Tumour evolution and microenvironment interactions in 2D and 3D space.

To study the spatial interactions among cancer and non-cancer cells1, we here examined a cohort of 131 tumour sections from 78 cases across 6 cancer types by Visium spatial transcriptomics (ST). This was combined with 48 matched single-nucleus RNA sequencing samples and 22 matched co-detection by indexing (CODEX) samples. To describe tumour structures and habitats, we defined 'tumour microregions' as spatially distinct cancer cell clusters separated by stromal components. They varied in size and density among cancer types, with the largest microregions observed in metastatic samples. We further grouped microregions with shared genetic alterations into 'spatial subclones'. Thirty five tumour sections exhibited subclonal structures. Spatial subclones with distinct copy number variations and mutations displayed differential oncogenic activities. We identified increased metabolic activity at the centre and increased antigen presentation along the leading edges of microregions. We also observed variable T cell infiltrations within microregions and macrophages predominantly residing at tumour boundaries. We reconstructed 3D tumour structures by co-registering 48 serial ST sections from 16 samples, which provided insights into the spatial organization and heterogeneity of tumours. Additionally, using an unsupervised deep-learning algorithm and integrating ST and CODEX data, we identified both immune hot and cold neighbourhoods and enhanced immune exhaustion markers surrounding the 3D subclones. These findings contribute to the understanding of spatial tumour evolution through interactions with the local microenvironment in 2D and 3D space, providing valuable insights into tumour biology.

Epigenetic regulation during cancer transitions across 11 tumour types.

Chromatin accessibility is essential in regulating gene expression and cellular identity, and alterations in accessibility have been implicated in driving cancer initiation, progression and metastasis1-4. Although the genetic contributions to oncogenic transitions have been investigated, epigenetic drivers remain less understood. Here we constructed a pan-cancer epigenetic and transcriptomic atlas using single-nucleus chromatin accessibility data (using single-nucleus assay for transposase-accessible chromatin) from 225 samples and matched single-cell or single-nucleus RNA-sequencing expression data from 206 samples. With over 1 million cells from each platform analysed through the enrichment of accessible chromatin regions, transcription factor motifs and regulons, we identified epigenetic drivers associated with cancer transitions. Some epigenetic drivers appeared in multiple cancers (for example, regulatory regions of ABCC1 and VEGFA; GATA6 and FOX-family motifs), whereas others were cancer specific (for example, regulatory regions of FGF19, ASAP2 and EN1, and the PBX3 motif). Among epigenetically altered pathways, TP53, hypoxia and TNF signalling were linked to cancer initiation, whereas oestrogen response, epithelial-mesenchymal transition and apical junction were tied to metastatic transition. Furthermore, we revealed a marked correlation between enhancer accessibility and gene expression and uncovered cooperation between epigenetic and genetic drivers. This atlas provides a foundation for further investigation of epigenetic dynamics in cancer transitions.

Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics.

Analyzing somatic evolution within a tumor over time and across space is a key challenge in cancer research. Spatially resolved transcriptomics (SRT) measures gene expression at thousands of spatial locations in a tumor, but does not directly reveal genomic aberrations. We introduce CalicoST, an algorithm to simultaneously infer allele-specific copy number aberrations (CNAs) and reconstruct spatial tumor evolution, or phylogeography, from SRT data. CalicoST identifies important classes of CNAs-including copy-neutral loss of heterozygosity and mirrored subclonal CNAs-that are invisible to total copy number analysis. Using nine patients' data from the Human Tumor Atlas Network, CalicoST achieves an average accuracy of 86%, approximately 21% higher than existing methods. CalicoST reconstructs a tumor phylogeography in three-dimensional space for two patients with multiple adjacent slices. CalicoST analysis of multiple SRT slices from a cancerous prostate organ reveals mirrored subclonal CNAs on the two sides of the prostate, forming a bifurcating phylogeography in both genetic and physical space.

Structural basis of G protein-Coupled receptor CMKLR1 activation and signaling induced by a chemerin-derived agonist.

