Column reference#

Every table below is built from vdjdb.schema while this page renders. There is no generated file to refresh, so no copy of these tables can disagree with the build.

vdjdb.txt - the release table#

Column

Title

Description

complex.id

complex.id

TCR alpha and beta chain records having the same complex identifier belong to the same T-cell clone.

gene

Gene

TCR chain: alpha or beta.

cdr3

CDR3

TCR complementarity determining region 3 (CDR3) amino acid sequence.

v.segm

V

TCR Variable segment allele.

j.segm

J

TCR Joining segment allele.

species

Species

TCR parent species.

mhc.a

MHC A

First MHC chain allele.

mhc.b

MHC B

Second MHC chain allele (defaults to Beta2Microglobulin for MHC class I).

mhc.class

MHC class

MHC class (I or II).

antigen.epitope

Epitope

Amino acid sequence of the epitope.

antigen.gene

Epitope gene

Representative parent gene of the epitope.

antigen.species

Epitope species

Representative parent species of the epitope.

reference.id

Reference

Pubmed reference / URL / or submitter details in case unpublished.

vdjdb.score

Info

VDJdb confidence score, the higher is the score the more confidence we have in the antigen specificity annotation of a given TCR clonotype/clone. Zero score indicates that there are insufficient method details to draw any conclusion.

TCR_hash

TCR hash

SHA256 hash of TCR structure used for structure visualization.

method

Method

Details on method used to assay TCR specificity.

meta

Meta

Various meta-information: cell subset, donor status, etc.

cdr3fix

CDR3fix

Details on CDR3 sequence fixing (if applied) and consistency between V, J and reported CDR3 sequence.

web.method

Internal

Internal: coarse identification method for fast filtering.

web.method.seq

Internal

Internal: coarse sequencing method for fast filtering.

web.cdr3fix.nc

Internal

Internal: CDR3 has non-canonical V or J anchor residues.

web.cdr3fix.unmp

Internal

Internal: CDR3 could not be mapped onto V or J germline.

vdjdb.txt - the vdjdb-web variant#

The variant served by vdjdb-web adds five evidence.* columns and a TCR_hash row that the release table does not have.

Column

Title

Description

complex.id

complex.id

TCR alpha and beta chain records having the same complex identifier belong to the same T-cell clone.

gene

Gene

TCR chain: alpha or beta.

cdr3

CDR3

TCR complementarity determining region 3 (CDR3) amino acid sequence.

v.segm

V

TCR Variable segment allele.

j.segm

J

TCR Joining segment allele.

species

Species

TCR parent species.

mhc.a

MHC A

First MHC chain allele.

mhc.b

MHC B

Second MHC chain allele (defaults to Beta2Microglobulin for MHC class I).

mhc.class

MHC class

MHC class (I or II).

antigen.epitope

Epitope

Amino acid sequence of the epitope.

antigen.gene

Epitope gene

Representative parent gene of the epitope.

antigen.species

Epitope species

Representative parent species of the epitope.

reference.id

Reference

Pubmed reference / URL / or submitter details in case unpublished.

vdjdb.score

Info

VDJdb confidence score, the higher is the score the more confidence we have in the antigen specificity annotation of a given TCR clonotype/clone. Zero score indicates that there are insufficient method details to draw any conclusion.

TCR_hash

TCR hash

SHA256 hash of TCR structure used for structure visualization.

method

Method

Details on method used to assay TCR specificity.

meta

Meta

Various meta-information: cell subset, donor status, etc.

cdr3fix

CDR3fix

Details on CDR3 sequence fixing (if applied) and consistency between V, J and reported CDR3 sequence.

web.method

Internal

Internal: coarse identification method for fast filtering.

web.method.seq

Internal

Internal: coarse sequencing method for fast filtering.

web.cdr3fix.nc

Internal

Internal: CDR3 has non-canonical V or J anchor residues.

web.cdr3fix.unmp

Internal

Internal: CDR3 could not be mapped onto V or J germline.

evidence.validation.same.study

Validation same study

Antigen specificity validated within the same study.

evidence.validation.independent

Validation independent

Antigen specificity independently validated in another study.

evidence.structure.contacts

Structure model with contacts

Structural model with annotated TCR-pMHC contacts available.

evidence.structure.quality

Structure good quality model

Good-quality structural model available.

