Methodology

neoag.ai is a research instrument for following the neoantigen cancer vaccine field. It is designed for readers who need a fast view of what changed, why it matters, and where the original source lives.

Source coverage

The collector layer favors public, citable sources over private rumor. Current source families include:

• ClinicalTrials.gov: trial registrations, status, phase, sponsor, indication, and update signals.
• arXiv: AI, machine-learning, immunoinformatics, and computational-methods preprints.
• Europe PMC: indexed papers and preprints relevant to neoantigen vaccines, TCR/pMHC modeling, and delivery.
• bioRxiv and medRxiv: recent preprints filtered for field relevance.
• Google News RSS: commercial, regulatory, deal, financing, and readout signals, with publisher links resolved when possible.

Filtering and ranking

Recall starts broad, then precision is added in stages. The pipeline first applies keyword and source-specific filters, then scores candidate items for field relevance. Items are tagged as clinical, commercial, or AI/methods so the daily brief does not collapse everything into one undifferentiated feed.

The aim is not to publish every mention of a keyword. The aim is to surface items that change the reader's view of the field, add useful context, or track a durable program, method, company, or trial.

Role of AI

AI is used to help score relevance, classify items, and synthesize short editorial summaries. It does not invent the source list. The URLs, titles, dates, and source records are stored in the local archive and passed into the generation step.

The site follows a practical rule: no citation, no publish. If the system cannot attach an item back to a public source record, it should not be presented as a sourced development.

Quality controls

• Archive deduplication: repeated items are collapsed where possible, including across source families.
• Quiet-day behavior: if no new item clears the bar, the homepage says that plainly instead of dressing old material as news.
• Inline citations: summary claims point down to the item cards and outbound original sources.
• Static publishing: the site builds to static HTML so readers see a reproducible snapshot rather than a live database query.

Limitations

neoag.ai can miss relevant developments, include low-value items, or summarize a source imperfectly. Public registries and publisher feeds can also lag, change, or contain errors. The site is a field-monitoring tool, not a clinical, scientific, investment, legal, or professional advice service.

For formal disclaimers, read the Terms. For analytics and data handling, read the Privacy Policy.

Where to go next

New to the field? Start with the Primer. Tracking programs? Open the Trials tracker. Want the daily feed? Use RSS or return to today's brief.