Meta-DHGNN: method for CRS-related cytokines analysis in CAR-T therapy based on meta-learning directed heterogeneous graph neural network.

Chimeric antigen receptor T-cell (CAR-T) immunotherapy, a novel approach for treating blood cancer, is associated with the production of cytokine release syndrome (CRS), which poses significant safety concerns for patients. Currently, there is limited knowledge regarding CRS-related cytokines and the intricate relationship between cytokines and cells. Therefore, it is imperative to explore a reliable and efficient computational method to identify cytokines associated with CRS. In this study, we propose Meta-DHGNN, a directed and heterogeneous graph neural network analysis method based on meta-learning. The proposed method integrates both directed and heterogeneous algorithms, while the meta-learning module effectively addresses the issue of limited data availability. This approach enables comprehensive analysis of the cytokine network and accurate prediction of CRS-related cytokines. Firstly, to tackle the challenge posed by small datasets, a pre-training phase is conducted using the meta-learning module. Consequently, the directed algorithm constructs an adjacency matrix that accurately captures potential relationships in a more realistic manner. Ultimately, the heterogeneous algorithm employs meta-photographs and multi-head attention mechanisms to enhance the realism and accuracy of predicting cytokine information associated with positive labels. Our experimental verification on the dataset demonstrates that Meta-DHGNN achieves favorable outcomes. Furthermore, based on the predicted results, we have explored the multifaceted formation mechanism of CRS in CAR-T therapy from various perspectives and identified several cytokines, such as IFNG (IFN-γ), IFNA1, IFNB1, IFNA13, IFNA2, IFNAR1, IFNAR2, IFNGR1 and IFNGR2 that have been relatively overlooked in previous studies but potentially play pivotal roles. The significance of Meta-DHGNN lies in its ability to analyze directed and heterogeneous networks in biology effectively while also facilitating CRS risk prediction in CAR-T therapy.

BertTCR: a Bert-based deep learning framework for predicting cancer-related immune status based on T cell receptor repertoire.

The T cell receptor (TCR) repertoire is pivotal to the human immune system, and understanding its nuances can significantly enhance our ability to forecast cancer-related immune responses. However, existing methods often overlook the intra- and inter-sequence interactions of T cell receptors (TCRs), limiting the development of sequence-based cancer-related immune status predictions. To address this challenge, we propose BertTCR, an innovative deep learning framework designed to predict cancer-related immune status using TCRs. BertTCR combines a pre-trained protein large language model with deep learning architectures, enabling it to extract deeper contextual information from TCRs. Compared to three state-of-the-art sequence-based methods, BertTCR improves the AUC on an external validation set for thyroid cancer detection by 21 percentage points. Additionally, this model was trained on over 2000 publicly available TCR libraries covering 17 types of cancer and healthy samples, and it has been validated on multiple public external datasets for its ability to distinguish cancer patients from healthy individuals. Furthermore, BertTCR can accurately classify various cancer types and healthy individuals. Overall, BertTCR is the advancing method for cancer-related immune status forecasting based on TCRs, offering promising potential for a wide range of immune status prediction tasks.

PrCRS: a prediction model of severe CRS in CAR-T therapy based on transfer learning.

BACKGROUND: CAR-T cell therapy represents a novel approach for the treatment of hematologic malignancies and solid tumors. However, its implementation is accompanied by the emergence of potentially life-threatening adverse events known as cytokine release syndrome (CRS). Given the escalating number of patients undergoing CAR-T therapy, there is an urgent need to develop predictive models for severe CRS occurrence to prevent it in advance. Currently, all existing models are based on decision trees whose accuracy is far from meeting our expectations, and there is a lack of deep learning models to predict the occurrence of severe CRS more accurately. RESULTS: We propose PrCRS, a deep learning prediction model based on U-net and Transformer. Given the limited data available for CAR-T patients, we employ transfer learning using data from COVID-19 patients. The comprehensive evaluation demonstrates the superiority of the PrCRS model over other state-of-the-art methods for predicting CRS occurrence. We propose six models to forecast the probability of severe CRS for patients with one, two, and three days in advance. Additionally, we present a strategy to convert the model's output into actual probabilities of severe CRS and provide corresponding predictions. CONCLUSIONS: Based on our findings, PrCRS effectively predicts both the likelihood and timing of severe CRS in patients, thereby facilitating expedited and precise patient assessment, thus making a significant contribution to medical research. There is little research on applying deep learning algorithms to predict CRS, and our study fills this gap. This makes our research more novel and significant. Our code is publicly available at https://github.com/wzy38828201/PrCRS . The website of our prediction platform is: http://prediction.unicar-therapy.com/index-en.html .

Contrast-enhanced phase-resolved second harmonic generation microscopy.

The characterization of inverted structures (crystallographic, ferroelectric, or magnetic domains) is crucial in the development and application of novel multi-state devices. However, determining these inverted structures needs a sensitive probe capable of revealing their phase correlation. Here a contrast-enhanced phase-resolved second harmonic generation (SHG) microscopy is presented, which utilizes a phase-tunable Soleil-Babinet compensator and the interference between the SHG fields from the inverted structures and a homogeneous reference. By this means, such inverted structures are correlated through the π-phase difference of SHG, and the phase difference is ultimately converted into the intensity contrast. As a demonstration, we have applied this microscopy in two scenarios to determine the inverted crystallographic domains in two-dimensional van der Waals material MoS2. Our method is particularly suitable for applying in vacuum and cryogenic environments while providing optical diffraction-limited resolution and arbitrarily adjustable contrast. Without loss of generality, this contrast-enhanced phase-resolved SHG microscopy can also be used to resolve other non-centrosymmetric inverted structures, e.g. ferroelectric, magnetic, or multiferroic phases.

