http://hdl.handle.net/10453/35217 2026-07-21T21:20:12Z http://hdl.handle.net/10453/195699 Title: Robust Non-Negative Matrix Tri-Factorization with Dual Hyper-Graph Regularization Authors: Yu, J; Che, H; Leung, MF; Liu, C; Wu, W; Yan, Z Abstract: Non-negative Matrix Factorization (NMF) has been an ideal tool for machine learning. Non-negative Matrix Tri-Factorization (NMTF) is a generalization of NMF that incorporates a third non-negative factorization matrix, and has shown impressive clustering performance by imposing simultaneous orthogonality constraints on both sample and feature spaces. However, the performance of NMTF dramatically degrades if the data are contaminated with noises and outliers. Furthermore, the high-order geometric information is rarely considered. In this paper, a Robust NMTF with Dual Hyper-graph regularization (namely RDHNMTF) is introduced. Firstly, to enhance the robustness of NMTF, an improvement is made by utilizing the l<inf>2,1</inf>-norm to evaluate the reconstruction error. Secondly, a dual hyper-graph is established to uncover the higher-order inherent information within sample space and feature spaces for clustering. Furthermore, an alternating iteration algorithm is devised, and its convergence is thoroughly analyzed. Additionally, computational complexity is analyzed among comparison algorithms. The effectiveness of RDHNMTF is verified by benchmarking against ten cuttina-edae alaorithms across seven datasets corrupted with four types of noise. 2025-01-01T00:00:00Z http://hdl.handle.net/10453/195696 Title: Sustainable design of hollow macroporous chitin carrier with enhanced loading capacity for high-performance lipase bioreactor. Authors: Zhang, J; Lei, J; Li, K; Feng, S; Liu, S; Xu, J; Wang, Z Abstract: A hollow macroporous chitin carrier (HMCC) was developed as a high-efficiency platform for lipase immobilization, offering superior specific surface area, biocompatibility, and protein affinity. Using a NaOH/urea/water treatment system, chitin was dissolved and processed into HMCC, leveraging its amphiphilic properties to achieve remarkable interfacial performance. Compared to the conventional chitin carrier (~0.34 mg/g), HMCC-OA exhibited a highest 3.39-fold increase in protein loading capacity. Immobilization of Geobacillus thermocatenulatus lipase 2 (GTL2) on HMCC-OA yielded a hydrolytic activity of 145.95 U/mg, surpassing free GTL2 by 1.14-fold. In addition, the HMCC demonstrated exceptional versatility by supporting various commercial lipases while showcasing optimal hydrolytic efficiency with GTL2. Notably, GTL2 bioreactor mitigated thermal inactivation at elevated temperatures (up to 90 °C), retained high activity across short- to medium-chain fatty acids (C4, C8, C10), and maintained significant residual activity (51.79 U/mg) after 10 recycling cycles. These results highlight the innovative design and robust performance of HMCC as a sustainable carrier for lipase bioreactor, enabling enhanced biocatalytic performance demanding thermal stability and interfacial activity. 2025-05-01T00:00:00Z http://hdl.handle.net/10453/195684 Title: Knowledge-aware contrastive heterogeneous molecular graph learning. Authors: Chen, M; Wu, J; Pan, S; Lin, F; Du, B; Gong, X; Hu, W Editors: Fariselli, P Abstract: Molecular representation learning is pivotal in predicting molecular properties and advancing drug design. Traditional methodologies, which predominantly rely on homogeneous graph encoding, are limited by their inability to integrate external knowledge and represent molecular structures across different levels of granularity. To address these limitations, we propose a paradigm shift by encoding molecular graphs into heterogeneous structures, introducing a novel framework: Knowledge-aware Contrastive Heterogeneous Molecular Graph Learning. This approach leverages contrastive learning to enrich molecular representations with embedded external knowledge. KCHML conceptualizes molecules through three distinct graph views-molecular, elemental, and pharmacological-enhanced by heterogeneous molecular graphs and a dual message-passing mechanism. This design offers a comprehensive representation for property prediction, as well as for downstream tasks such as drug-drug interaction prediction. Extensive benchmarking demonstrates KCHML's superiority over state-of-the-art molecular property prediction models, underscoring its ability to capture intricate molecular features. 2025-05-01T00:00:00Z http://hdl.handle.net/10453/195675 Title: Isolation of two novel anti-glycation peptides from highland barley (Hordeum vulgare L.) protein hydrolysates: Structural characteristics and mechanism of action against advanced glycation end-products (AGEs) Authors: Phyo, SH; Siddique, MS; Khan, I; Li, C; Zhao, W 2025-09-01T00:00:00Z