evidence.structure.native

Structure native

Native (experimental) TCR-pMHC structure available.

vdjdb.slim.txt#

Column

Title

Description

gene

Gene

TCR chain: alpha or beta.

cdr3

CDR3

TCR complementarity determining region 3 (CDR3) amino acid sequence.

species

Species

TCR parent species.

antigen.epitope

Epitope

Amino acid sequence of the epitope.

antigen.gene

Epitope gene

Representative parent gene of the epitope.

antigen.species

Epitope species

Representative parent species of the epitope.

complex.id

complex.id

TCR alpha and beta chain records having the same complex identifier belong to the same T-cell clone.

v.segm

V

TCR Variable segment allele.

j.segm

J

TCR Joining segment allele.

mhc.a

MHC A

First MHC chain allele.

mhc.b

MHC B

Second MHC chain allele (defaults to Beta2Microglobulin for MHC class I).

mhc.class

MHC class

MHC class (I or II).

reference.id

Reference

Pubmed reference / URL / or submitter details in case unpublished.

vdjdb.score

Info

VDJdb confidence score, the higher is the score the more confidence we have in the antigen specificity annotation of a given TCR clonotype/clone. Zero score indicates that there are insufficient method details to draw any conclusion.

TCR_hash

TCR hash

SHA256 hash of TCR structure used for structure visualization.

j.start

J start

First amino acid position of the J germline part of CDR3, 0-based, junction space.

v.end

V end

Last amino acid position of the V germline part of CDR3, 0-based, junction space.

vdjdb_full.txt#

Column

Title

cdr3.alpha

CDR3 alpha

v.alpha

V alpha

j.alpha

J alpha

cdr3.beta

CDR3 beta

v.beta

V beta

d.beta

D beta

j.beta

J beta

species

Species

mhc.a

MHC A

mhc.b

MHC B

mhc.class

MHC class

antigen.epitope

Epitope

antigen.gene

Epitope gene

antigen.species

Epitope species

reference.id

Reference

method.identification

Identification method

method.frequency

Frequency

method.singlecell

Single cell

method.sequencing

Sequencing

method.verification

Verification

meta.study.id

Study id

meta.cell.subset

Cell subset

meta.subject.cohort

Subject cohort

meta.subject.id

Subject id

meta.replica.id

Replica id

meta.clone.id

Clone id

meta.epitope.id

Epitope id

meta.tissue

Tissue

meta.donor.MHC

Donor MHC

meta.donor.MHC.method

Donor MHC method

meta.structure.id

Structure id

cdr3fix.alpha

CDR3fix alpha

cdr3fix.beta

CDR3fix beta

vdjdb.score

Info

TCR_hash

TCR hash

The definitive tables#

records and chains are the database. Every shipped file is a join and a pivot away from them.

records#

Column

Title

Description

record_id

Record id

Stable VDJdb record identifier. Assigned once, never reused, and it survives a content change – a curator fixing a typo amends a record rather than deleting one and creating another.

pmhc_id

pMHC id

Identifies a presented peptide: PM plus 16 hex digits of a sha256 over the epitope and the two curated MHC chains. Keys on the allele as curated, never on a predicted one, so a new mhcmatch model cannot renumber it.

epitope_id

Epitope id

Identifies the peptide alone, across every allele presenting it: EP plus 16 hex digits of a sha256 over antigen.epitope.

species

Species

TCR parent species.

mhc_a

MHC A

First MHC chain allele.

mhc_b

MHC B

Second MHC chain allele (defaults to Beta2Microglobulin for MHC class I).

mhc_class

MHC class

MHC class (I or II).

antigen_epitope

Epitope

Amino acid sequence of the epitope.

antigen_gene

Epitope gene

Representative parent gene of the epitope.

antigen_species

Epitope species

Representative parent species of the epitope.

reference_id

Reference

Pubmed reference / URL / or submitter details in case unpublished.

meta_study_id

Study id

meta_cell_subset

Cell subset

meta_subject_cohort

Subject cohort

meta_subject_id

Subject id

meta_replica_id

Replica id

meta_clone_id

Clone id

meta_tissue

Tissue

meta_epitope_id

Epitope id

meta_donor_MHC

Donor MHC

meta_donor_MHC_method

Donor MHC method

meta_structure_id

Structure id

meta_subset_frequency

Subset frequency

Kept for debugging; dropped by the legacy build.