Chemokine-like receptor 1 (CMKLR1), also known as chemerin receptor 23 (ChemR23) or chemerin receptor 1, is a chemoattractant G protein-coupled receptor (GPCR) that responds to the adipokine chemerin and is highly expressed in innate immune cells, including macrophages and neutrophils. The signaling pathways of CMKLR1 can lead to both pro- and anti-inflammatory effects depending on the ligands and physiological contexts. To understand the molecular mechanisms of CMKLR1 signaling, we determined a high-resolution cryo-electron microscopy (cryo-EM) structure of the CMKLR1-Gi signaling complex with chemerin9, a nanopeptide agonist derived from chemerin, which induced complex phenotypic changes of macrophages in our assays. The cryo-EM structure, together with molecular dynamics simulations and mutagenesis studies, revealed the molecular basis of CMKLR1 signaling by elucidating the interactions at the ligand-binding pocket and the agonist-induced conformational changes. Our results are expected to facilitate the development of small molecule CMKLR1 agonists that mimic the action of chemerin9 to promote the resolution of inflammation.

CRISPR screen in regulatory T cells reveals modulators of Foxp3.

Regulatory T (Treg) cells are required to control immune responses and maintain homeostasis, but are a significant barrier to antitumour immunity1. Conversely, Treg instability, characterized by loss of the master transcription factor Foxp3 and acquisition of proinflammatory properties2, can promote autoimmunity and/or facilitate more effective tumour immunity3,4. A comprehensive understanding of the pathways that regulate Foxp3 could lead to more effective Treg therapies for autoimmune disease and cancer. The availability of new functional genetic tools has enabled the possibility of systematic dissection of the gene regulatory programs that modulate Foxp3 expression. Here we developed a CRISPR-based pooled screening platform for phenotypes in primary mouse Treg cells and applied this technology to perform a targeted loss-of-function screen of around 500 nuclear factors to identify gene regulatory programs that promote or disrupt Foxp3 expression. We identified several modulators of Foxp3 expression, including ubiquitin-specific peptidase 22 (Usp22) and ring finger protein 20 (Rnf20). Usp22, a member of the deubiquitination module of the SAGA chromatin-modifying complex, was revealed to be a positive regulator that stabilized Foxp3 expression; whereas the screen suggested that Rnf20, an E3 ubiquitin ligase, can serve as a negative regulator of Foxp3. Treg-specific ablation of Usp22 in mice reduced Foxp3 protein levels and caused defects in their suppressive function that led to spontaneous autoimmunity but protected against tumour growth in multiple cancer models. Foxp3 destabilization in Usp22-deficient Treg cells could be rescued by ablation of Rnf20, revealing a reciprocal ubiquitin switch in Treg cells. These results reveal previously unknown modulators of Foxp3 and demonstrate a screening method that can be broadly applied to discover new targets for Treg immunotherapies for cancer and autoimmune disease.

ADCdb: the database of antibody-drug conjugates.

Antibody-drug conjugates (ADCs) are a class of innovative biopharmaceutical drugs, which, via their antibody (mAb) component, deliver and release their potent warhead (a.k.a. payload) at the disease site, thereby simultaneously improving the efficacy of delivered therapy and reducing its off-target toxicity. To design ADCs of promising efficacy, it is crucial to have the critical data of pharma-information and biological activities for each ADC. However, no such database has been constructed yet. In this study, a database named ADCdb focusing on providing ADC information (especially its pharma-information and biological activities) from multiple perspectives was thus developed. Particularly, a total of 6572 ADCs (359 approved by FDA or in clinical trial pipeline, 501 in preclinical test, 819 with in-vivo testing data, 1868 with cell line/target testing data, 3025 without in-vivo/cell line/target testing data) together with their explicit pharma-information was collected and provided. Moreover, a total of 9171 literature-reported activities were discovered, which were identified from diverse clinical trial pipelines, model organisms, patient/cell-derived xenograft models, etc. Due to the significance of ADCs and their relevant data, this new database was expected to attract broad interests from diverse research fields of current biopharmaceutical drug discovery. The ADCdb is now publicly accessible at: https://idrblab.org/adcdb/.

