Biological Validation Research Articles

AutoXAI4Omics: an automated explainable AI tool for omics and tabular data.

Machine learning (ML) methods offer opportunities for gaining insights into the intricate workings of complex biological systems, and their applications are increasingly prominent in the analysis of omics data to facilitate tasks, such as the identification of novel biomarkers and predictive modeling of phenotypes. For scientists and domain experts, leveraging user-friendly ML pipelines can be incredibly valuable, enabling them to run sophisticated, robust, and interpretable models without requiring in-depth expertise in coding or algorithmic optimization. By streamlining the process of model development and training, researchers can devote their time and energies to the critical tasks of biological interpretation and validation, thereby maximizing the scientific impact of ML-driven insights. Here, we present an entirely automated open-source explainable AI tool, AutoXAI4Omics, that performs classification and regression tasks from omics and tabular numerical data. AutoXAI4Omics accelerates scientific discovery by automating processes and decisions made by AI experts, e.g. selection of the best feature set, hyper-tuning of different ML algorithms and selection of the best ML model for a specific task and dataset. Prior to ML analysis AutoXAI4Omics incorporates feature filtering options that are tailored to specific omic data types. Moreover, the insights into the predictions that are provided by the tool through explainability analysis highlight associations between omic feature values and the targets under investigation, e.g. predicted phenotypes, facilitating the identification of novel actionable insights. AutoXAI4Omics is available at: https://github.com/IBM/AutoXAI4Omics.

Biological validation of peak-width of skeletonized mean diffusivity as a VCID biomarker: The MarkVCID Consortium.

Peak-width of skeletonized mean diffusivity (PSMD), a neuroimaging marker of cerebral small vessel disease (SVD), has shown excellent instrumental properties. Here, we extend our work to perform a biological validation of PSMD. We included 396 participants from the Biomarkers for Vascular Contributions to Cognitive Impairment and Dementia (MarkVCID-1) Consortium and three replication samples (Cohorts for Heart and Aging Research in Genomic Epidemiology =6172, Rush University Medical Center =287, University of California Davis Alzheimer's Disease Research Center =567). PSMD was derived from diffusion tensor imaging using an automated algorithm. We related PSMD to a composite measure of general cognitive function using linear regression models adjusting for confounders. Higher PSMD was associated with lower general cognition in MarkVCID-1 independent of age, sex, education, and intracranial volume (Beta [95% confidence interval], -0.8 [-1.2, -0.4], P<0.001). These findings were replicated in independent samples. Furthermore, PSMD explained cognitive status above and beyond white matter hyperintensities. Our biological validation work supports the pursuit of larger clinical validation studies evaluating PSMD as a susceptibility/risk biomarker of small vessel disease contributing to cognitive impairment and dementia. Peak-width of skeletonized mean diffusivity (PSMD) is a novel small vessel disease neuroimaging biomarker. A prior instrumental validation study demonstrated that PSMD is a robust biomarker. This biological validation study shows that high PSMD relates to worse cognition. PSMD explains cognitive function above and beyond white matter hyperintensities. Future clinical validation will assess PSMD as a vascular contribution to cognitive impairment and dementia biomarker in clinical trials.

Expression of Network Medicine-Predicted Genes in Human Macrophages Infected with Leishmania major.

Leishmania spp. commonly infects phagocytic cells of the immune system, particularly macrophages, employing various immune evasion strategies that enable their survival by altering the intracellular environment. In mammals, these parasites establish persistent infections by modulating gene expression in macrophages, thus interfering with immune signaling and response pathways, ultimately creating a favorable environment for the parasite's survival and reproduction. In this study, our objective was to use data mining and subsequent filtering techniques to identify the genes that play a crucial role in the infection process of Leishmania spp. We aimed to pinpoint genes that have the potential to influence the progression of Leishmania infection. To achieve this, we exploited prior, curated knowledge from major databases and constructed 16 datasets of human molecular information consisting of coding genes and corresponding proteins. We obtained over 400 proteins, identifying approximately 200 genes. The proteins coded by these genes were subsequently used to build a network of protein-protein interactions, which enabled the identification of key players; we named this set Predicted Genes. Then, we selected approximately 10% of Predicted Genes for biological validation. THP-1 cells, a line of human macrophages, were infected with Leishmania major in vitro for the validation process. We observed that L. major has the capacity to impact crucial genes involved in the immune response, resulting in macrophage inactivation and creating a conducive environment for the survival of Leishmania parasites.

