Revisiting drug-protein interaction prediction: a novel global-local perspective.

MOTIVATION: Accurate inference of potential drug-protein interactions (DPIs) aids in understanding drug mechanisms and developing novel treatments. Existing deep learning models, however, struggle with accurate node representation in DPI prediction, limiting their performance. RESULTS: We propose a new computational framework that integrates global and local features of nodes in the drug-protein bipartite graph for efficient DPI inference. Initially, we employ pre-trained models to acquire fundamental knowledge of drugs and proteins and to determine their initial features. Subsequently, the MinHash and HyperLogLog algorithms are utilized to estimate the similarity and set cardinality between drug and protein subgraphs, serving as their local features. Then, an energy-constrained diffusion mechanism is integrated into the transformer architecture, capturing interdependencies between nodes in the drug-protein bipartite graph and extracting their global features. Finally, we fuse the local and global features of nodes and employ multilayer perceptrons to predict the likelihood of potential DPIs. A comprehensive and precise node representation guarantees efficient prediction of unknown DPIs by the model. Various experiments validate the accuracy and reliability of our model, with molecular docking results revealing its capability to identify potential DPIs not present in existing databases. This approach is expected to offer valuable insights for furthering drug repurposing and personalized medicine research. AVAILABILITY AND IMPLEMENTATION: Our code and data are accessible at: https://github.com/ZZCrazy00/DPI.

Prediction of miRNA-disease associations in microbes based on graph convolutional networks and autoencoders.

MicroRNAs (miRNAs) are short RNA molecular fragments that regulate gene expression by targeting and inhibiting the expression of specific RNAs. Due to the fact that microRNAs affect many diseases in microbial ecology, it is necessary to predict microRNAs' association with diseases at the microbial level. To this end, we propose a novel model, termed as GCNA-MDA, where dual-autoencoder and graph convolutional network (GCN) are integrated to predict miRNA-disease association. The proposed method leverages autoencoders to extract robust representations of miRNAs and diseases and meantime exploits GCN to capture the topological information of miRNA-disease networks. To alleviate the impact of insufficient information for the original data, the association similarity and feature similarity data are combined to calculate a more complete initial basic vector of nodes. The experimental results on the benchmark datasets demonstrate that compared with the existing representative methods, the proposed method has achieved the superior performance and its precision reaches up to 0.8982. These results demonstrate that the proposed method can serve as a tool for exploring miRNA-disease associations in microbial environments.

An efficient model for predicting human diseases through miRNA based on multiple-types of contrastive learning.

Multiple studies have demonstrated that microRNA (miRNA) can be deeply involved in the regulatory mechanism of human microbiota, thereby inducing disease. Developing effective methods to infer potential associations between microRNAs (miRNAs) and diseases can aid early diagnosis and treatment. Recent methods utilize machine learning or deep learning to predict miRNA-disease associations (MDAs), achieving state-of-the-art performance. However, the problem of sparse neighborhoods of nodes due to lack of data has not been well solved. To this end, we propose a new model named MTCL-MDA, which integrates multiple-types of contrastive learning strategies into a graph collaborative filtering model to predict potential MDAs. The model adopts a contrastive learning strategy based on topology, which alleviates the damage to model performance caused by sparse neighborhoods. In addition, the model also adopts a semantic-based contrastive learning strategy, which not only reduces the impact of noise introduced by topology-based contrastive learning, but also enhances the semantic information of nodes. Experimental results show that our model outperforms existing models on all evaluation metrics. Case analysis shows that our model can more accurately identify potential MDA, which is of great significance for the screening and diagnosis of real-life diseases. Our data and code are publicly available at: https://github.com/Lqingquan/MTCL-MDA.

A valve-based microfluidic device for on-chip single cell treatments.