Twist Phase Matching in Two-Dimensional Materials.

Optical phase matching involves establishing a proper phase relationship between the fundamental excitation and generated waves to enable efficient optical parametric processes. It is typically achieved through birefringence or periodic polarization. Here, we report that the interlayer twist angle in two-dimensional (2D) materials creates a nonlinear geometric phase that can compensate for the phase mismatch, and the vertical assembly of the 2D layers with a proper twist sequence generates a nontrivial "twist-phase-matching" (twist-PM) regime. The twist-PM model provides superior flexibility in the design of optical crystals, which can be applied for twisted layers with either periodic or random thickness distributions. The designed crystal from the twisted rhombohedral boron nitride films within a thickness of only 3.2 µm is capable of producing a second-harmonic generation with conversion efficiency of ∼8% and facile polarization controllability that is absent in conventional crystals. Our methodology establishes a platform for the rational design and atomic manufacturing of nonlinear optical crystals based on abundant 2D materials.

Mechanisms of Actinidia chinensis Planch in treating colon cancer based on the integration of network pharmacology, molecular docking, and experimental verification.

BACKGROUND: As an anticancer Chinese herbal medicine, the effective components and mechanism of Actinidia chinensis Planch (ACP, Tengligen) in the treatment of colon cancer are still unclear. In the present study, the integration of network pharmacology, molecular docking, and cell experiments was employed to study the effective mechanism of ACP against colon cancer. METHODS: The Venn diagram and STRING database were used to construct the protein-protein interaction network (PPI) of ACP-colon cancer, and further topological analysis was used to obtain the key target genes of ACP in colon cancer. The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were used to visualize the related functions and pathways. Molecular docking between key targets and compounds was determined using software such as AutoDockTools. Finally, the effect of ACP on CT26 cells was observed in vitro. RESULTS: The study identified 40 ACP-colon key targets, including CASP3, CDK2, GSK3B, and PIK3R1. GO and KEGG enrichment analyses found that these genes were involved in 211 biological processes and 92 pathways, among which pathways in cancer, PI3K-Akt, p53, and cell cycle might be the main pathways of ACP against colon cancer. Molecular docking verified that the key components of ACP could stably bind to the corresponding targets. The experimental results showed that ACP could inhibit proliferation, induce apoptosis, and downregulate the phosphorylation of PIK3R1, Akt, and GSK3B in CT26 cells. CONCLUSION: ACP is an anti-colon cancer herb with multiple components, and involvement of multiple target genes and signaling pathways. ACP can significantly inhibit proliferation and induce apoptosis of colon cancer cells, which may be closely related to the regulation of PI3K/AKT/GSK3B signal transduction.

Creep level qualitative evaluating of crushed rock based on uncertainty measurement theory and hierarchical analysis.

A large number of tectonically mixed rock belts and complex tectonic zones are distributed in the southwestern part of China. In these areas, high geostress and tectonic stresses have caused some underground rock layers to be crushed and broken, eventually forming crushed rock zones. Which may undergo creep deformation under long-term loads. The manuscript is based on a typical crushed rock in the southwestern China. Firstly, the factors affecting creep deformation were analysed, and the response law of each influencing factor to rock creep is demonstrated. Then, the theory of uncorroborated measures and hierarchical analysis were used to systematically correlate the factors influencing creep. Thereby, a creep level qualitative evaluating model of crushed rock is established. Finally, this model was used to qualitatively evaluate the creep level of the crushed rock in the study area. It is concluded that the creep level qualitative evaluating of this crushed rock is rated as Class II, which is characterised by a low creep level and small creep deformations (0-10 mm). The research results can provide a reference for the creep analysis of crushed rock and provide a basis for the safe construction of engineering slopes.

An efficient ultrasonic wavenumber-domain plane wave imaging method towards the inspection of curved structures.

Ultrasonic phased array testing is commonly employed for inspecting curved structures. Conventional plane wave imaging techniques, based on delay-and-sum in the time-domain, offer high image quality and inspection accuracy but suffer from low frame rates due to their high computational complexity. In this work, an efficient wavenumber-domain imaging method that combines non-stationary wavefield extrapolation and f-k migration is proposed for curved structure inspection. Special emission focal laws are designed to generate a sequence of steered plane waves through the curved interface. The raw data is then extrapolated to the top boundary of the region of interest, followed by f-k migration to reconstruct images with high time efficiency. Simulation and experimental evaluations demonstrate a time reduction by a factor of up to 32.24 compared to conventional time-domain plane wave image reconstruction with equivalent image quality, highlighting its potential for monitoring flaws in real-time.

Nonlinear optical signatures of topological Dirac fermion.

Dirac fermion in topological materials exhibits intriguing nonlinear optical responses. However, their direct correlation with the linearly dispersed band remains elusive experimentally. Here, we take topological semimetal ZrSiS as a paradigm, unveiling three unique nonlinear optical signatures of Dirac fermion. These signatures include strong quadrupolar response, quantum interference effect, and exponential divergent four-wave mixing (FWM), all of which are described by the prominent third-order nonlinear optical susceptibility. Resonantly enhanced by linear bands, quadrupolar second harmonic generation in centrosymmetric bulk overwhelms the electric-dipole contribution at the surface with inherent inversion symmetry breaking. Furthermore, owing to the interference between multiple resonant transition pathways within linear bands, difference-frequency FWM is several orders of magnitude stronger than sum-frequency FWM and third harmonic generation. The difference-frequency FWM further displays an inverse-square divergence toward degenerate excitation, whose scaling law perfectly matches with the long-sought behavior of Dirac fermion. These signatures lay the solid foundation toward the practical applications of topological materials in nonlinear optoelectronics and photonics.