, it is crucial to have rigidly controlled conditions ensuring the therapeutic benefit of isolated stem cells. In the present study, we investigated the impact of short-term storage on human CD133 + cells. CD133 + cells were isolated from human bone marrow and kept at standardized nonfreezing storage conditions for up to 72 h. Cell viability (apoptosis/necrosis) and expression of CD133 and CXCR4 were analyzed by flow cytometry. Metabolic activity was determined using an MTT assay; colony-forming ability, as well as endothelial-like differentiation, was further evaluated. A qRT-PCR array was employed to investigate the expression of sternness genes. CD133 and CXCR4 expressions were preserved at all time points. After 30 h, cell number and metabolic activity decreased, although no significant changes were detected in cell viability and proliferation as well as endothelial-like differentiation. Cell viability and proliferation decreased significantly only after 72 h of storage. Our results indicate that storage of isolated human CD133 + bone marrow stem cells in liquid allows for high viability and functionality. However, storage time should be limited in order to avoid cell loss.

The emerging edge computing paradigm has given rise to a new promising mobile network architecture, which can address a number of challenges that the operators are facing while trying to support growing end user’s needs by shifting the computation from the base station to the edge cloud computing facilities. With such powerfully computational power, traditional unpractical resource allocation algorithms could be feasible. However, even with near optimal algorithms, the allocation result could still be far from optimal due to the inaccurate modeling of interference among sensor nodes. Such a dilemma calls for a measurement data-driven resource allocation to improve the total capacity. Meanwhile, the measurement process of inter-nodes’ interference could be tedious, time-consuming and have low accuracy, which further compromise the benefits brought by the edge computing paradigm. To this end, we propose a measurement-based estimation solution to obtain the interference efficiently and intelligently by dynamically controlling the measurement and estimation through an accuracy-driven model. Basically, the measurement cost is reduced through the link similarity model and the channel derivation model. Compared to the exhausting measurement method, it can significantly reduce the time cost to the linear order of the network size with guaranteed accuracy through measurement scheduling and the accuracy control process, which could also balance the tradeoff between accuracy and measurement overhead. Extensive experiments based on real data traces are conducted to show the efficiency of the proposed solutions.

Online social networks such as Facebook, Twitter and Gowalla allow people to communicate and interact across borders. In past years online social networks have become increasingly important for studying the behavior of individuals, group formation, and the emergence of online societies. Here we focus on the characterization of the average growth of online social networks and try to understand which are possible processes behind seemingly long-range temporal correlated collective behavior. In agreement with recent findings, but in contrast to Gibrat's law of proportionate growth, we find scaling in the average growth rate and its standard deviation. In contrast, Renren and Twitter deviate, however, in certain important aspects significantly from those found in many social and economic systems. Whereas independent methods suggest no significance for temporally long-range correlated behavior for Renren and Twitter, a scaling analysis of the standard deviation does suggest long-range temporal correlated growth in Gowalla. However, we demonstrate that seemingly long-range temporal correlations in the growth of online social networks, such as in Gowalla, can be explained by a decomposition into temporally and spatially independent growth processes with a large variety of entry rates. Our analysis thus suggests that temporally or spatially correlated behavior does not play a major role in the growth of online social networks.

This study assessed the concept of whether delivery of magnetic nanobeads (MNBs)/adenoviral vectors (Ad)-encoded hVEGF gene (Ad(hVEGF)) could regenerate ischaemically damaged hearts in a rat acute myocardial infarction model under the control of an external magnetic field. Adenoviral vectors were conjugated to MNBs with the Sulfo-NHS-LC-Biotin linker. In vitro transduction efficacy of MNBs/Ad-encoded luciferase gene (Ad(luc)) was compared with Ad(luc) alone in human umbilical vein endothelial cells (HUVECs) under magnetic field stimulation. In vivo, in a rat acute myocardial infarction (AMI) model, MNBs/Ad(hVEGF) complexes were injected intravenously and an epicardial magnet was employed to attract the circulating MNBs/Ad(hVEGF) complexes. In vitro, compared with Ad(luc) alone, MNBs/Ad(luc) complexes had a 50-fold higher transduction efficiency under the magnetic field. In vivo, epicardial magnet effectively attracted MNBs/Ad(hVEGF) complexes and resulted in strong therapeutic gene expression in the ischemic zone of the infarcted heart. When compared to other MI-treated groups, the MI-M(+)/Ad(hVEGF) group significantly improved left ventricular function (p

