RESUMEN
Networks offer a powerful approach to modeling complex systems by representing the underlying set of pairwise interactions. Link prediction is the task that predicts links of a network that are not directly visible, with profound applications in biological, social, and other complex systems. Despite intensive utilization of the topological feature in this task, it is unclear to what extent a feature can be leveraged to infer missing links. Here, we aim to unveil the capability of a topological feature in link prediction by identifying its prediction performance upper bound. We introduce a theoretical framework that is compatible with different indexes to gauge the feature, different prediction approaches to utilize the feature, and different metrics to quantify the prediction performance. The maximum capability of a topological feature follows a simple yet theoretically validated expression, which only depends on the extent to which the feature is held in missing and nonexistent links. Because a family of indexes based on the same feature shares the same upper bound, the potential of all others can be estimated from one single index. Furthermore, a feature's capability is lifted in the supervised prediction, which can be mathematically quantified, allowing us to estimate the benefit of applying machine learning algorithms. The universality of the pattern uncovered is empirically verified by 550 structurally diverse networks. The findings have applications in feature and method selection, and shed light on network characteristics that make a topological feature effective in link prediction.
RESUMEN
Objective.Accurate and automated segmentation of lung tumors from computed tomography (CT) images is critical yet challenging. Lung tumors are of various sizes and locations and have indistinct boundaries adjacent to other normal tissues.Approach.We propose a new segmentation model that can integrate the topological structure and global features of image region nodes to address the challenges. Firstly, we construct a weighted graph with image region nodes. The graph topology reflects the complex spatial relationships among these nodes, and each node has its specific attributes. Secondly, we propose a node-wise topological feature learning module based on a new graph convolutional autoencoder (GCA). Meanwhile, a node information supplementation (GNIS) module is established by integrating specific features of each node extracted by a convolutional neural network (CNN) into each encoding layer of GCA. Afterwards, we construct a global feature extraction model based on multi-layer perceptron (MLP) to encode the features learnt from all the image region nodes which are crucial complementary information for tumor segmentation.Main results.Ablation study results over the public lung tumor segmentation dataset demonstrate the contributions of our major technical innovations. Compared with other segmentation methods, our new model improves the segmentation performance and has generalization ability on different 3D image segmentation backbones. Our model achieved Dice of 0.7827, IoU of 0.6981, and HD of 32.1743 mm on the public dataset 2018 Medical Segmentation Decathlon challenge, and Dice of 0.7004, IoU of 0.5704 and HD of 64.4661 mm on lung tumor dataset from Shandong Cancer Hospital.Significance. The novel model improves automated lung tumor segmentation performance especially the challenging and complex cases using topological structure and global features of image region nodes. It is of great potential to apply the model to other CT segmentation tasks.
Asunto(s)
Carcinoma de Pulmón de Células no Pequeñas , Neoplasias Pulmonares , Humanos , Carcinoma de Pulmón de Células no Pequeñas/diagnóstico por imagen , Neoplasias Pulmonares/diagnóstico por imagen , Tomografía Computarizada por Rayos X/métodos , Redes Neurales de la Computación , Imagenología Tridimensional , Procesamiento de Imagen Asistido por Computador/métodosRESUMEN
The identification of long noncoding RNA (lncRNA)-disease associations is of great value for disease diagnosis and treatment, and it is now commonly used to predict potential lncRNA-disease associations with computational methods. However, the existing methods do not sufficiently extract key features during data processing, and the learning model parts are either less powerful or overly complex. Therefore, there is still potential to achieve better predictive performance by improving these two aspects. In this work, we propose a novel lncRNA-disease association prediction method LDAformer based on topological feature extraction and Transformer encoder. We construct the heterogeneous network by integrating the associations between lncRNAs, diseases and micro RNAs (miRNAs). Intra-class similarities and inter-class associations are presented as the lncRNA-disease-miRNA weighted adjacency matrix to unify semantics. Next, we design a topological feature extraction process to further obtain multi-hop topological pathway features latent in the adjacency matrix. Finally, to capture the interdependencies between heterogeneous pathways, a Transformer encoder based on the global self-attention mechanism is employed to predict lncRNA-disease associations. The efficient feature extraction and the intuitive and powerful learning model lead to ideal performance. The results of computational experiments on two datasets show that our method outperforms the state-of-the-art baseline methods. Additionally, case studies further indicate its capability to discover new associations accurately.