Bubble: a fast single-cell RNA-seq imputation using an autoencoder constrained by bulk RNA-seq data.

Single-cell RNA-sequencing technology (scRNA-seq) brings research to single-cell resolution. However, a major drawback of scRNA-seq is large sparsity, i.e. expressed genes with no reads due to technical noise or limited sequence depth during the scRNA-seq protocol. This phenomenon is also called 'dropout' events, which likely affect downstream analyses such as differential expression analysis, the clustering and visualization of cell subpopulations, cellular trajectory inference, etc. Therefore, there is a need to develop a method to identify and impute these dropout events. We propose Bubble, which first identifies dropout events from all zeros based on expression rate and coefficient of variation of genes within cell subpopulation, and then leverages an autoencoder constrained by bulk RNA-seq data to only impute those values. Unlike other deep learning-based imputation methods, Bubble fuses the matched bulk RNA-seq data as a constraint to reduce the introduction of false positive signals. Using simulated and several real scRNA-seq datasets, we demonstrate that Bubble enhances the recovery of missing values, gene-to-gene and cell-to-cell correlations, and reduces the introduction of false positive signals. Regarding some crucial downstream analyses of scRNA-seq data, Bubble facilitates the identification of differentially expressed genes, improves the performance of clustering and visualization, and aids the construction of cellular trajectory. More importantly, Bubble provides fast and scalable imputation with minimal memory usage.

PEA-m6A: an ensemble learning framework for accurately predicting N6-methyladenosine modifications in plants.

N 6-methyladenosine (m6A), which is the mostly prevalent modification in eukaryotic mRNAs, is involved in gene expression regulation and many RNA metabolism processes. Accurate prediction of m6A modification is important for understanding its molecular mechanisms in different biological contexts. However, most existing models have limited range of application and are species-centric. Here we present PEA-m6A, a unified, modularized and parameterized framework that can streamline m6A-Seq data analysis for predicting m6A-modified regions in plant genomes. The PEA-m6A framework builds ensemble learning-based m6A prediction models with statistic-based and deep learning-driven features, achieving superior performance with an improvement of 6.7% to 23.3% in the area under precision-recall curve compared with state-of-the-art regional-scale m6A predictor WeakRM in 12 plant species. Especially, PEA-m6A is capable of leveraging knowledge from pretrained models via transfer learning, representing an innovation in that it can improve prediction accuracy of m6A modifications under small-sample training tasks. PEA-m6A also has a strong capability for generalization, making it suitable for application in within- and cross-species m6A prediction. Overall, this study presents a promising m6A prediction tool, PEA-m6A, with outstanding performance in terms of its accuracy, flexibility, transferability, and generalization ability. PEA-m6A has been packaged using Galaxy and Docker technologies for ease of use and is publicly available at https://github.com/cma2015/PEA-m6A.

Tropoelastin Improves Post-Infarct Cardiac Function.

BACKGROUND: Myocardial infarction (MI) is among the leading causes of death worldwide. Following MI, necrotic cardiomyocytes are replaced by a stiff collagen-rich scar. Compared to collagen, the extracellular matrix protein elastin has high elasticity and may have more favorable properties within the cardiac scar. We sought to improve post-MI healing by introducing tropoelastin, the soluble subunit of elastin, to alter scar mechanics early after MI. METHODS AND RESULTS: We developed an ultrasound-guided direct intramyocardial injection method to administer tropoelastin directly into the left ventricular anterior wall of rats subjected to induced MI. Experimental groups included shams and infarcted rats injected with either PBS vehicle control or tropoelastin. Compared to vehicle treated controls, echocardiography assessments showed tropoelastin significantly improved left ventricular ejection fraction (64.7±4.4% versus 46.0±3.1% control) and reduced left ventricular dyssynchrony (11.4±3.5 ms versus 31.1±5.8 ms control) 28 days post-MI. Additionally, tropoelastin reduced post-MI scar size (8.9±1.5% versus 20.9±2.7% control) and increased scar elastin (22±5.8% versus 6.2±1.5% control) as determined by histological assessments. RNA sequencing (RNAseq) analyses of rat infarcts showed that tropoelastin injection increased genes associated with elastic fiber formation 7 days post-MI and reduced genes associated with immune response 11 days post-MI. To show translational relevance, we performed immunohistochemical analyses on human ischemic heart disease cardiac samples and showed an increase in tropoelastin within fibrotic areas. Using RNA-seq we also demonstrated the tropoelastin gene ELN is upregulated in human ischemic heart disease and during human cardiac fibroblast-myofibroblast differentiation. Furthermore, we showed by immunocytochemistry that human cardiac fibroblast synthesize increased elastin in direct response to tropoelastin treatment. CONCLUSIONS: We demonstrate for the first time that purified human tropoelastin can significantly repair the infarcted heart in a rodent model of MI and that human cardiac fibroblast synthesize elastin. Since human cardiac fibroblasts are primarily responsible for post-MI scar synthesis, our findings suggest exciting future clinical translation options designed to therapeutically manipulate this synthesis.

