Community Call for Open Repertoire Benchmarks: VBASE3 exclusively incorporates authentic, peer-reviewed, open-access datasets. If your laboratory has published single-cell BCR/TCR, longitudinal, or multi-species repertoire datasets (under CC-BY or Public Domain) that you would like showcased as tutorial benchmarks on our analysis pages, we warmly invite you to contribute!
User Repertoire Data IngestionClient-Side SIMD Engine
Provide your own experimental data via drag-and-drop file upload or direct text paste. Supports AIRR Community TSV, 10x Genomics CellRanger CSV, Simple TSV/CSV, or Paired FASTA. Zero network transmission.
๐ Expected Data Format Specifications & Column Requirements:Click any tab below to review the required columns or copy a ready-to-use sample snippet
AIRR Community standard: Tab-delimited file with one record per contig. Rows sharing the same sequence_id or cell barcode are automatically paired across heavy (IGH/TRB) and light (IGK/IGL/TRA) chains.
sequence_id v_call j_call junction_aa duplicate_count locus timepoint target
donor_cell_01_H humIGHV1-69*01 humIGHJ4*01 CARDGYYDILTGYYFDYW 112 IGH Day 0 Healthy Baseline
donor_cell_01_L humIGKV1-39*01 humIGKJ1*01 CQQSYSTPPTF 112 IGK Day 0 Healthy Baseline
donor_cell_02_H humIGHV1-69*01 humIGHJ6*01 CARDKGYSYGYW 94 IGH Day 28 SARS-CoV-2 S2
donor_cell_02_L humIGKV3-20*01 humIGKJ2*01 CQQYGSSPLTF 94 IGK Day 28 SARS-CoV-2 S2
10x Genomics CellRanger: Directly drag-and-drop your filtered_contig_annotations.csv file. Contigs sharing the identical barcode are paired into heterodimers.
Non-productive contigs (stop codons, frameshifts) are automatically audited and filtered.
Simple Paired Table: Exactly one line per single-cell heterodimer. Perfect for copying and pasting columns from Microsoft Excel, Apple Numbers, or Google Sheets.
Can be tab-delimited (TSV) or comma-delimited (CSV). Delimiter is auto-detected.
Paired Multi-FASTA: Standard FASTA format where paired chains are identified either sequentially or via header suffixes (>clone1_heavy, >clone1_light, >clone1_H, >clone1_L, or >clone1_TRB / >clone1_TRA). The engine accepts nucleotide DNA in any reading frame (with automated productive frame and mature FR1 anchor detection) or amino acid protein sequences.
Non-linear manifold projections (PaCMAP/UMAP) optimize relative pairwise distance hierarchies rather than absolute Cartesian lengths. Relative distances preserve local sequence neighborhoods and global phylogeny; numerical values on Dimensions 1 & 2 reflect this continuous topology.
Quantitative tracking of somatic hypermutation rate, clonal lineage expansion, and affinity maturation trajectories during vaccination or immune interventions.
Somatic Hypermutation Accumulation CurveMean SHM vs. Time
Lineage Abundance & Clonal ExpansionClonal Frequency (%)