Asunto(s)
MicroARNs , Neoplasias , ARN Largo no Codificante , Humanos , ARN Largo no Codificante/genética , ARN Largo no Codificante/metabolismo , Biología Computacional/métodos , Neoplasias/genética , MicroARNs/genéticaRESUMEN
Introduction: Sleep deprivation (SD) has a detrimental effect on cognitive functions. Numerous studies have indicated the mechanisms underlying cognitive impairments after SD in brain networks. However, the findings based on the functional connectivity (FC) and topological architecture of brain networks are inconsistent. Methods: In this study, we recruited 30 healthy participants with regular sleep (aged 25.20 ± 2.20 years). All participants performed the repeatable battery for the assessment of neuropsychological status and resting-state fMRI scans twice, during the rested wakefulness (RW) state and after 24 h of total SD. Using the Dosenbach atlas, both large-scale FC and topological features of brain networks (ie nodal, global and local efficiency) were calculated for the RW and SD states. Furthermore, the correlation analysis was conducted to explore the relationship between the changes in FC and topological features of brain networks and cognitive performances. Results: Compared to the RW state, the large-scale brain network results showed decreased between-network FC in somatomotor network (SMN)-default mode network (DMN), SMN-frontoparietal network (FPN), and SMN-ventral attention network (VAN), and increased between-network FC in the dorsal attention network (DAN)-VAN, DAN-SMN after SD. The clustering coefficient, characteristic path length and local efficiency decreased after SD. Moreover, the decreased attention score positively correlated with the decreased topological measures and negatively correlated with the FC of DAN-SMN. Conclusion: Our results suggested that the increased FC of DAN-SMN and decreased topological features of brain networks may act as neural indicators for the decrease in attention after SD. Clinical Trial Registration: The study was registered at the Chinese Clinical Trial Registry, registration ID: ChiCTR2000039858, China.
RESUMEN
Salient object detection can be applied as a critical preprocessing step in many computer vision tasks. Recent studies of salient object detection mainly employed convolutional neural networks (CNNs) for mining high-level semantic properties. However, the existing methods can still be improved to find precise semantic information in different scenarios. In particular, in the two main methods employed for salient object detection, the patchwise detection models might ignore spatial structures among regions and the fully convolution-based models mainly consider semantic features in a global manner. In this paper, we proposed a salient object detection framework by embedding topological features into a deep neural network for extracting semantics. We segment the input image and compute weight for each region with low-level features. The weighted segmentation result is called a topological map and it provides an additional channel for the CNN to emphasize the structural integrity and locality during the extraction of semantic features. We also utilize the topological map for saliency refinement based on a conditional random field at the end of our model. Experimental results on six benchmark datasets demonstrated that our proposed framework achieves competitive performance compared to other state-of-the-art methods.
Asunto(s)
Redes Neurales de la Computación , Estimulación Luminosa/métodos , Reconocimiento en Psicología/fisiología , Semántica , HumanosRESUMEN
BACKGROUND: Functional remodeling may vary with tumor aggressiveness of glioma. Investigation of the functional remodeling is expected to provide scientific relevance of tumor characterization and disease management of glioma. In this study, we aimed to investigate the functional remodeling of the contralesional hemisphere and its utility in predicting the malignant grade of glioma at the individual level with multivariate logistic regression (MLR) analysis. SUBJECTS AND METHODS: One hundred and twenty-six right-handed subjects with histologically confirmed cerebral glioma were included with 80 tumors located in the left hemisphere (LH) and 46 tumors located in the right hemisphere (RH). Resting-state functional networks of the contralesional hemisphere were constructed using the human brainnetome atlas based on resting-state fMRI data. Functional connectivity and topological features of functional networks were quantified. The performance of functional features in predicting the glioma grade was evaluated using area under (AUC) the receiver operating characteristic curve (ROC). The dataset was divided into training and validation datasets. Features with high AUC values in malignancy classification in the training dataset were determined as predictive features. An MLR model was constructed based on predictive features and its classification performance was evaluated on the training and validation datasets with 10-fold cross validation. RESULTS: Predictive functional features showed apparent hemispheric specifications. MLR classification models constructed with age and predictive functional connectivity features (AUC of 0.853 ± 0.079 and 1.000 ± 0.000 for LH and RH group, respectively) and topological features (AUC of 0.788 ± 0.150 and 0.897 ± 0.165 for LH and RH group, respectively) achieved efficient performance in predicting the malignant grade of gliomas. CONCLUSION: Functional remodeling of the contralesional hemisphere was hemisphere-specific and highly predictive of the malignant grade of glioma. Network approach provides a novel pathway that may innovate glioma characterization and management at the individual level.
