AIRRDB

A searchable index of the known human paratope space.

What it is

The human B-cell repertoire is, by some measures, the most diverse molecular system on Earth. Searching it efficiently — across billions of sequences, with biologically meaningful tolerance for mutation — is a hard problem. Our flagship database, AIRRDB, solves this: a searchable index of 2.1 billion antibody sequences, served at sub-second latency, exportable in standard AIRR formats.

Researchers query it by CDR-H3 — the most diverse loop of the antibody binding site — and get back hits with controlled mismatch tolerance, as AIRR-TSV or VDJfasta. Average response time per query: under one second. That makes the database usable for interactive workflows — comparing a candidate antibody against the global repertoire while you design it, not in an overnight batch.

Features

Large scale

2.1 billion entries, continuously curated from public sources.

Ultra fast

Sub-second similarity search across the full index, with up to two mismatches per CDR.

Plug-and-play export

Results download as standardised AIRR-TSV or VDJfasta — drop straight into your existing pipeline.

Python & R API

Query the index programmatically from Python or R — integrates directly into your analysis workflow.

Patented core

The N-Hamming search method behind AIRRDB is protected by patent WO 2023/274497.

Use cases

  • Precision antibody engineering — check whether a designed paratope already exists naturally.
  • Vaccine response analysis — compare a study cohort's repertoire against the public space.
  • Therapeutic antibody surveys — search for analogues of a clinical lead across millions of donors.
  • Population-level repertoire studies — quantify rarity, clonality, public-vs-private signatures.

Get access

AIRRDB runs as a standalone service at airrdb.com. Academic and commercial access are available — please reach out via the contact page for credentials and terms.

Visit airrdb.com →