DNA methylation classifier to diagnose pancreatic ductal adenocarcinoma metastases from different anatomical sites

BackgroundWe have recently constructed a DNA methylation classifier that can discriminate between pancreatic ductal adenocarcinoma (PAAD) liver metastasis and intrahepatic cholangiocarcinoma (iCCA) with high accuracy (PAAD-iCCA-Classifier). PAAD is one of the leading causes of cancer of unknown primary and diagnosis is based on exclusion of other malignancies. Therefore, our focus was to investigate whether the PAAD-iCCA-Classifier can be used to diagnose PAAD metastases from other sites.MethodsFor this scope, the anomaly detection filter of the initial classifier was expanded by 8 additional mimicker carcinomas, amounting to a total of 10 carcinomas in the negative class. We validated the updated version of the classifier on a validation set, which consisted of a biological cohort (n = 3579) and a technical one (n = 15). We then assessed the performance of the classifier on a test set, which included a positive control cohort of 16 PAAD metastases from various sites and a cohort of 124 negative control samples consisting of 96 breast cancer metastases from 18 anatomical sites and 28 carcinoma metastases to the brain.ResultsThe updated PAAD-iCCA-Classifier achieved 98.21% accuracy on the biological validation samples, and on the technical validation ones it reached 100%. The classifier also correctly identified 15/16 (93.75%) metastases of the positive control as PAAD, and on the negative control, it correctly classified 122/124 samples (98.39%) for a 97.85% overall accuracy on the test set. We used this DNA methylation dataset to explore the organotropism of PAAD metastases and observed that PAAD liver metastases are distinct from PAAD peritoneal carcinomatosis and primary PAAD, and are characterized by specific copy number alterations and hypomethylation of enhancers involved in epithelial-mesenchymal-transition.ConclusionsThe updated PAAD-iCCA-Classifier (available at https://classifier.tgc-research.de/) can accurately classify PAAD samples from various metastatic sites and it can serve as a diagnostic aid.

Network pharmacology, molecular docking and biological verification to explore the potential anti-prostate cancer mechanisms of Tripterygium wilfordii Hook. F

Ethnopharmacological relevanceTripterygium wilfordii Hook. f. (TW) is extensively utilized in clinical practice for its effective anti-inflammatory and anti-cancer properties. Aim of the studyThis study aims to elucidate the processes of TW in combating prostate cancer through a comprehensive strategy that integrates network pharmacology, molecular docking and molecular biology validation. Materials and methodsA drug-target network and protein-protein interaction network were constructed established to predict the potential targets of TW for prostate cancer treatment. The interaction between active components and targets was confirmed using molecular docking. Moreover, prostate cancer cells were used to examine the anti-tumor effects of active ingredients in vitro. The xenograft animal model was constructed to evaluate the anti-tumor effect of triptonoterpene in vivo. ResultsTwenty-nine active components interact with 226 corresponding targets, and 112 disease targets specifically related with prostate cancer were identified. The primary targets (AKT1, TP53, RELA) were chosen, and kaempferol, triptolide, and triptonoterpene exhibited probable binding affinity with these targets, respectively. Triptonoterpene was subsequently confirmed to inhibit the growth of prostate cancer cells and induce apoptosis in vitro and in vivo. ConclusionOverall, this study demonstrated that TW may serve as a viable therapeutic agent for prostate cancer. Triptonoterpene is a specific inhibitor of p-AKT1 and p65, making it an attractive contender for prostate cancer therapy.

