Researchers Develop Deep Learning Platform to Identify HLA-E Cancer Antigens for Immunotherapy
Scientists created an integrated platform combining deep learning and high-throughput screening to discover cancer peptides presented by HLA-E, a protein minimally polymorphic across human populations. The approach, called MHC Attention, identified novel HLA-E-presented cancer peptides from genes including ETV4, WT1, RNF43, and BMP8A. This work could enable development of broadly applicable cancer immunotherapies targeting HLA-E, which is widely expressed across cancer types.
Researchers developed an antigen discovery platform to systematically identify cancer-associated peptides presented by HLA-E, a major histocompatibility complex protein that is minimally polymorphic and widely expressed across human populations and cancer types. The platform integrates a novel deep learning prediction model called MHC Attention—a neural network that learns allele-level attention patterns in multi-allele immunopeptidomics datasets—with pooled mammalian cell screening, peptide-HLA-E stability validation, and mass spectrometry. By screening approximately 6,000 peptides in an HLA-E library, the team generated HLA-E-specific training data that improved the algorithm's prediction performance. The integrated approach identified novel HLA-E-presented cancer peptides with orthogonal validation, including candidates derived from ETV4, WT1, RNF43, and BMP8A. The researchers have made MHC Attention 2.0 publicly accessible online, establishing a scalable framework for HLA-E peptide target discovery that could support development of peptide-HLA-directed cancer immunotherapies applicable across diverse patient populations.
What's missing
The study does not discuss clinical validation timelines, potential immunogenicity challenges, or how the identified peptides compare in therapeutic potential to existing HLA-E-targeted approaches. Additionally, the generalizability of the approach to HLA-E variants and the computational requirements for scaling the platform are not detailed.
What different sources said
- bioRxivCenter
MHC Attention: Identifying HLA-E presented cancer antigens through deep learning and high-throughput screening
Related
Study Shows Statins Reduce Coenzyme Q in Brain Cells, Impairing Mitochondrial Function
A laboratory study found that statin drugs decrease coenzyme Q levels in astrocytes (brain support cells) by 30-40%, reducing their mitochondrial energy production and increasing oxidative stress. Astrocytes are critical for maintaining brain health and protecting neurons from damage. The findings suggest CoQ10 supplementation may help counteract these effects, though human clinical evidence remains limited.
Study reveals zebrafish larvae exhibit slowly fluctuating directional swim biases driven by internal dynamics
Researchers found that 5-day-old zebrafish larvae display changing directional swim preferences over many hours even in stable environments, contrary to classical models assuming constant individual biases. Computational analysis suggests these fluctuations arise from a non-stationary Markovian process with two independent internal input streams modulating swim direction repetition. The findings suggest animals possess intrinsic mechanisms for generating behavioral variability independent of external stimuli, with implications for understanding how internal states shape adaptive behavior.
Two Small Molecules Show Promise as Broad-Spectrum Coronavirus Inhibitors in Laboratory Study
Researchers found that sennoside A and ceftazidime, two small molecules, can inhibit RNA binding in the nucleocapsid proteins of SARS-CoV-2, SARS-CoV, and MERS-CoV in laboratory experiments. The nucleocapsid protein is highly conserved across coronaviruses and essential for viral replication, making it a potential drug target. The findings suggest these compounds could form the basis for pan-coronavirus antiviral therapies, though further development and clinical testing would be needed.