Assays toward single-cell analysis have attracted the attention in biological and biomedical researches to reveal cellular mechanisms as well as heterogeneity. Yet nowadays microfluidic devices for single-cell analysis have several drawbacks: some would cause cell damage due to the hydraulic forces directly acting on cells, while others could not implement biological assays since they could not immobilize cells while manipulating the reagents at the same time. In this work, we presented a two-layer pneumatic valve-based platform to implement cell immobilization and treatment on-chip simultaneously, and cells after treatment could be collected non-destructively for further analysis. Target cells could be encapsulated in sodium alginate droplets which solidified into hydrogel when reacted with Ca2+ . The size of hydrogel beads could be precisely controlled by modulating flow rates of continuous/disperse phases. While regulating fluid resistance between the main channel and passages by the integrated pneumatic valves, on-chip capture and release of hydrogel beads was implemented. As a proof of concept for on-chip single-cell treatments, we showed cellular live/dead staining based on our devices. This method would have potential in single cell manipulation for biochemical cellular assays.

Erythrocyte membrane-coated gold nanocages for targeted photothermal and chemical cancer therapy.

Recently, red blood cell (RBC) membrane-coated nanoparticles have attracted much attention because of their excellent immune escapability; meanwhile, gold nanocages (AuNs) have been extensively used for cancer therapy due to their photothermal effect and drug delivery capability. The combination of the RBC membrane coating and AuNs may provide an effective approach for targeted cancer therapy. However, few reports have shown the utilization of combining these two technologies. Here, we design erythrocyte membrane-coated gold nanocages for targeted photothermal and chemical cancer therapy. First, anti-EpCam antibodies were used to modify the RBC membranes to target 4T1 cancer cells. Second, the antitumor drug paclitaxel (PTX) was encapsulated into AuNs. Then, the AuNs were coated with the modified RBC membranes. These new nanoparticles were termed EpCam-RPAuNs. We characterized the capability of the EpCam-RPAuNs for selective tumor targeting via exposure to near-infrared irradiation. The experimental results demonstrate that EpCam-RPAuNs can effectively generate hyperthermia and precisely deliver the antitumor drug PTX to targeted cells. We also validated the biocompatibility of the EpCam-RAuNs in vitro. By combining the molecularly modified targeting RBC membrane and AuNs, our approach provides a new way to design biomimetic nanoparticles to enhance the surface functionality of nanoparticles. We believe that EpCam-RPAuNs can be potentially applied for cancer diagnoses and therapies.

Microfluidic Electroporation-Facilitated Synthesis of Erythrocyte Membrane-Coated Magnetic Nanoparticles for Enhanced Imaging-Guided Cancer Therapy.

Biomimetic cell membrane-coated nanoparticles (CM-NPs) with superior biochemical properties have been broadly utilized for various biomedical applications. Currently, researchers primarily focus on using ultrasonic treatment and mechanical extrusion to improve the synthesis of CM-NPs. In this work, we demonstrate that microfluidic electroporation can effectively facilitate the synthesis of CM-NPs. To test it, Fe3O4 magnetic nanoparticles (MNs) and red blood cell membrane-derived vesicles (RBC-vesicles) are infused into a microfluidic device. When the mixture of MNs and RBC-vesicles flow through the electroporation zone, the electric pulses can effectively promote the entry of MNs into RBC-vesicles. After that, the resulting RBC membrane-capped MNs (RBC-MNs) are collected from the chip and injected into experimental animals to test the in vivo performance. Owing to the superior magnetic and photothermal properties of the MN cores and the long blood circulation characteristic of the RBC membrane shells, core-shell RBC-MNs were used for enhanced tumor magnetic resonance imaging (MRI) and photothermal therapy (PTT). Due to the completer cell membrane coating, RBC-MNs prepared by microfluidic electroporation strategy exhibit significantly better treatment effect than the one fabricated by conventional extrusion. We believe the combination of microfluidic electroporation and CM-NPs provides an insight into the synthesis of bioinpired nanoparticles to improve cancer diagnosis and therapy.