RESUMEN
The interactions of microRNAs (miRNAs), transcription factors (TFs) and their common target long noncoding RNAs (lncRNAs) can lead to the production of TFmiRNAlncRNA (TML) network motifs. These motifs are functional regulators that perform a wide range of biological processes, such as carcinogenesis. However, TML network motifs have not been systematically identified, and their roles in lung adenocarcinoma (LUAD) and lung squamous carcinoma (LUSC) are largely unknown. In the present study, a computational integration approach was performed using multiple sources in order to construct a global TML network for LUAD and LUSC. The analysis revealed several dysregulated TML network motifs, which were common between the two lung cancer subtypes or specific to a single cancer subtype. In addition, functional analysis further indicated that the TML network motifs may potentially serve as putative biomarkers in LUAD and LUSC. The associations between drug treatments and dysregulated TML network motifs were also examined. Collectively, the present study elucidated the roles of TML network motifs in LUAD and LUSC, which may be beneficial for understanding the pathogenesis of lung cancer and its potential treatment.
Asunto(s)
Adenocarcinoma del Pulmón/genética , Antineoplásicos/uso terapéutico , Carcinoma de Pulmón de Células no Pequeñas/genética , Regulación Neoplásica de la Expresión Génica , Neoplasias Pulmonares/genética , MicroARNs/genética , ARN Largo no Codificante/genética , Factores de Transcripción/genética , Adenocarcinoma del Pulmón/tratamiento farmacológico , Adenocarcinoma del Pulmón/metabolismo , Adenocarcinoma del Pulmón/patología , Antineoplásicos/clasificación , Carcinoma de Pulmón de Células no Pequeñas/tratamiento farmacológico , Carcinoma de Pulmón de Células no Pequeñas/metabolismo , Carcinoma de Pulmón de Células no Pequeñas/patología , Línea Celular Tumoral , Biología Computacional/métodos , Bases de Datos Genéticas , Perfilación de la Expresión Génica , Ontología de Genes , Redes Reguladoras de Genes , Humanos , Neoplasias Pulmonares/tratamiento farmacológico , Neoplasias Pulmonares/metabolismo , Neoplasias Pulmonares/patología , MicroARNs/metabolismo , Anotación de Secuencia Molecular , Farmacogenética/métodos , Mapeo de Interacción de Proteínas , ARN Largo no Codificante/metabolismo , Transducción de Señal , Factores de Transcripción/metabolismoRESUMEN
The lath martensite structure contains hierarchical substructures, such as blocks, packets and prior austenite grains. Generally, high-angle grain boundaries in the lath martensite structure, i.e. block boundaries, are correlated to mechanical properties. On the other hand, low-angle grain boundaries play an important role in morphological development. However, it is difficult to understand their nature because of the difficulty associated with the characterization of the complex morphologies by two-dimensional techniques. This study aims to identify the morphologies of low-angle boundaries in ultra-low carbon lath martensite. A serial-sectioning method and electron backscatter diffraction analysis are utilized to reconstruct three-dimensional objects and analyse their grain boundaries. A packet comprizes two low-angle grain boundaries - sub-block and fine packet boundaries. Sub-blocks exhibit porous morphology, with two large sub-blocks predominantly occupying a block. Several fine packets with different habit planes from the surrounding regions are observed. Fine packets are present in blocks, which frequently share a close-packed direction with the neighbouring fine packets. In addition, fine packets are in contact with the sub-block boundaries.