MFA-DTI: Drug-target interaction prediction based on multi-feature fusion adopted framework.

The identification of drug-target interactions (DTI) is a valuable step in the drug discovery and repositioning process. However, traditional laboratory experiments are time-consuming and expensive. Computational methods have streamlined research to determine DTIs. The application of deep learning methods has significantly improved the prediction performance for DTIs. Modern deep learning methods can leverage multiple sources of information, including sequence data that contains biological structural information, and interaction data. While useful, these methods cannot be effectively applied to each type of information individually (e.g., chemical structure and interaction network) and do not take into account the specificity of DTI data such as low- or zero-interaction biological entities. To overcome these limitations, we propose a method called MFA-DTI (Multi-feature Fusion Adopted framework for DTI). MFA-DTI consists of three modules: an interaction graph learning module that processes the interaction network to generate interaction vectors, a chemical structure learning module that extracts features from the chemical structure, and a fusion module that combines these features for the final prediction. To validate the performance of MFA-DTI, we conducted experiments on six public datasets under different settings. The results indicate that the proposed method is highly effective in various settings and outperforms state-of-the-art methods.

Chemoenzymatic Installation of Site-Specific Chemical Groups on DNA Enhances the Catalytic Activity.

Functional DNAs are valuable molecular tools in chemical biology and analytical chemistry but suffer from low activities due to their limited chemical functionalities. Here, we present a chemoenzymatic method for site-specific installation of diverse functional groups on DNA, and showcase the application of this method to enhance the catalytic activity of a DNA catalyst. Through chemoenzymatic introduction of distinct chemical groups, such as hydroxyl, carboxyl, and benzyl, at specific positions, we achieve significant enhancements in the catalytic activity of the RNA-cleaving deoxyribozyme 10-23. A single carboxyl modification results in a 100-fold increase, while dual modifications (carboxyl and benzyl) yield an approximately 700-fold increase in activity when an RNA cleavage reaction is catalyzed on a DNA-RNA chimeric substrate. The resulting dually modified DNA catalyst, CaBn, exhibits a kobs of 3.76 min-1 in the presence of 1 mM Mg2+ and can be employed for fluorescent imaging of intracellular magnesium ions. Molecular dynamics simulations reveal the superior capability of CaBn to recruit magnesium ions to metal-ion-binding site 2 and adopt a catalytically competent conformation. Our work provides a broadly accessible strategy for DNA functionalization with diverse chemical modifications, and CaBn offers a highly active DNA catalyst with immense potential in chemistry and biotechnology.

Dynamic changes of peripheral inflammatory markers link with disease severity and predict short-term poor outcome of myasthenia gravis.

