Search

Constructing accurate gene regulatory network s (GRNs), which reflect the dynamic governing process between genes, is critical to understanding the diverse cellular process and unveiling the complexities in biological systems. With the development of computer sciences, computational-based approaches have been applied to the GRNs inference task. However, current methodologies face challenges in effectively utilizing existing topological information and prior knowledge of gene regulatory relationships, hindering the comprehensive understanding and accurate reconstruction of GRNs. In response, we propose a novel graph neural network (GNN)-based Multi-Task Learning framework for GRN reconstruction, namely MTLGRN. Specifically, we first encode the gene promoter sequences and the gene biological features and concatenate the corresponding feature representations. Then, we construct a multi-task learning framework including GRN reconstruction , Gene knockout predict , and Gene expression matrix reconstruction. With joint training, MTLGRN can optimize the gene latent representations by integrating gene knockout information, promoter characteristics, and other biological attributes. Extensive experimental results demonstrate superior performance compared with state-of-the-art baselines on the GRN reconstruction task, efficiently leveraging biological knowledge and comprehensively understanding the gene regulatory relationships. MTLGRN also pioneered attempts to simulate gene knockouts on bulk data by incorporating gene knockout information. [ABSTRACT FROM AUTHOR]

As the quality of life rises, the demand for flowers has increased significantly, leading to higher expectations for flower sorting system efficiency and speed. This paper presents a real-time, high-precision end-to-end method, which can complete three key tasks in the sorting system: flower localization, flower classification, and flower grading. In order to improve the challenging maturity detection, red–green–blue depth (RGBD) images were captured. The multi-task and multi-dimension-You Only Look Once (MTMD-YOLO) network was proposed to complete these three tasks in an end-to-end manner. The feature fusion was simplified to increase training speed, and the detection head and non-maximum suppression (NMS) were optimized for the dataset. This optimization allowed the loss function for the grading task to be added to train each task separately. The results showed that the use of RGBD and multi-task improved by 3.63% and 1.87% of mean average precision (mAP) on flower grading task, respectively. The final mAP of the flower classification and grading task reached 98.19% and 97.81%, respectively. The method also achieved real-time speed on embedded Jetson Orin NX, with 37 frames per second (FPS). This method provided essential technical support to determine the automatic flower picking times, in combination with a picking robot. [ABSTRACT FROM AUTHOR]

Accurate segmentation of the pulmonary airway tree is crucial for diagnosing lung diseases. To tackle the issues of low segmentation accuracy and frequent leaks in existing methods, this paper proposes a precise segmentation method using quasi-spherical region-constrained wavefront propagation with tracheal wall gap sealing. Based on the characteristic that the surface formed by seed points approximates the airway cross-section, the width of the unsegmented airway is calculated, determining the initial quasi-spherical constraint region. Using the wavefront propagation method, seed points are continuously propagated and segmented along the tracheal wall within the quasi-spherical constraint region, thus overcoming the need to determine complex segmentation directions. To seal tracheal wall gaps, a morphological closing operation is utilized to extract the characteristics of small holes and locate low-brightness tracheal wall gaps. By filling the CT values at these gaps, the method seals the tracheal wall gaps. Extensive experiments on the EXACT09 dataset demonstrate that our algorithm ranks third in segmentation completeness. Moreover, its performance in preventing airway leaks is significantly better than the top-two algorithms, effectively preventing large-scale leak-induced spread. [ABSTRACT FROM AUTHOR]

Objectives: Quantitative ultrasound (QUS) is a noninvasive imaging technique that can be used for assessing response to anticancer treatment. In the present study, tumor cell death response to the ultrasound‐stimulated microbubbles (USMB) and hyperthermia (HT) treatment was monitored in vivo using QUS. Methods: Human breast cancer cell lines (MDA‐MB‐231) were grown in mice and were treated with HT (10, 30, 50, and 60 minutes) alone, or in combination with USMB. Treatment effects were examined using QUS with a center frequency of 25 MHz (bandwidth range: 16 to 32 MHz). Backscattered radiofrequency (RF) data were acquired from tumors subjected to treatment. Ultrasound parameters such as average acoustic concentration (AAC) and average scatterer diameter (ASD), were estimated 24 hours prior and posttreatment. Additionally, texture features: contrast (CON), correlation (COR), energy (ENE), and homogeneity (HOM) were extracted from QUS parametric maps. All estimated parameters were compared with histopathological findings. Results: The findings of our study demonstrated a significant increase in QUS parameters in both treatment conditions: HT alone (starting from 30 minutes of heat exposure) and combined treatment of HT plus USMB finally reaching a maximum at 50 minutes of heat exposure. Increase in AAC for 50 minutes HT alone and USMB +50 minutes was found to be 5.19 ± 0.417% and 5.91 ± 1.11%, respectively, compared to the control group with AAC value of 1.00 ± 0.44%. Furthermore, between the treatment groups, ΔASD‐ENE values for USMB +30 minutes HT significantly reduced, depicting 0.00062 ± 0.00096% compared to 30 minutes HT only group, showing 0.0058 ± 0.0013%. Further, results obtained from the histological analysis indicated greater cell death and reduced nucleus size in both HT alone and HT combined with USMB. Conclusion: The texture‐based QUS parameters indicated a correlation with microstructural changes obtained from histological data. This work demonstrated the use of QUS to detect HT treatment effects in breast cancer tumors in vivo. [ABSTRACT FROM AUTHOR]