The Significance of Artificial Intelligence and Machine Learning in the Identification of Immunotherapy Targets for Cancer: Advances, Challenges, and Future Directions

Cancer immunotherapy has revolutionized cancer treatment by leveraging the immune system to target malignant cells, yet resistance in many cancers highlights the need for novel therapeutic targets. Artificial intelligence (AI) and machine learning (ML) have emerged as transformative tools for identifying new immunotherapy targets by analyzing vast datasets from genomics, proteomics, and clinical studies. This review explores the role of AI and ML in advancing the discovery of cancer-specific immunotherapy targets, such as tumor antigens and immune pathways. Key advances include the integration of big data, neoantigen prediction, biomarker discovery, and single-cell analysis. Despite these advancements, challenges remain, including data quality and standardization, interpretability of AI models, computational costs, and the need for biological validation of AI-driven discoveries. As AI and ML technologies continue to evolve, they hold the potential to overcome these barriers, leading to personalized immunotherapy solutions. This review also discusses future directions for AI-driven immunotherapy, emphasizing the need for improved models, ethical considerations, and clinical integration. Keywords: Artificial Intelligence, Machine Learning, Immunotherapy, Cancer, Advances, Challenges, Future Directions

Bioorthogonal Monomycolate of Trehalose Disclosed the O-Mycoloylation of Mycoloyltransferases and Other Cell Envelope Proteins in C. glutamicum.

Protein mycoloylation is a recently identified unusual post-translational modification (PTM) exclusively observed in Mycobacteriales, an order of bacteria that includes several human pathogens. These bacteria possess a distinctive outer membrane, known as the mycomembrane, composed of very long-chain fatty acids called mycolic acids. It has been demonstrated that a few mycomembrane proteins undergo covalent modification with mycolic acids in the model organism Corynebacterium glutamicum through the action of mycoloyltransferase MytC. This PTM represents the first example of protein O-acylation in prokaryotes and also the first example of protein modification by mycolic acid. Many questions about the specificity of protein O-mycoloylation remain crucial for understanding its evolutionary significance in Mycobacteriales and its role in cell physiology. We have developed the first bioorthogonal mycolate donor featuring the natural mycolic acid pattern, enabling direct, unambiguous transfer of the lipid moiety to its acceptors and efficient metabolic labeling and enrichment of MytC protein substrates. Mass spectrometry analysis of the labeled proteins and comparative proteomic analysis of the cell envelope proteome between wild-type and ΔmytC strains identified an unbiased list of 21 proteins likely mycoloylated in the cell. The robustness of our approach is demonstrated by the successful biological validation of mycoloylation in 6 candidate proteins within wild-type cells, revealing the characteristic profile of proteins modified with natural mycolates. These findings provide interesting insights into the significance of this new lipidation pathway and pave the way for understanding their function, especially concerning the mycoloyltransferase family that includes the essential Antigen85 enzymes in Mycobacteria.

Deficiency of DDX3X results in neurogenesis defects and abnormal behaviors via dysfunction of the Notch signaling

The molecular mechanisms underlying the neurodevelopmental disorders (NDDs) caused by DDX3X variants remain poorly understood. In this study, we validated that de novo DDX3X variants are enriched in female developmental delay (DD) patients and mainly affect the evolutionarily conserved amino acids based on a meta-analysis of 46,612 NDD trios. We generated a ddx3x deficient zebrafish allele, which exhibited reduced survival rate, DD, microcephaly, adaptation defects, anxiolytic behaviors, social interaction deficits, and impaired spatial recognitive memory. As revealed by single-nucleus RNA sequencing and biological validations, ddx3x deficiency leads to reduced neural stem cell pool, decreased total neuron number, and imbalanced differentiation of excitatory and inhibitory neurons, which are responsible for the behavioral defects. Indeed, the supplementation of L-glutamate or glutamate receptor agonist ly404039 could partly rescue the adaptation and social deficits. Mechanistically, we reveal that the ddx3x deficiency attenuates the stability of the crebbp mRNA, which in turn causes downregulation of Notch signaling and defects in neurogenesis. Our study sheds light on the molecular pathology underlying the abnormal neurodevelopment and behavior of NDD patients with DDX3X mutations, as well as providing potential therapeutic targets for the precision treatment.

