The theory of electron spin has been proposed for a century, but the study of quantum effects in biological molecules is still in its infancy. Chirality-induced spin selectivity (CISS) is a very modern theory that can explain many biochemical phenomena. In this paper, we propose a new theoretical model based on CISS theory and quantum chemistry theory, which can well explain the theoretical explanation of the chiral selectivity of chiral proteins. Moreover, this theory can predict the spin state of corresponding chiral molecules. Taking L-DOPA and AADC enzyme as examples, this theoretical model solves the theoretical explanation of AADC enzyme's chiral catalysis selectivity problem and successfully predicts the spin state of L-DOPA and D-DOPA's valence electrons.

Two-dimensional (2D) ice I is an atomic-level ice that is composed of two interlocked atom-ic layers saturated with hydrogen bonds. It has recently been experimentally observed but its properties have yet to be clarified. Accordingly, we theoretically studied the hydrophobic properties of 2D ice I. On the contrary, simulation of a hydrogen fluoride molecule on a 2D ice surface manifested that it destroyed the 2d ice structure and connected new hydrogen bonds with water molecules. Investigations of the interfacial effect between 2D and three-dimensional (3D) ice films indicated that the network structure of 2D ice was not de-stroyed by a 3D ice surface, as the former was saturated with hydrogen bonds. However, the surface of 3D ice reorganized to form as many hydrogen bonds as possible. Thus, the 2D ice film was hydrophobic and inhibited the growth of 3D ice. This shows that if 2D ice can be produced on an industrial scale, it could be used as an anti-3D-icing agent under low tem-perature.

Avian infectious bronchitis virus (IBV) is the causative agent of infectious bronchitis, which causes considerable economic losses to the poultry industry worldwide. It is imperative to develop safe and efficient candidate vaccines to control IBV infection. In the current study, recombinant baculoviruses co-expressing S1 and N proteins, mono-expressing S1 or N proteins alone of IBV were constructed and prepared into subunit vaccines rHBM-S1-N, rHBM-S1 and rHBM-N. The levels of immune protection of these subunit vaccines were evaluated by inoculating specific pathogen-free (SPF) chickens at 14 days of age, boosting with the same dose 14 days later, and following challenge with a virulent GX-YL5 strain of IBV 14 days post-booster (dpb). The commercial vaccine strain H120 was used as a control. The IBV-specific antibody levels as well as the percentages of CD4+ and CD8+ T lymphocytes were detected within 28 days post-vaccination (dpv). The morbidity, mortality, and re-isolation of virus from the tracheas and kidneys of challenged birds were evaluated at 5 days post-challenge (dpc). The results showed that the IBV-specific antibody levels and the percentages of CD4+ and CD8+ T lymphocyte in rHBM-S1-N group were higher than those of rHBM-S1 and rHBM-N groups, especially the cellular immunity response. At 5 dpc, the mortality, morbidity and virus re-isolation rate of rHBM-S1-N were slightly higher than those of H120 group, but were lower than those of rHBM-S1 group and rHBM-N group. The present study demonstrated that the protection of recombinant baculovirus co-expressing S1 and N proteins was better than that of recombinant baculoviruses mono-expressing S1 or N protein alone. Thus, the recombinant baculovirus co-expressing S1 and N proteins could serve as a potential IBV vaccine and this demonstrates that the bivalent subunit vaccine including the S1 and N proteins might be a strategy for the development of an IBV subunit vaccine.

Real-time object detection remains a pivotal topic in the realm of computer vision. Balancing the accuracy and speed of object detectors poses a formidable challenge for both academic researchers and industry practitioners. While recent transformer-based models have showncase the prowess of the attention mechanism, delivering remarkable performance improvements over CNNs, their computational demands can hinder their effectiveness in real-time detection tasks. In this paper, we elect to use YOLOX as our robust starting point and introduce a series of effective enhancements, culminating in a new high-performance detector named YOLOAX. To further exploit the power of the attention mechanism, we devise multi-dimensional attention-based modules that activate CNNs, emphasizing the most pertinent regions and bolstering the capacity to learn the most informative image representations from feature maps. Moreover, we introduce a novel leading label assignment strategy called STA, along with a groundbreaking loss function named GEIOU Loss, to further refine our detector's performance. We provide extensive ablation studies on the COCO and PASCAL VOC 2012 detection datasets to validate our proposed methods. Remarkably, our YOLOAX is trained solely on the MS-COCO dataset from scratch, without leveraging any prior knowledge, which achieves an impressive 54.2% AP on the COCO 2017 test set while maintaining a real-time speed of 72.3 fps, surpassing YOLOX by a significant margin of 3.0% AP. Our source code and pre-trained models are openly accessible at https://github.com/KejianXu/yoloax.

Few-shot object detection (FSOD) aims to address the challenge of requiring a substantial amount of annotations for training in conventional object detection, which is very labor-intensive. However, existing few-shot methods achieve high precision with the sacrifice of time-consuming for exhaustive fine-tuning, or take poor performance in novel-class adaptation. We presume the major reason is that the valuable correlation feature among different categories is insufficiently exploited, hindering the generalization of knowledge from base to novel categories for object detection. In this paper, we propose Few-Shot object detection via Correlation-RPN and Transformer Encoder-Decoder (CRTED), a novel training network to learn object-relevant features of inter-class correlation and intra-class compactness while suppressing object-agnostic features in the background with limited annotated samples. And we also introduce a 4-way tuple-contrast training strategy to positively activate the training progress of our object detector. Experiments over two few-shot benchmarks (Pascal VOC, MS-COCO) demonstrate that, our proposed CRTED without further fine-tuning can achieve comparable performance with current state-of-the-art fine-tuned works. The codes and pre-trained models will be released.