MARiO is a machine-learning model that predicts the pathogenicity of missense cancer variants by combining variant-level in silico features with patient-level mutational context, addressing the limitation of existing tools that ignore clinical context. This integration aims to improve variant interpretation in comprehe
MediumChembiochem : a European journal of chemical biology2026-11-12
This review examines DNAzyme-based biosensors for detecting single-nucleotide polymorphisms (SNPs), the most common form of human genetic variation and key biomarkers for disease risk and precision medicine. The authors argue that these catalytic DNA sensors offer advantages over conventional genotyping methods by comb
The authors present bronko, an alignment-free framework that detects viral genome variation directly from sequencing reads using locality-sensitive bucketing and k-mer counting, avoiding the computational cost of traditional alignment-based variant calling. This approach enables scaling to massive viral sequencing data
MediumThe Journal of infectious diseases2026-11-10
This study re-examined the capsule polysaccharides of pneumococcal serogroup 15 and found that O-acetylation generates four distinct structural variants of the 15A capsule, not just the previously recognized 15A/15F pair. The work shows that conventional serological and molecular typing can miss O-acetylation differenc
MediumJournal of animal science and biotechnology2026-11-08
This review examines how artificial intelligence tools are being applied to livestock genetics and breeding, noting that while animal breeding has long used data-driven prediction, modern AI approaches remain at an early stage with substantial untapped potential. The authors survey current applications and likely futur
This study systematically evaluates and optimizes a long-read metagenomic workflow for gut microbiome samples, spanning sample processing through computational assembly, to improve recovery of complete microbial genomes. By integrating experimental and computational optimizations, the authors show that long-read approa
This study characterizes the microbiome of an unusual acidic, suboxic submerged 'lake' within the Luymes Sinkholes in the Caribbean, using 59 size-fractionated samples analyzed with multi-omics approaches. The authors resolve a specialized microbial community enriched in nitrification and halogen cycling potential, sho
A large-scale field survey across China's rice-growing regions, combined with 13CH4-DNA stable isotope probing, reveals a positive association between aerobic methane oxidation and microbial mercury methylation in paddy soils. This suggests that methanotrophs contribute to the production of neurotoxic methylmercury tha
This prospective birth cohort study in Switzerland characterized antibiotic exposure during the first two years of life, identifying patterns and determinants of prescription. The findings highlight how early-life antibiotic use varies by factors such as season, childcare attendance, and parental characteristics, provi
This paper describes the protocol for a prospective birth cohort in Vancouver, Canada, designed to investigate how the age at which infants are first exposed to respiratory syncytial virus (RSV) affects their immune response and long-term respiratory health. The study aims to clarify whether early RSV infection leads t
Researchers reconstructed draft genomes of Desulfobulbus oralis, an oral bacterium, from ancient dental calculus (calcified plaque) of Pacific Islanders using metagenomic methods. The genomes are highly similar to a modern reference strain, showing the bacterium's long presence in human oral microbiomes. This demonstra
PEONY is an updated global reference database of DNA viral sequences (vOTUs, mostly dsDNA bacteriophages) derived exclusively from virus size-fractionated metagenomes (viromes) across global ecosystems. The update rebrands the former Phages and Integrated Genomes Encapsidated or Not database to reflect its stricter inc
This review examines why computational methods for discovering subclass B1 metallo-β-lactamase inhibitors (targeting NDM-1, VIM-2, IMP-1) often fail to translate into clinical candidates, despite guiding much of current inhibitor discovery. It catalogs progress and pitfalls in computational prediction and highlights op
This study uses atomistic molecular dynamics simulations to investigate how a naphthalene-diimide organic semiconductor crystal transitions between two polymorphs (α at room temperature, γ at higher temperatures). The simulations reproduce both stable phases and refine unit-cell parameters and side-chain conformations
MediumThe journal of physical chemistry. B2026-11-12
This study uses molecular dynamics simulations to compute potential of mean force (PMF) profiles that predict how diverse PFAS molecules partition into phospholipid bilayers. By integrating interaction energies along defined pathways, the approach offers a physics-based way to estimate membrane partitioning for PFAS wi
The authors used AlphaFold protein structure prediction combined with FoldSeek structural similarity search to build a whole-proteome structural atlas of the syphilis pathogen Treponema pallidum Nichols strain. This approach allowed them to assign likely functions to previously uncharacterized proteins, particularly th
MediumThe journal of physical chemistry letters2026-11-12
This study investigates how two structurally similar antenna subunits of photosystem II, CP29 and LHCII, achieve functional specialization despite their similarity. By combining 2D electronic-vibrational spectroscopy with kinetic modeling, the authors assign spectral features to specific exciton states and show that mi
This paper develops a machine learning framework that iteratively refines molecular dynamics simulations to predict how PFAS coatings on nylon repel water and oil. By systematically varying PFAS head group chemistries, the approach aims to explain how molecular structure and surface organization control wettability, su
The authors present scstruc, an R package for learning Bayesian gene regulatory networks from single-cell transcriptomic data and, importantly, for evaluating how accurately the learned network structures capture causal gene-gene relationships. This addresses a gap in available software for both Bayesian network infere
COAST is a new computational method that physically aligns consecutive tissue sections from spatial multi-omics experiments using only their associated images, without requiring shared molecular features or prior annotations. It was validated on spatial transcriptomics slides across different technologies, tissues, and
This study uses single-nucleus RNA sequencing to build a transcriptomic atlas of the Arabidopsis stem and identify the endodermal gene regulatory network controlling gravitropism. It focuses on the transcription factor SGR5/IDD15, which is essential for shoot gravitropism, and reveals how gravity perception in endoderm
PIANO is a deep learning framework that uses a generative model of gene expression to integrate single-cell multi-omics data, addressing strong batch effects and scaling to 100 million cells. It also models how covariates influence expression, enabling more accurate comparisons across large transcriptomic atlases.