Discovery and Total Synthesis of Anhydrotuberosin as a STING Antagonist for Treating Autoimmune Diseases

Excessive activation of the stimulator of the interferon gene (STING) pathway has been identified as a significant contributor to various autoimmune diseases, such as STING‐associated vasculopathy with infantile‐onset (SAVI) and inflammatory bowel disease (IBD). However, discovering effective STING antagonists for treating STING‐mediated autoimmune disorders remains challenging. Herein, we identified the natural product anhydrotuberosin (ATS) as a potent STING antagonist by a high‐throughput chemical screen and follow‐up biological validations. However, the limited supply from natural product isolation impeded the pharmacological evaluations of ATS. Accordingly, we developed a concise and scalable total synthesis of ATS in 6 steps. Enabled by total synthesis, we further extensively investigated ATS’s mode of action and evaluated its therapeutic potential. Remarkably, ATS inhibits STING signaling in PBMCs derived from three SAVI patients. ATS showed decent pharmacokinetic parameters and strongly alleviated tissue inflammation in DSS‐induced IBD colitis and Trex1‐/‐ autoimmune animal models with low toxicity. Collectively, this research lays the foundation for developing novel STING antagonists as an effective therapy for autoinflammatory and autoimmune diseases.

Discovery and Total Synthesis of Anhydrotuberosin as a STING Antagonist for Treating Autoimmune Diseases.

Excessive activation of the stimulator of the interferon gene (STING) pathway has been identified as a significant contributor to various autoimmune diseases, such as STING-associated vasculopathy with infantile-onset (SAVI) and inflammatory bowel disease (IBD). However, discovering effective STING antagonists for treating STING-mediated autoimmune disorders remains challenging. Herein, we identified the natural product anhydrotuberosin (ATS) as a potent STING antagonist by a high-throughput chemical screen and follow-up biological validations. However, the limited supply from natural product isolation impeded the pharmacological evaluations of ATS. Accordingly, we developed a concise and scalable total synthesis of ATS in 6 steps. Enabled by total synthesis, we further extensively investigated ATS's mode of action and evaluated its therapeutic potential. Remarkably, ATS inhibits STING signaling in PBMCs derived from three SAVI patients. ATS showed decent pharmacokinetic parameters and strongly alleviated tissue inflammation in DSS-induced IBD colitis and Trex1-/- autoimmune animal models with low toxicity. Collectively, this research lays the foundation for developing novel STING antagonists as an effective therapy for autoinflammatory and autoimmune diseases.

First-in-Class Small Molecule Degrader of Pregnane X Receptor Enhances Chemotherapy Efficacy.

Pregnane X receptor (PXR) is a ligand-activated transcription factor that binds diverse compounds and upregulates drug metabolism machinery in response. PXR activation is detrimental to drug efficacy and safety because it reduces active drug concentrations and increases reactive metabolites, leading to toxicity and/or drug-drug interactions. Thus, effort must be expended in drug development pipelines to assess PXR activation by lead candidates and chemically modify agonists to reduce PXR liabilities while maintaining on-target potencies. Coadministration of drugs with PXR antagonists could prevent PXR-mediated metabolism events, but such compounds are rare and may themselves be converted to agonists by metabolic enzymes or PXR mutations. Here, we report the design, synthesis, optimization, and biological validation of proteolysis targeting chimeras that induce PXR degradation through E3 ubiquitin ligase recruitment. PXR degradation blocks agonist-induced gene expression and enhances anticancer effects of the chemotherapy paclitaxel, a known PXR agonist and substrate of downstream metabolic enzymes.

Guided structure learning of DAGs for count data

In this paper, we tackle structure learning of directed acyclic graphs (DAGs), with the idea of exploiting available prior knowledge of the domain at hand to guide the search of the best structure. In particular, we assume to know the topological ordering of variables in addition to the given data. We study a new algorithm for learning the structure of DAGs, proving its theoretical consistency in the limit of infinite observations. Furthermore, we experimentally compare the proposed algorithm to several popular competitors, to study its behaviour in finite samples. Biological validation of the algorithm is presented through the analysis of non-small cell lung cancer data.

