Jiangxi Academy of Sciences

Accurate prediction of drug metabolism and pharmacokinetics (ADMET) properties is crucial in drug discovery. Here, we present a novel approach to enhance ADMET property predictions using Cross-Aligned Multimodal Attention (CMA) mechanisms, pretrained models, and multimodal techniques. ADMET data is collected and processed using image processing, graph neural networks, and chemical fingerprinting. Pretrained models like GROVER and ResNet generate a multi-channel data format, and the CMA mechanism aligns and correlates the data modalities. Grad-CAM technology interprets the model's predictions, visually demonstrating the relationship between compound properties and fragments. Our ADMET property prediction server ( http://guolab.mpu.edu.mo/CMA ) implements the CMA-based model and a substantial language model for ADMET property prediction. The innovation lies in the integration of multimodal data, the application of pretrained models, and the development of cross-modal alignment. This approach improves the efficiency and accuracy of ADMET property predictions and opens new avenues for research in molecular science, particularly in drug design and evaluation.

Against the backdrop of the United Nations Sustainable Development Goals (SDGs), innovation-driven development serves as the core engine of long-term sustainable economic development, while resource misallocation has emerged as a critical bottleneck constraining sustainable innovation and coordinated regional development. Grounded in the neoclassical theory of factor allocation, this paper incorporates capital misallocation, labor misallocation and the digital economy into a unified analytical framework. Using China’s provincial panel data spanning 2001 to 2024, we systematically investigate the impact effects, underlying transmission mechanisms and regional heterogeneity of resource misallocation and the digital economy on scientific and technological innovation through benchmark regression, robustness tests and heterogeneity analysis. The results show that resource misallocation exerts a significant and robust inhibitory effect on technological innovation, with the inhibitory effect of capital misallocation being more pronounced than that of labor misallocation. The digital economy has a significant positive driving effect on technological innovation, and it can also indirectly boost technological innovation by alleviating resource misallocation, with its mitigating effect on resource misallocation presenting obvious structural differences and a stronger correction effect on capital misallocation than on labor misallocation. Economic growth and technological innovation form a mutually reinforcing “growth-innovation” virtuous cycle. In addition, the innovation effects of both resource misallocation and the digital economy exhibit significant regional heterogeneity, where the digital economy’s innovation-driven effect and misallocation-mitigating effect are notably stronger in eastern China than in the central and western regions. The theoretical contribution of this paper lies in constructing a transmission mechanism framework of “digital economy to resource misallocation to technological innovation”, which enriches the connotations of factor allocation and innovation theories. Its practical value is to provide policymakers with differentiated development paths for the digital economy and optimization strategies for factor allocation, thus facilitating the effective implementation of the innovation-driven development strategy.

ABSTRACT Laser‐enhanced contact optimization (LECO) technology has attracted growing attention to further improve the photovoltaic conversion efficiency (PCE) of tunnelling oxide passivated contact (TOPCon) solar cells. Although LECO is known to improve metallization via current‐induced firing, the specific interaction between interfacial microstructure evolution, contact resistance, and carrier recombination remains not fully understood. Herein, we clarify the underlying mechanisms of silver‐silicon contact formation on the p + emitter during fast firing and subsequent LECO processing. We identify that a relatively low peak firing temperature followed by a secondary LECO treatment is more promising than increasing the firing temperature. Specifically, while the secondary LECO process leads to a slight increase in localized specific contact resistivity, it significantly increases the areal density of effective Ag–Si contact points, including Ag–Si alloy. This increased contact density compensates for the local resistivity rise through a parallel resistance network mechanism, ensuring efficient carrier extraction. Furthermore, unlike high‐temperature firing, which aggravates passivation damage, the secondary LECO acts as a localized annealing process, simultaneously minimizing metal‐induced recombination. Consequently, this optimized strategy endows the champion cell achieving a PCE of 26.69%. This work provides critical insight into balancing carrier transport and recombination, offering a promising route toward the development of high‐efficiency TOPCon solar cells.

Prunus tomentosa is a fruit tree native to China, with a broad distribution and high conservation value from ornamental, economic, and ecological perspectives. The germplasm resource is valuable for developing new cherry cultivars due to its high adaptability to almost all soil types and climatic conditions. However, little is known about the genetic diversity, phylogeographic structure and historical events, with its abundant genetic information. Here, we clarified the phylogeography of P. tomentosa based on chloroplast DNA (cpDNA) fragments and nuclear ribosomal internal transcribed spacer (ITS) across 779 individuals from 42 populations. We found 26 unique chloroplast haplotypes and 23 unique ITS haplotypes. Haplotype diversity was high based on cpDNA (Hd = 0.689) and ITS (Hd = 0.877) haplotypes. We observed a significant difference in the genetic differentiation based on Wright’s fixation index (cpDNA: Fst = 0.6077; ITS: Fst = 0.3447). Comparing two genetic differentiation indices, Nst and Gst, it was inferred that the P. tomentosa populations have a phylogeographical structure. Combination of network map, phylogenetic tree, population distribution, and population structure of P. tomentosa haplotype revealed two phylogeographic groups—eastern phylogeographic group (EG) and western phylogeographic group (WG). The divergence time of EG and WG occurred about 12.31 Ma and the common ancestor of P. tomentosa may have emerged during the Miocene period, and a significant correlation was observed between genetic distance and geographical distance of the populations. Signals from cpDNA and ITS data indicate that P. tomentosa has undergone recombination and population expansion. The expansion time was about 5.74 Ka based on the chloroplast data. Our findings illuminated the genetic diversity, species divergence and phylogeographical history of P. tomentosa and provided robust genetic evidence to support that Xiaolong Mountains, Hua Mountains, and Qian Mountains might be the glacial refuges of P. tomentosa. Our findings will not only guide conservation and utilization strategies for P. tomentosa but also contribute to the development of germplasm for breeding novel cherry cultivars.

The laser-clad Fe45 alloy coating inherently comprises multiple crystalline phases, resulting in a heterogeneous microstructural distribution that influences its performance. In this study, the rare earth yttria (Y2O3) was employed to modify laser-clad Fe45 alloy coatings, and the effects of Y2O3 addition on their microstructure, microhardness, and tribological properties were investigated. As the Y2O3 content increases from 0% to 0.3wt.%, the dominant microstructure transforms from columnar crystals to fine cellular and equiaxed crystals. The modified coating with 0.3wt.% Y2O3 achieves a surface hardness of 568 HV0.3 and a wear volume of 1,735.41 µm3, representing a 14.06% increase in hardness and a 51.16% reduction in wear volume compared to the undoped coating. Further increasing the Y2O3 content from 0.3wt.% to 0.9wt.% gradually leads to the emergence of a coarser feather-like microstructure, characterized by a dendritic framework with inter-dendritic equiaxed crystals. Concurrently, both the hardness and wear resistance of the coating decrease. Nevertheless, all Y2O3-modified coatings surpass the undoped Fe45 coating in both hardness and wear resistance. Appropriate Y2O3 doping effectively refines the Fe45 alloy coating’s microstructure and induces lattice distortion, thereby enhancing its hardness and wear resistance.