RESUMEN
Long non-coding RNAs (lncRNAs), transcription factors and microRNAs can form lncRNA-mediated feed-forward loops (L-FFLs), which are functional network motifs that regulate a wide range of biological processes, such as development and carcinogenesis. However, L-FFL network motifs have not been systematically identified, and their roles in human cancers are largely unknown. In this study, we computationally integrated data from multiple sources to construct a global L-FFL network for six types of human cancer and characterized the topological features of the network. Our approach revealed several dysregulated L-FFL motifs common across different cancers or specific to particular cancers. We also found that L-FFL motifs can take part in other types of regulatory networks, such as mRNA-mediated FFLs and ceRNA networks, and form the more complex networks in human cancers. In addition, survival analyses further indicated that L-FFL motifs could potentially serve as prognostic biomarkers. Collectively, this study elucidated the roles of L-FFL motifs in human cancers, which could be beneficial for understanding cancer pathogenesis and treatment.
Asunto(s)
Biomarcadores de Tumor/genética , MicroARNs/genética , Neoplasias/genética , ARN Largo no Codificante/genética , Factores de Transcripción/genética , Transcriptoma , Perfilación de la Expresión Génica , Redes Reguladoras de Genes/genética , Humanos , Estimación de Kaplan-Meier , Neoplasias/mortalidad , PronósticoRESUMEN
BACKGROUND: In the context of drug discovery, drug target interactions (DTIs) can be predicted based on observed topological features of a semantic network across the chemical and biological space. In a semantic network, the types of the nodes and links are different. In order to take into account the heterogeneity of the semantic network, meta-path-based topological patterns were investigated for link prediction. RESULTS: Supervised machine learning models were constructed based on meta-path topological features of an enriched semantic network, which was derived from Chem2Bio2RDF, and was expanded by adding compound and protein similarity neighboring links obtained from the PubChem databases. The additional semantic links significantly improved the predictive performance of the supervised learning models. The binary classification model built upon the enriched feature space using the Random Forest algorithm significantly outperformed an existing semantic link prediction algorithm, Semantic Link Association Prediction (SLAP), to predict unknown links between compounds and protein targets in an evolving network. In addition to link prediction, Random Forest also has an intrinsic feature ranking algorithm, which can be used to select the important topological features that contribute to link prediction. CONCLUSIONS: The proposed framework has been demonstrated as a powerful alternative to SLAP in order to predict DTIs using the semantic network that integrates chemical, pharmacological, genomic, biological, functional, and biomedical information into a unified framework. It offers the flexibility to enrich the feature space by using different normalization processes on the topological features, and it can perform model construction and feature selection at the same time.
Asunto(s)
Diseño de Fármacos , Aprendizaje Automático , Preparaciones Farmacéuticas/química , Minería de Datos , Bases de Datos Factuales , Modelos Químicos , Fenómenos Farmacológicos , Proteínas/químicaRESUMEN
We investigate the mechanism of tactile transduction during active exploration of finely textured surfaces using a tactile sensor mimicking the human fingertip. We focus in particular on the role of exploratory conditions in shaping the subcutaneous mechanical signals. The sensor has been designed by integrating a linear array of MEMS micro-force sensors in an elastomer layer. We measure the response of the sensors to the passage of elementary topographical features at constant velocity and normal load, such as a small hole on a flat substrate. Each sensor's response is found to strongly depend on its relative location with respect to the substrate/skin contact zone, a result which can be quantitatively understood within the scope of a linear model of tactile transduction. The modification of the response induced by varying other parameters, such as the thickness of the elastic layer and the confining load, are also correctly captured by this model. We further demonstrate that the knowledge of these characteristic responses allows one to dynamically evaluate the position of a small hole within the contact zone, based on the micro-force sensors signals, with a spatial resolution an order of magnitude better than the intrinsic resolution of individual sensors. Consequences of these observations on robotic tactile sensing are briefly discussed.