Deep convolutional neural network-based identification and biological evaluation of MAO-B inhibitors

Parkinson's disease (PD) is one of the most prominent motor disorder of adult-onset dementia connected to memory and other cognitive abilities. Individuals with this vicious neurodegenerative condition tend to have an elevated expression of Monoamine Oxidase-B (MAO-B) that catalyzes the oxidative deamination of aryalkylamines neurotransmitters with concomitant reduction of oxygen to hydrogen peroxide. This oxidative stress damages mitochondrial DNA and contributes to the progression of PD. To address this, we have developed a deep learning (DL)-based virtual screening protocol for the identification of promising MAO-B inhibitors using Convolutional neural network (ConvNet) based image classification technique by dealing with two unique kinds of image datasets associated with MACCS fingerprints. Following model building and prediction on the Maybridge library, our approach shortlisted the top 11 compounds at the end of molecular docking protocol. Further, the biological validation of the hits identified 4 compounds as promising MAO-B inhibitors. Among these, the compound RF02426 was found to have >50 % inhibition at 10 μM. Additionally, the study also underscored the utility of scaffold analysis as an effective way for evaluating the significance of structurally diverse compounds in data-driven investigations. We believe that our models are able to pick up diverse chemotype and this can be a starting scaffold for further structural optimization with medicinal chemistry efforts in order to improve their inhibition efficacy and be established as novel MAO-B inhibitors in the furture.

Multiarm, non-randomised, single-centre feasibility study—investigation of the differential biology between benign and malignant renal masses using advanced magnetic resonance imaging techniques (IBM-Renal): protocol

IntroductionLocalised renal masses are an increasing burden on healthcare due to the rising number of cases. However, conventional imaging cannot reliably distinguish between benign and malignant renal masses, and renal...

Validation of a Commercial ELISA Kit for Non-Invasive Measurement of Biologically Relevant Changes in Equine Cortisol Concentrations

The measurement of fecal cortisol/corticosterone metabolites (FCMs) is often used to quantify the stress response. The sampling method is relatively non-invasive, reduces concern for elevation of cortisol from the sampling method, and has been shown to measure cortisol more consistently without the daily diurnal rhythm observed in blood. Commercial ELISA (enzyme-linked immunoassay) kits offer benefits over previously validated immunoassay methods but lack validation. The objective of this study was to evaluate a commercial ELISA kit (Arbor AssaysTM DetectX® Cortisol ELISA kit, K003-H1, Ann Arbor, MI, USA) and provide analytical and biologic validation of equine fecal and plasma samples. Horses (4 male, 4 female, mean ± SD: 4 ± 5 yr) were transported for 15 min with limited physical and visual contact via a livestock trailer. Blood and fecal samples were collected pre- and post-transportation. Parallelism, accuracy, and precision tests were used to analytically validate this kit. Data were analyzed using PROC MIXED in SAS 9.4. Plasma cortisol concentrations increased in response to trailering (254.5 ± 26.4 nmol/L, 0 min post-transportation) compared to pre-transportation (142.8 ± 26.4 nmol/L). FCM concentrations increased 24 h post-trailering (10.8 ± 1.7 ng/g) when compared to pre-transportation (7.4 ± 1.7 ng/g). These data support that changes in FCMs can be observed 24 h post-stressor. In conclusion, the Arbor AssaysTM DetectX® Cortisol ELISA kit is a reliable, economic option for the measurement of biologically relevant changes in cortisol in equine plasma and FCMs.

Machine learning and biological validation identify sphingolipids as potential mediators of paclitaxel-induced neuropathy in cancer patients

