neoag
Neoantigen × AI
Daily research signal
№38Brief2026-08-07

BioNTech mRNA-4157 shows 49% recurrence risk reduction in melanoma

BioNTech’s mRNA-4157 shows a 49% reduction in recurrence risk when combined with pembrolizumab in high-risk melanoma, reinforcing the clinical viability of personalized neoantigen vaccines.

On the methodological front, new AI frameworks TransNRank and MHChron aim to improve neoantigen prediction accuracy and pMHC binding robustness through transformer architectures and diversity-balanced datasets.arXivbioRxiv

4new stories1 in context
ClinicaltodayNew

Personalized Neoantigen-pulsed Autologous Dendritic Cell Injections for Malignant Solid Tumors

ZSky Biotech Inc has launched a recruiting, single-center study for ZSNeo-DC, a personalized neoantigen-pulsed autologous dendritic cell therapy for malignant solid tumors. The trial will enroll approximately 100 patients across multiple tumor cohorts, administering seven subcutaneous injections either as monotherapy or with immune checkpoint inhibitors. This real-world study aims to evaluate safety, efficacy, and immunogenicity under GMP manufacturing standards.

Dendritic-cell vaccines
clinicaltrials · today · ZSky Biotech Inc

Also new today03

01
Clinical3 days agoNew

Identification of pre-existing ubiquitous neoantigen-reactive tumor-infiltrating T-cells in a patient with metastatic pancreatic neuroendocrine tumor

A preprint investigates whether ubiquitous neoantigens—mutations shared across all metastatic sites—can overcome inter-site heterogeneity in pancreatic neuroendocrine tumors. Using whole-exome and RNA sequencing on 14 samples from a treatment-naive patient, researchers identified ubiquitous mutations and validated immunogenicity via IFN-gamma ELISpot and single-cell TCR sequencing. This approach targets durable immune responses by focusing on neoantigens present in both primary and metastatic lesions.

TCR & T-cell receptorPancreatic cancer
biorxiv · 3 days ago · Tanis, J.-B.; McCann, K.; Castaneda-Castro, F. E.; Thomas, J.; Bailey, A.; Singh, P.; Curr
02
AI / Methods3 days agoNew

MHChron: diversity-balanced dataset design for robust peptide-MHC binding prediction across MHC class I and II

MHChron introduces a unified peptide-MHC binding prediction framework built on a diverse dataset covering 214 class I and 98 class II alleles, with peptides ranging from 8 to 36 residues. The study emphasizes rigorous data curation and leakage-controlled splits to ensure robust generalization. Both sequence-based and structure-aware models trained on this balanced dataset outperformed state-of-the-art predictors, particularly in extrapolating to unseen alleles.

Peptide vaccines
biorxiv · 3 days ago · Chronowska, M.; Shrimpton-Phoenix, E.; Kluonis, T.
03
AI / Methods4 days agoNew

TransNRank: Towards Accurate Neoantigen Ranking with Transformer

TransNRank is a new Transformer-based deep learning framework designed to improve neoantigen ranking by capturing long-range dependencies and handling class imbalance via positive-aware training. Tested on NCI, TESLA, and HiTIDE datasets, it increased the top 20 recall rate from 46.9% to 53.1% while reducing training epochs significantly. The model addresses limitations of prior linear regression and XGBoost methods in modeling complex immunogenicity features.

arxiv · 4 days ago · Zhiyin An, Yuenan Hou, Shumeng Duan, Yiming Zhou, Yuanting Zheng, Leming Shi

Catching up01

Relevant work from earlier we're still surfacing — not today's news.

Explore the field

TopicsTrialsLearn