The relationship between peripheral inflammatory markers, their dynamic changes, and the disease severity of myasthenia gravis (MG) is still not fully understood. Besides, the possibility of using it to predict the short-term poor outcome of MG patients have not been demonstrated. This study aims to investigate the relationship between peripheral inflammatory markers and their dynamic changes with Myasthenia Gravis Foundation of America (MGFA) classification (primary outcome) and predict the short-term poor outcome (secondary outcome) in MG patients. The study retrospectively enrolled 154 MG patients from June 2016 to December 2021. The logistic regression was used to investigate the relationship of inflammatory markers with MGFA classification and determine the factors for model construction presented in a nomogram. Finally, net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were utilized to evaluate the incremental capacity. Logistic regression revealed significant associations between neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), aggregate index of systemic inflammation (AISI) and MGFA classification (p = 0.013, p = 0.032, p = 0.017, respectively). Incorporating dynamic changes of inflammatory markers into multivariable models improved their discriminatory capacity of disease severity, with significant improvements observed for NLR, systemic immune-inflammation index (SII) and AISI in NRI and IDI. Additionally, AISI was statistically associated with short-term poor outcome and a prediction model incorporating dynamic changes of inflammatory markers was constructed with the area under curve (AUC) of 0.953, presented in a nomograph. The inflammatory markers demonstrate significant associations with disease severity and AISI could be regarded as a possible and easily available predictive biomarker for short-term poor outcome in MG patients.

Dual Function of Naphthalenediimide Supramolecular Photocatalyst with Giant Internal Electric Field for Efficient Hydrogen and Oxygen Evolution.

Organic supramolecular photocatalysts have garnered widespread attention due to their adjustable structure and exceptional photocatalytic activity. Herein, a novel bis-dicarboxyphenyl-substituent naphthalenediimide self-assembly supramolecular photocatalyst (SA-NDI-BCOOH) with efficient dual-functional photocatalytic performance is successfully constructed. The large molecular dipole moment and short-range ordered stacking structure of SA-NDI-BCOOH synergistically create a giant internal electric field (IEF), resulting in a remarkable 6.7-fold increase in its charge separation efficiency. Additionally, the tetracarboxylic structure of SA-NDI-BCOOH greatly enhances its hydrophilicity. Thus, SA-NDI-BCOOH demonstrates efficient dual-functional activity for photocatalytic hydrogen and oxygen evolution, with rates of 372.8 and 3.8 µmol h-1, respectively. Meanwhile, a notable apparent quantum efficiency of 10.86% at 400 nm for hydrogen evolution is achieved, prominently surpassing many reported supramolecular photocatalysts. More importantly, with the help of dual co-catalysts, it exhibits photocatalytic overall water splitting activity with H2 and O2 evolution rates of 3.2 and 1.6 µmol h-1. Briefly, this work sheds light on enhancing the IEF by controlling the molecular polarity and stacking structure to dramatically improve the photocatalytic performance of supramolecular materials.

Artificial intelligence-driven prediction of multiple drug interactions.

When a drug is administered to exert its efficacy, it will encounter multiple barriers and go through multiple interactions. Predicting the drug-related multiple interactions is critical for drug development and safety monitoring because it provides foundations for practical, safe compatibility and rational use of multiple drugs. With the progress of artificial intelligence (AI) technology, a variety of novel prediction methods for single interaction have emerged and shown great advantages compared to the traditional, expensive and time-consuming laboratory research. To promote the comprehensive and simultaneous predictions of multiple interactions, we systematically reviewed the application of AI in drug-drug, drug-food (excipients) and drug-microbiome interactions. We began by outlining the model methods, evaluation indicators, algorithms and databases commonly used to build models for three types of drug interactions. The models based on the metabolic enzyme P450, drug similarity and drug targets have empathized among the machine learning models of drug-drug interactions. In particular, we discussed the limitations of current approaches and identified potential areas for future research. It is anticipated the in-depth review will be helpful for the development of the next-generation of systematic prediction models for simultaneous multiple interactions.

Discovery unbalanced DNA mixtures and evaluation mixing ratio via a droplet digital PCR method.