Nanog is a pivotal transcription factor in embryonic stem (ES) cells and is essential for maintaining the pluripotency and self-renewal of ES cells. SUMOylation has been proved to regulate several stem cell markers' function, such as Oct4 and Sox2. Nanog is strictly regulated by Oct4/Sox2 heterodimer. However, the direct effects of SUMOylation on Nanog expression remain unclear. In this study, we reported that SUMOylation repressed Nanog expression. Depletion of Sumo1 or its conjugating enzyme Ubc9 increased the expression of Nanog, while high SUMOylation reduced its expression. Interestingly, we found that SUMOylation of Oct4 and Sox2 regulated Nanog in an opposing manner. SUMOylation of Oct4 enhanced Nanog expression, while SUMOylated Sox2 inhibited its expression. Moreover, SUMOylation of Oct4 by Pias2 or Sox2 by Pias3 impaired the interaction between Oct4 and Sox2. Taken together, these results indicate that SUMOylation has a negative effect on Nanog expression and provides new insights into the mechanism of SUMO modification involved in ES cells regulation.

The possible different therapeutic efficacy of human mesenchymal stem cells (hMSC) derived from umbilical cord blood (CB), adipose tissue (AT) or bone marrow (BM) for the treatment of myocardial infarction (MI) remains unexplored. This study was to assess the regenerative potential of hMSC from different origins and to evaluate the role of CD105 in cardiac regeneration. Male SCID mice underwent LAD-ligation and received the respective cell type (400.000/per animal) intramyocardially. Six weeks post infarction, cardiac catheterization showed significant preservation of left ventricular functions in BM and CD105(+)-CB treated groups compared to CB and nontreated MI group (MI-C). Cell survival analyzed by quantitative real time PCR for human GAPDH and capillary density measured by immunostaining showed consistent results. Furthermore, cardiac remodeling can be significantly attenuated by BM-hMSC compared to MI-C. Under hypoxic conditions in vitro, remarkably increased extracellular acidification and apoptosis has been detected from CB-hMSC compared to BM and CD105 purified CB-derived hMSC. Our findings suggests that hMSC originating from different sources showed a different healing performance in cardiac regeneration and CD105(+) hMSC exhibited a favorable survival pattern in infarcted hearts, which translates into a more robust preservation of cardiac function.

Bone marrow-derived mesenchymal stem cells (MSCs) are multipotent cells characterized by their self-renewal and differentiation potential. Accumulating clinical and preclinical evidence indicate MSCs are a promising cell source for regenerative medical therapies. However, undesirable immortalization, spontaneous transformation, and tumorigenic potential from long-term cultured MSCs have been reported in human and mouse. We report rat MSCs isolated from young donors could undergo transformation in early passage culture. We aimed to characterize the transformed population and determine their therapeutic effects after intracardiac transplantation in the infarcted myocardium. MSCs were isolated from bone marrow of Lewis rats according to standard protocols and cultured under standard conditions. Phenotype of growing cells was assessed by flow cytometry. Following acute myocardial infarction in rats, cells were delivered by intracardiac injection. Cardiac functions were assessed by pressure–volume loops. Infarction size and pathologic effects were evaluated after 6 weeks. The abnormal colonies were detected in culture as early at passage 3. They were noted to appear as distinctly different morphology from typical MSCs, which changed from a normal elongated spindle shape to a compact abnormal morphology. They exhibited rapid cell proliferation. Some subclones lost contact inhibition of cell division and formed multilayer aggregates. Chromosomal instability was detected. They were devoid of surface markers CD29, CD44, CD90, and CD117. Furthermore, there was no significant improvement on infarction size and cardiac function 6 weeks after cell transplantation. Our study highlights the need for establishment of biosafety criteria in regulating culture-expanded MSCs to achieve the full clinical therapeutic benefits.