3D Object Detection and Instance Segmentation from 3D Range and 2D Color Images.

Instance segmentation and object detection are significant problems in the fields of computer vision and robotics. We address those problems by proposing a novel object segmentation and detection system. First, we detect 2D objects based on RGB, depth only, or RGB-D images. A 3D convolutional-based system, named Frustum VoxNet, is proposed. This system generates frustums from 2D detection results, proposes 3D candidate voxelized images for each frustum, and uses a 3D convolutional neural network (CNN) based on these candidates voxelized images to perform the 3D instance segmentation and object detection. Results on the SUN RGB-D dataset show that our RGB-D-based system's 3D inference is much faster than state-of-the-art methods, without a significant loss of accuracy. At the same time, we can provide segmentation and detection results using depth only images, with accuracy comparable to RGB-D-based systems. This is important since our methods can also work well in low lighting conditions, or with sensors that do not acquire RGB images. Finally, the use of segmentation as part of our pipeline increases detection accuracy, while providing at the same time 3D instance segmentation.

COVID-19 Multi-Targeted Drug Repurposing Using Few-Shot Learning.

The life-threatening disease COVID-19 has inspired significant efforts to discover novel therapeutic agents through repurposing of existing drugs. Although multi-targeted (polypharmacological) therapies are recognized as the most efficient approach to system diseases such as COVID-19, computational multi-targeted compound screening has been limited by the scarcity of high-quality experimental data and difficulties in extracting information from molecules. This study introduces MolGNN, a new deep learning model for molecular property prediction. MolGNN applies a graph neural network to computational learning of chemical molecule embedding. Comparing to state-of-the-art approaches heavily relying on labeled experimental data, our method achieves equivalent or superior prediction performance without manual labels in the pretraining stage, and excellent performance on data with only a few labels. Our results indicate that MolGNN is robust to scarce training data, and hence a powerful few-shot learning tool. MolGNN predicted several multi-targeted molecules against both human Janus kinases and the SARS-CoV-2 main protease, which are preferential targets for drugs aiming, respectively, at alleviating cytokine storm COVID-19 symptoms and suppressing viral replication. We also predicted molecules potentially inhibiting cell death induced by SARS-CoV-2. Several of MolGNN top predictions are supported by existing experimental and clinical evidence, demonstrating the potential value of our method.

GdPO4-Based Nanoprobe for Bioimaging and Selective Recognition of Dipicolinic Acid and Cysteine by a Sensing Ensemble Approach.

Multiple functions incorporated in one single-component nanoplatform pave the way for important biomedicine applications. Herein, a multifunctional terbium-doped gadolinium orthophosphate (GdPO4:Tb-EDTA) nanoplatform was prepared through a simple, ecofriendly, one-step hydrothermal method. Results showed that dipicolinic acid (DPA), the biomarker of bacterial spores, significantly increased the fluorescence intensity of this nanoplatform and conferred it with rapid response and excellent selectivity. Subsequently, the fluorescence of the ensemble GdPO4:Tb-EDTA-DPA can be remarkably quenched by Cu2+, which led to a rewritable nanosensor used in the detection of cysteine (Cys) with excellent sensitivity. In addition, GdPO4:Tb-EDTA can also be a potential T1-weighted magnetic resonance imaging (MRI) contrast agent, which indicated a satisfactory in vitro MRI with r1 relaxivity values of 13.9 mM-1 s-1 and in vivo MRI through intravenous administration on a rat model. Overall, the proposed assay may have great theoretical and practical significance for designing multifunctional biomaterials.

A lanthanide-based magnetic nanosensor as an erasable and visible platform for multi-color point-of-care detection of multiple targets and the potential application by smartphone.

The sensitive, selective and point-of-care detection of dipicolinic acid (DPA) is of great significance for the prevention of the anthrax virus and the containment of bioterrorism. In this work, a multi-color fluorescent nanoprobe composed of lanthanides and magnetic nanoparticles (Fe3O4@CePO4:Tb-EDTA-Eu) has been designed, in which the portion of Fe3O4@CePO4:Tb can be used as the internal stable signal of green fluorescence, while the EDTA-Eu part can be used as the sensitive reaction signal for monitoring DPA. Upon the addition of DPA, the red fluorescence of Eu3+ ions is significantly enhanced, while the fluorescent color of the nanoprobes can change from green to red (such as yellow-green, orange-yellow and orange-red), achieving visual multi-color fluorescent detection even by the naked eye. By using the magnetic separation method, the composites can be easily purified for point-of-care testing. More importantly, the nanoprobe fixed test pieces enable real-time analysis of DPA by using an easy-to-access color-scanning application on a smartphone. Furthermore, the fluorescence intensity can be quenched by the addition of Cu2+, which leads to a rewritable nanosensor and can be used in the detection of cysteine (Cys) with high sensitivity. With the addition of Cys, this erasable nano detection platform can also display the original multi-color visual point-of-care detection. With further optimization, this new type of multi-color fluorescent assay is promising in point-of-care clinics for multi-target diagnostics.

A novel visual ratiometric fluorescent sensing platform for highly-sensitive visual detection of tetracyclines by a lanthanide- functionalized palygorskite nanomaterial.

