FDA approves Bayer's Hyrnuo for first-line HER2-mutant lung cancer
The FDA has granted first-line approval to Bayer's tyrosine kinase inhibitor Hyrnuo for patients with HER2-mutant non-small cell lung cancer.
News
Publications
Benchmark reveals conservative bias and cell-type limits across perturbation models
A systematic evaluation of 13 computational methods across 25 perturbation datasets found that current models struggle to generalize cellular drug and gene responses across different cell types.
Machine-learning tool reconstructs cell-fate changes from single-cell snapshots
DiffusionOT applies neural networks and optimal transport to model how cells differentiate over time by accounting for random variations in gene expression.
Multi-modal model predicts patient-specific cancer drug responses
UniCure links biological and chemical foundation models to forecast drug-induced gene expression changes, aiming to bridge the gap between cancer cell lines and clinical responses.
Graph learning on membrane proximity atlases identifies EGFR-CDCP1 for multispecific targeting
A proximity-guided graph learning approach mapped tumor-associated cell-surface protein communities to uncover targetable membrane proximity antigens.
Tree-aware DNA language model outperforms classical evolutionary methods in variant effect prediction
Researchers introduced GPN-Star, a genomic foundation model that integrates whole-genome alignments and evolutionary species trees across vertebrate, mammal, and primate lineages.
Virtual cell model maps dynamic protein shifts across 38 million measurements to predict drug response
Scientists developed ProteinTalks, an operational virtual cell model pretrained on more than 38 million time-resolved protein abundance measurements from perturbed breast cancer cell lines.
Generative path sampling maps rare protein transitions without predefined variables
A generative committor-guided path-sampling framework reconstructs rare structural transitions in biomolecules without requiring researchers to define collective reaction variables in advance.
Screening 1,700 AI-designed protein binders pinpoints key failure modes in CAR T cells
Bioengineers tested 1,758 de novo protein binders designed with generative AI against tumor targets BCMA, CD19, and CD22 to evaluate their real-world function as chimeric antigen receptors (CARs).
Preprints
Image-to-gene models fail to predict spatial transcriptomics in human skin
An independent benchmark of 16 computational models found that methods designed to predict single-cell gene expression directly from standard tissue microscopy images perform poorly across skin disease samples.
Geometric algorithm matches deep learning on collagen structures at a fraction of the cost
Researchers built the Collagen-specific Deterministic Structure Modeler (CDSM), an empirical geometric modeling pipeline that predicts all-atom collagen triple-helical structures without large neural networks.