This paper proposes "Longevitomics" as a unified systems-science framework for studying healthy longevity, arguing that its determinants are currently fragmented across separate molecular layers (genomics, epigenomics, transcriptomics, proteomics, metabolomics, microbiome, digital phenotyping, AI). It defines the "Long
Using transcriptome-integrated genome-scale metabolic modeling, the authors show that host-like temperature (37°C) enables pathogenic Leptospira to take up and metabolize glucose, contrary to the long-held view that this bacterium relies primarily on long-chain fatty acids. This temperature-dependent metabolic shift ma
This study applied machine learning to plasma proteomic data to find proteins that distinguish atopic dermatitis from ulcerative colitis, two inflammatory diseases that both involve aberrant Th2 immune activity but affect different tissues. The analysis identified candidate blood-based biomarkers that differentiate the
This multicenter registry study developed a deep learning model to predict mortality among critically ill COVID-19 patients admitted to the ICU, using a deliberately minimal set of readily available clinical variables to enable rapid risk stratification. The work addresses a gap in AI-based mortality prediction, which
MARiO is a machine-learning model that predicts the pathogenicity of missense cancer variants by combining variant-level features with patient-level mutational context from comprehensive genomic profiling. Unlike most existing tools that rely only on variant-level information, MARiO incorporates clinical context to imp
This study develops and internally validates a deep-learning model that analyzes peripheral blood smear images to predict acute inflammatory states and chronic comorbidities. The work suggests that morphological features of immune cells captured on routine blood smears contain systemic pathophysiological information be
This study presents a single-cell multi-omic atlas of histone modifications and transcriptomes across eight human basal ganglia regions, revealing cell-type- and region-specific regulatory elements and gene regulatory networks. The authors identify combinatorial homeobox transcription factor expression as a key driver
This study integrates single-cell and bulk multi-omics data to dissect malignant epithelial heterogeneity in head and neck squamous cell carcinoma, identifying C16orf74 as a novel regulator of epithelial-mesenchymal transition (EMT). The authors show that C16orf74 expression is associated with EMT-associated malignant
Single-Cell & Spatial Omics
MediumbioRxiv : the preprint server for biology2026-11-12
The authors present RAIDS, a software toolkit that infers genetic ancestry from human molecular data, including sequence data that are difficult to genotype. Its data-synthesis approach lets it assess and optimize ancestry inference accuracy for each individual sample without needing cohort data.
SigBridgeR is an R toolkit that standardizes the identification and benchmarking of phenotype-associated cell subpopulations by integrating bulk transcriptomic data with single-cell RNA-seq data. It addresses the current fragmentation in the field—where methods differ in algorithms, parameters, and output formats—by pr
gcPathogen 2.0 is a major update to a pathogen genomic surveillance database, expanding coverage from 1,134 to 1,506 pathogenic species across bacteria, fungi, viruses, and parasites, with over 2.3 million genomic records. The update adds integrated transcriptomic profiles, systematic tracking of cross-habitat distribu
Met-SC is a new integrated metabolomics database and analysis platform that supports standardized data deposition, spectral resource management, online analysis, and disease-focused knowledge mining. It captures detailed metadata across chromatography, CE-MS, and NMR workflows and contains over 7.7 million reference sp
The authors developed a ligand-based virtual screening workflow combining Thompson Sampling, a pharmacophore model, and a machine learning classifier to search 51 million virtual compounds from Enamine's REAL database for new Cryptosporidium parvum inhibitors. The approach identified novel candidate compounds, addressi
This review traces the historical evolution of QSAR (quantitative structure-activity relationships) from Hansch analysis to modern machine learning and AI-driven drug discovery. It highlights how changes in molecular representation led to an explosion of descriptors and how growing data volumes necessitated advanced st
This review traces the evolution of AI in early-stage drug discovery, from deep-learning generative models for designing small molecules and biologics to emerging agentic AI systems that can autonomously plan and execute multi-step tasks. It highlights that while state-of-the-art generative models exist, their behavior
Pepti-DrugPhAI is a multimodal AI tool that predicts pharmacokinetic properties and toxicity for peptide drug candidates, addressing a major bottleneck in peptide therapeutic development. It combines multiple data modalities to evaluate functional PK and safety profiles, aiming to reduce late-stage failures of promisin
The authors synthesized a new organic-inorganic copper(II) hybrid halometalate, (C6H8BrN2)2[CuBr4], and characterized it using X-ray crystallography, optical spectroscopy, DFT calculations, molecular docking, and biological assays. The crystal structure features discrete square-planar [CuBr4]2- anions held together by
The authors synthesized a novel indenonaphthodiazepine[5.3.3]propellane compound via a one-pot multicomponent reaction and characterized it spectroscopically. DFT calculations probed its electronic properties, and molecular docking against the anti-apoptotic protein Mcl-1 gave a binding score of -5.678 kcal/mol, sugges