A palygorskite (Pal)-based ratiometric fluorescent nanoprobe is designed in order to establish a real time, on-site visual, and highly sensitive detection method for tetracyclines (TCs). The nanoprobe comprises the green emissive dye molecules embedded in the natural Pal, which serve as the internal reference signal. The potential red-emissive seed-europium (Eu3+) ions are covalently bound on the surface of modified Pal, and they can act as the specific recognition element. The emission intensity of Eu3+ ions significantly increases upon TC addition. The nanoprobe fluorescence changes from green to yellow, orange, or red, thereby accomplishing the visual ratiometric fluorescent detection. This nanoprobe exhibits a high sensitivity with a detection limit of 7.1nM and an excellent selectivity in monitoring the levels of TCs in milk samples. In addition, this nanoprobe is useful for quantitative determination of TCs, and it is not affected with intensity fluctuations due to instrumental or environmental factors. The nanoprobe-immobilized test paper realizes real-time TCs analysis by using a smartphone with an easy-to-access color-scanning APP as the detection platform. Moreover, the reported construction of visual ratiometric detection system follows the sustainable development idea, that is, from nature, for nature, and into the nature.

AuPd Bimetallic Nanocrystals Embedded in Magnetic Halloysite Nanotubes: Facile Synthesis and Catalytic Reduction of Nitroaromatic Compounds.

In this research, a facile and effective approach was developed for the preparation of well-designed AuPd alloyed catalysts supported on magnetic halloysite nanotubes (HNTs@Fe3O4@AuPd). The microstructure and the magnetic properties of HNTs@Fe3O4@AuPd were confirmed by transmission electron microscopy (TEM), high resolution TEM (HRTEM), energy-dispersive X-ray spectroscopy (EDS), and vibrating sample magnetometry (VSM) analyses. The catalysts, fabricated by a cheap, environmentally friendly, and simple surfactant-free formation process, exhibited high activities during the reduction of 4-nitrophenol and various other nitroaromatic compounds. Moreover, the catalytic activities of the HNTs@Fe3O4@AuPd nanocatalysts were tunable via adjusting the atomic ratio of AuPd during the synthesis. As compared with the monometallic nanocatalysts (HNTs@Fe3O4@Au and HNTs@Fe3O4@Pd), the bimetallic alloyed HNTs@Fe3O4@AuPd nanocatalysts exhibited excellent catalytic activities toward the reduction of 4-nitrophenol (4-NP) to 4-aminophenol. Furthermore, the as-obtained HNTs@Fe3O4@AuPd can be recycled several times, while retaining its functionality due to the stability and magnetic separation property.

Using fluorescently-labeled magnetic nanocomposites as a dual contrast agent for optical and magnetic resonance imaging.

Dual-modality imaging probes synergistically combine magnetic resonance (MR) and fluorescence into a single nanocomposite. This promising technique affords a new level of flexibility for molecular imaging uses in biomedical research. In this study, we report a new strategy for the synthesis of a novel attapulgite nanorod-based atta@Fe3O4@[Ru(bpy)2(fmp)]Cl2 nanocomposite (atta@Fe3O4@Ru NC). Our synthesized NC has both photoluminescent and magnetic properties, bright fluorescence, as well as significant magnetic resonance. Transmission electron microscopy, energy dispersive spectroscopy, fluorescence spectrometry, and magnetization measurements were all used to validate its properties. In vitro studies showed that our functionalized NC had high cellular biocompatibility and was successfully used to label living cells through endocytosis of cells. Moreover, a CCK8 assay showed that even high concentrations of the atta@Fe3O4@Ru NC had low toxicity. Finally, the intravenous administration of the atta@Fe3O4@Ru NC to a rabbit model of hepatic carcinoma resulted in a marked and negatively enhanced T2-weighted MRI in both normal liver and tumor, which can further enhance the visibility of the liver cancer tissue and normal liver tissue. Collectively, these results suggest that the atta@Fe3O4@Ru NC can be used for tumor discovery and diagnosis.

Facile ratiometric fluorapatite nanoprobes for rapid and sensitive bacterial spore biomarker detection.

Since bacterial spores, such as Bacillus anthracis spores, are extremely hazardous to human beings and animals, efforts have focused on the development of bacterial spore detector with rapid response and high selectivity and sensitivity. Therefore, we reported a facile one-step chelating-reagent-assisted hydrothermal synthesis of lanthanide-doped fluorapatite (FA) nanoprobes for detecting the biomarker of bacterial spores. In FA synthesis, ethylenediaminetetraacetic acid (EDTA) can serve not only as a shape controller and a stabilizer but also as a chelating reagent for lanthanide ions. After the fabrication, terbium (or europium) ions were not only embedded into lattice models of FA nanocrystals, but also chelated by EDTA on the surface of the FA nanocrystals, which can be available for detecting the biomarker (i.e., dipicolinic acid) of bacterial spores. The obtained FA: Ln-EDTA nanoprobes can provide reference fluorescence which play crucial roles in the calibration and correction of concentration. The fluorescence intensities of FA: Tb-EDTA or FA: Eu-EDTA from Tb3+ or Eu3+ emission were highly sensitive and increased gradually with increasing DPA concentrations. The limit of detection of FA: Tb-EDTA was 8.2nM, which is lower than that of FA: Eu-EDTA (20.9nM). It can be attributed to the fact that DPA can achieve a more effective energy transfer to Tb3+ than to Eu3+. Additionally, the nanoprobes were successfully applied for the determination of DPA in fetal calf serum samples. Given their rapid response and high sensitivity and selectivity, these ratiometric fluorescent nanoprobes are promising tools for the detection of bacterial spores.