DANCE: a deep learning library and benchmark platform for single-cell analysis.

DANCE is the first standard, generic, and extensible benchmark platform for accessing and evaluating computational methods across the spectrum of benchmark datasets for numerous single-cell analysis tasks. Currently, DANCE supports 3 modules and 8 popular tasks with 32 state-of-art methods on 21 benchmark datasets. People can easily reproduce the results of supported algorithms across major benchmark datasets via minimal efforts, such as using only one command line. In addition, DANCE provides an ecosystem of deep learning architectures and tools for researchers to facilitate their own model development. DANCE is an open-source Python package that welcomes all kinds of contributions.

Thermosensitive hydrogel with emodin-loaded triple-targeted nanoparticles for a rectal drug delivery system in the treatment of chronic non-bacterial prostatitis.

BACKGROUND: The complex etiology and pathogenesis underlying Chronic Non-Bacterial Prostatitis (CNP), coupled with the existence of a Blood Prostate Barrier (BPB), contribute to a lack of specificity and poor penetration of most drugs. Emodin (EMO), a potential natural compound for CNP treatment, exhibits commendable anti-inflammatory, anti-oxidant, and anti-fibrosis properties but suffers from the same problems as other drugs. METHODS: By exploiting the recognition properties of lactoferrin (LF) receptors that target intestinal epithelial cells (NCM-460) and prostate epithelial cells (RWPE-1), a pathway is established for the transrectal absorption of EMO to effectively reach the prostate. Additionally, hyaluronic acid (HA) is employed, recognizing CD44 receptors which target macrophages within the inflamed prostate. This interaction facilitates the intraprostatic delivery of EMO, leading to its pronounced anti-inflammatory effects. A thermosensitive hydrogel (CS-Gel) prepared from chitosan (CS) and ß-glycerophosphate disodium salt (ß-GP) was used for rectal drug delivery with strong adhesion to achieve effective drug retention and sustained slow release. Thus, we developed a triple-targeted nanoparticle (NPs)/thermosensitive hydrogel (Gel) rectal drug delivery system. In this process, LF, with its positive charge, was utilized to load EMO through dialysis, producing LF@EMO-NPs. Subsequently, HA was employed to encapsulate EMO-loaded LF nanoparticles via electrostatic adsorption, yielding HA/LF@EMO-NPs. Finally, HA/LF@EMO-NPs lyophilized powder was added to CS-Gel (HA/LF@EMO-NPs Gel). RESULTS: Cellular assays indicated that NCM-460 and RWPE-1 cells showed high uptake of both LF@EMO-NPs and HA/LF@EMO-NPs, while Raw 264.7 cells exhibited substantial uptake of HA/LF@EMO-NPs. For LPS-induced Raw 264.7 cells, HA/LF@EMO-NPs can reduce the inflammatory responses by modulating TLR4/NF-κB signaling pathways. Tissue imaging corroborated the capacity of HA/LF-modified formulations to breach the BPB, accumulating within the gland's lumen. Animal experiments showed that rectal administration of HA/LF@EMO-NPs Gel significantly reduced inflammatory cytokine expression, oxidative stress levels and fibrosis in the CNP rats, in addition to exerting anti-inflammatory effects by inhibiting the NF-κB signaling pathway without obvious toxicity. CONCLUSION: This triple-targeted NPs/Gel rectal delivery system with slow-release anti-inflammatory, anti-oxidant, and anti-fibrosis properties shows great potential for the effective treatment of CNP.

Single-Cell Multimodal Prediction via Transformers.

The recent development of multimodal single-cell technology has made the possibility of acquiring multiple omics data from individual cells, thereby enabling a deeper understanding of cellular states and dynamics. Nevertheless, the proliferation of multimodal single-cell data also introduces tremendous challenges in modeling the complex interactions among different modalities. The recently advanced methods focus on constructing static interaction graphs and applying graph neural networks (GNNs) to learn from multimodal data. However, such static graphs can be suboptimal as they do not take advantage of the downstream task information; meanwhile GNNs also have some inherent limitations when deeply stacking GNN layers. To tackle these issues, in this work, we investigate how to leverage transformers for multimodal single-cell data in an end-to-end manner while exploiting downstream task information. In particular, we propose a scMoFormer framework which can readily incorporate external domain knowledge and model the interactions within each modality and cross modalities. Extensive experiments demonstrate that scMoFormer achieves superior performance on various benchmark datasets. Remarkably, scMoFormer won a Kaggle silver medal with the rank of 24/1221 (Top 2%) without ensemble in a NeurIPS 2022 competition. Our implementation is publicly available at Github.