BACKGROUND: Small amounts of DNA from a perpetrator collected during crime-scene investigations can be masked by large amounts of DNA from the victim. These samples can provide important information for the perpetrator's conviction. Short tandem repeat (STR) detection system is not sensitive enough to detect trace amounts of minor components in unbalanced mixed DNA. We developed a system using droplet digital polymerase chain reaction (ddPCR) capable of discovering trace components and accurately determining the ratio of mixed DNA in extremely unbalanced mixtures. METHODS: The non-recombining regions of the X chromosome and Y chromosome were quantified in the DNA of male and female mixtures using duplex ddPCR. Absolute quantification of low-abundance portions of trace samples and unbalanced mixtures was done using different mixing ratios. RESULTS: The ddPCR system could be used to detect low-abundance samples with < 5 copies of DNA components in an extremely unbalanced mixture at a mixing ratio of 10000:1. The high sensitivity and specificity of the system could identify the mixing ratio of mixed DNA accurately. CONCLUSIONS: A ddPCR system was developed for evaluation of mixed samples of male DNA and female DNA. Our system could detect DNA quantities as low as 5 copies in extremely unbalanced mixed samples with good specificity and applicability. This method could assist forensic investigators in avoiding the omission of important physical evidence, and evaluating the ratio of mixed male/female trace samples.

Internal validation of the GA118-24B Genetic Analyzer, a stable capillary electrophoresis system for forensic DNA identification.

The GA118-24B Genetic Analyzer (hereafter, "GA118-24B") is an independently developed capillary electrophoresis instrument. In the present research, we designed a series of validation experiments to test its performance at detecting DNA fragments compared to the Applied Biosystems 3500 Genetic Analyzer (hereafter, "3500"). Three commercially available autosomal short tandem repeat multiplex kits were used in this validation. The results showed that GA118-24B had acceptable spectral calibration for three kits. The results of accuracy and concordance studies were also satisfactory. GA118-24B showed excellent precision, with a standard deviation of less than 0.1 bp. Sensitivity and mixture studies indicated that GA118-24B could detect low-template DNA and complex mixtures as well as the results generated by 3500 in parallel experiments. Based on the experimental results, we set specific analytical and stochastic thresholds. Besides, GA118-24B showed superiority than 3500 within certain size ranges in the resolution study. Instead of conventional commercial multiplex kits, GA118-24B performed stably on a self-developed eight-dye multiplex system, which were not performed on 3500 Genetic Analyzer. We compared our validation results with those of previous research and found our results to be convincing. Overall, we conclude that GA118-24B is a stable and reliable genetic analyzer for forensic DNA identification.

Validation of the PowerPlex®35GY System: a novel eight-dye STR multiplex kit on the Spectrum Compact CE System.

The PowerPlex® 35GY System (Promega, USA) is an advanced eight-dye multiplex STR kit, incorporating twenty-three autosomal STR loci, eleven Y chromosome STR loci, one sex determining marker Amelogenin, and two quality indicators. This multiplex system includes 20 CODIS loci and up to 15 mini-STR loci with sizing values less than 250 bases. In this study, validation for PowerPlex® 35GY System was conducted following the guidelines of SWGDAM, encompassing sensitivity, precision, accuracy, concordance, species specificity, stutter, mixture, stability, and degraded DNA. The results from experiments demonstrated that the PowerPlex® 35GY System could effectively amplify DNA samples, with complete allele detection achieved at 125 pg. Moreover, over 90% of alleles from minor contributors were detected at a mixed ratio of 1:4. Additionally, the system was found to yield full profiles even in the presence of hematin, humic acid, and indigo. The PowerPlex® 35GY System demonstrated superior performance in the sensitivity and degraded DNA studies compared to a six-dye STR kit. Hence, it is evident that the PowerPlex® 35GY System is well-suited for forensic practice, whether in casework or for database samples. These findings provide strong support for the efficacy and reliability of the PowerPlex® 35GY System in forensic applications.

LOXL1 promotes tumor cell malignancy and restricts CD8 + T cell infiltration in colorectal cancer.