Background:Chemotherapy-induced peripheral neuropathy (CIPN) is a serious therapy-limiting side effect of commonly used anticancer drugs. Previous studies suggest that lipids may play a role in CIPN. Therefore, the present study aimed to identify the particular types of lipids that are regulated as a consequence of paclitaxel administration and may be associated with the occurrence of post-therapeutic neuropathy.Methods:High-resolution mass spectrometry lipidomics was applied to quantify d=255 different lipid mediators in the blood of n=31 patients drawn before and after paclitaxel therapy for breast cancer treatment. A variety of supervised statistical and machine-learning methods was applied to identify lipids that were regulated during paclitaxel therapy or differed among patients with and without post-therapeutic neuropathy.Results:Twenty-seven lipids were identified that carried relevant information to train machine learning algorithms to identify, in new cases, whether a blood sample was drawn before or after paclitaxel therapy with a median balanced accuracy of up to 90%. One of the top hits, sphinganine-1-phosphate (SA1P), was found to induce calcium transients in sensory neurons via the transient receptor potential vanilloid 1 (TRPV1) channel and sphingosine-1-phosphate receptors.SA1P also showed different blood concentrations between patients with and without neuropathy.Conclusions:Present findings suggest a role for sphinganine-1-phosphate in paclitaxel-induced biological changes associated with neuropathic side effects. The identified SA1P, through its receptors, may provide a potential drug target for co-therapy with paclitaxel to reduce one of its major and therapy-limiting side effects.Funding:This work was supported by the Deutsche Forschungsgemeinschaft (German Research Foundation, DFG, Grants SFB1039 A09 and Z01) and by the Fraunhofer Foundation Project: Neuropathic Pain as well as the Fraunhofer Cluster of Excellence for Immune-Mediated Diseases (CIMD). This work was also supported by the Leistungszentrum Innovative Therapeutics (TheraNova) funded by the Fraunhofer Society and the Hessian Ministry of Science and Arts. Jörn Lötsch was supported by the Deutsche Forschungsgemeinschaft (DFG LO 612/16-1).

Machine learning and biological validation identify sphingolipids as potential mediators of paclitaxel-induced neuropathy in cancer patients.

Chemotherapy-induced peripheral neuropathy (CIPN) is a serious therapy-limiting side effect of commonly used anticancer drugs. Previous studies suggest that lipids may play a role in CIPN. Therefore, the present study aimed to identify the particular types of lipids that are regulated as a consequence of paclitaxel administration and may be associated with the occurrence of post-therapeutic neuropathy. High-resolution mass spectrometry lipidomics was applied to quantify d=255 different lipid mediators in the blood of n=31 patients drawn before and after paclitaxel therapy for breast cancer treatment. A variety of supervised statistical and machine-learning methods was applied to identify lipids that were regulated during paclitaxel therapy or differed among patients with and without post-therapeutic neuropathy. Twenty-seven lipids were identified that carried relevant information to train machine learning algorithms to identify, in new cases, whether a blood sample was drawn before or after paclitaxel therapy with a median balanced accuracy of up to 90%. One of the top hits, sphinganine-1-phosphate (SA1P), was found to induce calcium transients in sensory neurons via the transient receptor potential vanilloid 1 (TRPV1) channel and sphingosine-1-phosphate receptors.SA1P also showed different blood concentrations between patients with and without neuropathy. Present findings suggest a role for sphinganine-1-phosphate in paclitaxel-induced biological changes associated with neuropathic side effects. The identified SA1P, through its receptors, may provide a potential drug target for co-therapy with paclitaxel to reduce one of its major and therapy-limiting side effects. This work was supported by the Deutsche Forschungsgemeinschaft (German Research Foundation, DFG, Grants SFB1039 A09 and Z01) and by the Fraunhofer Foundation Project: Neuropathic Pain as well as the Fraunhofer Cluster of Excellence for Immune-Mediated Diseases (CIMD). This work was also supported by the Leistungszentrum Innovative Therapeutics (TheraNova) funded by the Fraunhofer Society and the Hessian Ministry of Science and Arts. Jörn Lötsch was supported by the Deutsche Forschungsgemeinschaft (DFG LO 612/16-1).

G-Protein Signaling in Alzheimer's Disease: Spatial Expression Validation of Semi-supervised Deep Learning-Based Computational Framework.

