The front page of AI-driven biology
Thursday, October 15, 2026Thu, Oct 15
Updated 6:32 PM PT
Newly published peer-reviewed research across journals and publishers, with the original publication date and what was actually tested.
Researchers introduced GenAI-Net, a generative framework that automates the design of chemical reaction networks by pairing reaction proposal with simulation against user-specified dynamical goals.
Researchers combined high-throughput yeast display screening with AlphaFold3 structural modeling to discover single-domain antibodies that neutralize neurotoxins from multiple snake species.
A new live-cell imaging workflow combines single-shot stimulated emission depletion microscopy with a neural network to track lipid movement across the endoplasmic reticulum, lipid droplets, and mitochondria simultaneously.
Researchers introduced GeoEPred, a geometric deep learning model that identifies effector proteins secreted by Gram-negative bacteria by fusing linear sequence embeddings with predicted three-dimensional structures.
Google researchers introduced SynthIDBio, a set of methods that embeds detectable digital watermarks directly into AI-generated protein sequences and molecular structures.
Researchers have developed CytoVI, a probabilistic deep generative model that integrates single-cell protein measurements across flow cytometry, mass cytometry, and CITE-seq.
Researchers have built CROWN, a foundation model designed to analyze cell samples across diverse organs, diseases, and clinical settings.
Scientists built Generative Attribution Metric Analysis (GAMA), an interpretability framework based on Integrated Gradients for autoregressive biological sequence generators.
A hybrid framework pairing population-state reconstruction with sparse identification of nonlinear dynamics (SINDy) mapped unobserved cell states in Clostridium co-cultures.
Researchers built an active machine learning system called AGENT that navigated formulation space in six rounds of testing to produce solid-state mRNA lipid nanoparticles that survive storage at 37 °C for over two months.
Researchers built SpaCEy, an explainable graph neural network that analyzes spatial proteomics data to forecast disease progression and overall survival without requiring prior cell-type labels.
Researchers applied rapid computational screening to forecast how existing Ebola Zaire antibody therapies perform against the 2026 Bundibugyo virus outbreak strain.
A hybrid computational framework called DEFT couples protein language models with 3D structural alignment to predict precise four-digit Enzyme Commission functional numbers.
Researchers trained a deep learning model to classify antibiotic mechanisms of action directly from high-throughput microscopy images of Corynebacterium glutamicum, a surrogate for Mycobacterium tuberculosis.
Researchers developed a machine-learning classifier that analyzes label-free optical microscopy images of surgical tissue margins to predict where glioblastoma tumors will recur.
Researchers used the generative protein-design model RFdiffusion alongside affinity maturation to create artificial proteins that selectively bind unanchored ubiquitin, an elusive signaling molecule lacking high-affinity probes.
Researchers combined cryo-electron tomography with computational structure prediction and integrative modeling to map a complete bacterial protein-translocation complex inside Mycoplasma pneumoniae cells at sub-nanometer resolution.
Researchers have developed UpTCR, a progressive transfer-learning model designed to predict whether T-cell receptors will bind specific peptide targets presented by human leukocyte antigen molecules.
Researchers designed DePass, a graph learning framework that combines paired single-cell and spatial multi-omics data using an iterative denoising process.
A new mathematical framework recasts spectral information theory into an interpretable metric called Leff, which calculates the effective number of fully independent positions that capture sequence variation.
Researchers applied the protein design algorithms PROSS and FuncLib to redesign human T cell receptor constant domains, creating a variant called structurally enhanced TCR that boosts surface expression and prevents faulty pairing.
Scientists built iBrainMap, an open-source computational framework based on knowledge-guided graph neural networks to model personalized gene regulatory networks from single-nucleus RNA sequencing data.
Investigators built predictive machine learning models to infer cell-type-specific gene expression from single-nucleus data generated in the multi-ancestry PsychAD cohort.
Researchers guided large language models to suggest specific chemical modifications that make computationally generated molecules easier to synthesize in the laboratory.
A computational pipeline builds personalized protein-protein interaction networks for glioblastoma patients by integrating patient-specific target profiles, dependency data, and network controllability analysis.
Researchers computationally designed synthetic transmembrane peptides that bind directly to the apolar, membrane-spanning domain of Toll-like receptor 4 to block inflammatory signaling.
Researchers combined the generative models ESM3, ProteinMPNN, and EvoDiff to design 76 functional carrier domains for non-ribosomal peptide synthetases, the giant multi-part enzymes that bacteria use to assemble complex peptide drugs.
Researchers have developed NanoTS, a deep learning tool designed to identify single nucleotide polymorphisms, or single-letter DNA variations, directly from nanopore RNA sequencing data.
RetroChimera combines models with complementary inductive biases through a learned ensembling method to predict chemical synthesis pathways, outperforming leading retrosynthesis baselines across varying training data sizes.
LIV2TRANS uses machine learning to map transcriptional data from microphysiological tissue models onto human clinical profiles to identify media supplements that maximize biological relevance.