BACKGROUND: Colorectal cancer (CRC) is a leading cause of cancer mortality globally. Lymph node metastasis and immunosuppression are main factors of poor prognosis in CRC patients. Lysyl oxidase like 1 (LOXL1), part of the lysyl oxidase (LOX) family, plays a yet unclear role in CRC. This study aimed to identify effective biomarkers predictive of prognosis and efficacy of immunotherapy in CRC patients, and to elucidate the prognostic value, clinical relevance, functional and molecular features, and immunotherapy predictive role of LOXL1 in CRC and pan-cancer. METHODS: Weighted gene co-expression network analysis (WGCNA) was employed to explore gene modules related to tumor metastasis and CD8 + T cell infiltration. LOXL1 emerged as a hub gene through differential gene expression and survival analysis. The molecular signatures, functional roles, and immunological characteristics affected by LOXL1 were analyzed in multiple CRC cohorts, cell lines and clinical specimens. Additionally, LOXL1's potential as an immunotherapy response indicator was assessed, along with its role in pan-cancer. RESULTS: Turquoise module in WGCNA analysis was identified as the hub module associated with lymph node metastasis and CD8 + T cell infiltration. Aberrant elevated LOXL1 expression was observed in CRC and correlated with poorer differentiation status and prognosis. Molecular and immunological characterization found that LOXL1 might mediate epithelial-mesenchymal transition (EMT) process and immunosuppressive phenotypes of CRC. Functional study found that LOXL1 enhanced tumor cell proliferation, migration and invasion. Moreover, high LOXL1 levels corresponded to reduced CD8 + T cell infiltration and predicted poor clinical outcomes of immunotherapy. Similar trends were also observed at the pan-cancer level. CONCLUSIONS: Our findings underscore the critical role of LOXL1 in modulating both malignancy and immunosuppression in CRC. This positions LOXL1 as a promising biomarker for predicting prognosis and the response to immunotherapy in CRC patients.

DrugSK: A Stacked Ensemble Learning Framework for Predicting Drug Combinations of Multiple Diseases.

Combination therapy is an important direction of continuous exploration in the field of medicine, with the core goals of improving treatment efficacy, reducing adverse reactions, and optimizing clinical outcomes. Machine learning technology holds great promise in improving the prediction of drug synergy combinations. However, most studies focus on single disease-oriented collaborative predictive models or involve excessive feature categories, making it challenging to predict the majority of new drugs. To address these challenges, the DrugSK comprehensive model was developed, which utilizes SMILES-BERT to extract structural information from 3492 drugs and trains on reactions from 48,756 drug combinations. DrugSK is an integrated learning model capable of predicting interactions among various drug categories. First, the primary learner is trained from the initial data set. Random forest, support vector machine, and XGboost model are selected as primary learners and logistic regression as secondary learners. A new data set is then "generated" to train level 2 learners, which can be thought of as a prediction for each model. Finally, the results are filtered using logistic regression. Furthermore, the combination of the new antibacterial drug Drafloxacin with other antibacterial agents was tested. The synergistic effect of Drafloxacin and Isavuconazonium in the fight against Candida albicans has been confirmed, providing enlightenment for the clinical treatment of skin infection. DrugSK's prediction is accurate in practical application and can also predict the probability of the outcome. In addition, the tendency of Drafloxacin and antifungal drugs to be synergistic was found. The development of DrugSK will provide a new blueprint for predicting drug combination synergies.

DNN-PNN: A parallel deep neural network model to improve anticancer drug sensitivity.

With the rapid development of deep learning techniques and large-scale genomics database, it is of great potential to apply deep learning to the prediction task of anticancer drug sensitivity, which can effectively improve the identification efficiency and accuracy of therapeutic biomarkers. In this study, we propose a parallel deep learning framework DNN-PNN, which integrates rich and heterogeneous information from gene expression and pharmaceutical chemical structure data. With the proposal of DNN-PNN, a new and more effective drug data representation strategy is introduced, that is, the correlation between features is represented by product, which alleviates the limitations of high-dimensional discrete data in deep learning. Furthermore, the framework is optimized to reduce the time complexity of the model. We conducted extensive experiments on the CCLE datasets to compare DNN-PNN with its variant DNN-FM representing the traditional feature correlation model, the component DNN or PNN alone, and the common machine learning models. It is found that DNN-PNN not only has high prediction accuracy, but also has significant advantages in stability and convergence speed.