Systemic study of pathogenic pathways and interrelationships underlying genes associated with Alzheimer's disease (AD) facilitates the identification of new targets for effective treatments. Recently available large-scale multiomics datasets provide opportunities to use computational approaches for such studies. Here, we devised a novel disease gene identification (digID) computational framework that consists of a semi-supervised deep learning classifier to predict AD-associated genes and a protein-protein interaction (PPI) network-based analysis to prioritize the importance of these predicted genes in AD. digID predicted 1,529 AD-associated genes and revealed potentially new AD molecular mechanisms and therapeutic targets including GNAI1 and GNB1, two G-protein subunits that regulate cell signaling, and KNG1, an upstream modulator of CDC42 small G-protein signaling and mediator of inflammation and candidate coregulator of amyloid precursor protein (APP). Analysis of mRNA expression validated their dysregulation in AD brains but further revealed the significant spatial patterns in different brain regions as well as among different subregions of the frontal cortex and hippocampi. Super-resolution STochastic Optical Reconstruction Microscopy (STORM) further demonstrated their subcellular colocalization and molecular interactions with APP in a transgenic mouse model of both sexes with AD-like mutations. These studies support the predictions made by digID while highlighting the importance of concurrent biological validation of computationally identified gene clusters as potential new AD therapeutic targets.

BioPathNet: Enhancing Link Prediction in Biomedical Knowledge Graphs through Path Representation Learning.

Understanding complex interactions in biomedical networks is crucial for advancements in biomedicine, but traditional link prediction (LP) methods are limited in capturing this complexity. Representation-based learning techniques improve prediction accuracy by mapping nodes to low-dimensional embeddings, yet they often struggle with interpretability and scalability. We present BioPathNet, a novel graph neural network framework based on the Neural Bellman-Ford Network (NBFNet), addressing these limitations through path-based reasoning for LP in biomedical knowledge graphs. Unlike node-embedding frameworks, BioPathNet learns representations between node pairs by considering all relations along paths, enhancing prediction accuracy and interpretability. This allows visualization of influential paths and facilitates biological validation. BioPathNet leverages a background regulatory graph (BRG) for enhanced message passing and uses stringent negative sampling to improve precision. In evaluations across various LP tasks, such as gene function annotation, drug-disease indication, synthetic lethality, and lncRNA-mRNA interaction prediction, BioPathNet consistently outperformed shallow node embedding methods, relational graph neural networks and task-specific state-of-the-art methods, demonstrating robust performance and versatility. Our study predicts novel drug indications for diseases like acute lymphoblastic leukemia (ALL) and Alzheimer's, validated by medical experts and clinical trials. We also identified new synthetic lethality gene pairs and regulatory interactions involving lncRNAs and target genes, confirmed through literature reviews. BioPathNet's interpretability will enable researchers to trace prediction paths and gain molecular insights, making it a valuable tool for drug discovery, personalized medicine and biology in general.

MHIPM: Accurate Prediction of Microbe-Host Interactions Using Multiview Features from a Heterogeneous Microbial Network.

Current studies have demonstrated that microbe-host interactions (MHIs) play important roles in human public health. Therefore, identifying the interactions between microbes and hosts is beneficial to understanding the role of the microbiome and their underlying mechanisms. However, traditional wet-lab experimental approaches are insufficient for large-scale exploration of candidate microbes, as they are costly, laborious, and time-consuming. Thus, it is critical to prioritize microbe-interacting hosts by computational approaches for further biological experimental validation. In this work, we proposed a novel deep learning-based method called MHIPM, to predict MHIs by utilizing multisource biological information. Specifically, we first constructed a heterogeneous microbial network that consisted of human proteins, viruses, bacteriophages (phages), and pathogenic bacteria. Next, we used one of the largest protein language models, ESM-2, and a document embedding model, doc2vec, combined with a self-attention mechanism to extract the interview features from protein sequences. Then, an inductive learning-based model, GraphSAGE, was used to capture the intraview features from the heterogeneous network. Experimental results on three prediction tasks indicated that the MHIPM model consistently achieved better performance than seven baseline algorithms and its four variants. In addition, case studies and molecular docking experiments for two human proteins further confirmed the effectiveness of our model. In conclusion, MHIPM is an efficient and robust method in predicting MHIs and provides plausible candidate microbes for biological experiments. MHIPM is available at https://github.com/JIENWU/MHIPM.