method_identification

Identification method

method_frequency

Frequency

Frequency of this clonotype in the isolated epitope-reactive population, as the submitter wrote it: `x/X`, `X%` or a bare float. Free text on purpose - it is what every release has shipped, and some studies report only the ratio.

method_singlecell

Single cell

method_sequencing

Sequencing

method_verification

Verification

method_frequency_count

Frequency count

Reads, UMIs or cells supporting this clonotype, parsed from an unambiguous `x/X` `method.frequency`. Null where the submitted value is a float or a percentage: the count is absent, not zero.

method_frequency_total

Frequency total

The sample total `method.frequency.count` is out of, from the same `x/X`.

method_pairing

Pairing

How alpha and beta chains were paired. Kept for debugging; dropped by the legacy build.

vdjdb_score

Info

VDJdb confidence score, the higher is the score the more confidence we have in the antigen specificity annotation of a given TCR clonotype/clone. Zero score indicates that there are insufficient method details to draw any conclusion.

chunk_file

Chunk file

The chunk the record was read from. One chunk is one publication.

chunk_row

Chunk row

0-based row within the chunk; with chunk.file it points at the curated line.

chunk_id

Chunk id

Kept for debugging; dropped by the legacy build.

submitter

Submitter

Kept for debugging; dropped by the legacy build.

comment

Comment

Kept for debugging; dropped by the legacy build.

chains#

Column

Title

Description

record_id

Record id

Stable VDJdb record identifier. Assigned once, never reused, and it survives a content change – a curator fixing a typo amends a record rather than deleting one and creating another.

gene

Gene

TCR chain: alpha or beta.

clonotype_id

Clonotype id

Identifies a receptor chain: CT plus 16 hex digits of a sha256 over species, gene, CDR3, V and J. Records reporting the same chain share it, and motif evidence attaches at this level. Derived from its own key, so no chunk being added or removed can change it.

clone_id

Clone id

Identifies an alpha/beta pair: CX plus 16 hex digits of a sha256 over the record’s two sorted clonotype ids. Empty string on a record reporting one chain, which is the curated state and not missing data.

cdr3

CDR3

TCR complementarity determining region 3 (CDR3) amino acid sequence.

v_segm

V

TCR Variable segment allele.

d_segm

D

TCR Diversity segment allele.

j_segm

J

TCR Joining segment allele.

v_end

V end

Last amino acid position of the V germline part of CDR3, 0-based, junction space.

j_start

J start

First amino acid position of the J germline part of CDR3, 0-based, junction space.

cdr3nt

CDR3 nucleotide

Most plausible nucleotide junction behind the amino-acid one, inferred from the recombination model – not observed. Two models agree on only 7.2 % of these, so it is a representative history, never evidence.

cdr3nt_pgen

CDR3nt Pgen

Generation probability of the inferred nucleotide junction.

cdr3nt_margin

CDR3nt margin

How far the inferred junction beat the runner-up: its Pgen divided by the next candidate’s. Near 1 means the choice among synonymous histories was near-arbitrary.

v_inferred

V inferred

V call proposed by the recombination model, filled only where the curator named none. Recovers the curated V on 23.8 % of human TRB and 50.1 % of TRA when it is hidden – the junction carries little V. Never overwrites v.segm.

j_inferred

J inferred

J call proposed by the recombination model, filled only where the curator named none. Recovers the curated J on 97.5 % of human TRB and 95.8 % of TRA.

d_inferred

D inferred

D allele of the recombination scenario that produced cdr3nt. Not the curated d.segm, which it matches at gene level on 76.5 % of beta chains.

d_start

D start

First nucleotide of the D segment in cdr3nt, 0-based, half-open with d.end.

d_end

D end

One past the last nucleotide of the D segment in cdr3nt.

d_posterior

D posterior

Posterior probability of the gene d.inferred names, from the same recombination scenario weights that named it – so the number is the probability of that call. Naming the D is 74.35 % correct at gene level on human TRB against nucleotide truth; filter on this rather than reading d.inferred alone.

v_end_inferred

V end inferred

V/J boundary from a germline alignment over the junction, in v.end’s space. A fallback, never an override: filled only where the markup engine declined and -1 everywhere else, so coalescing it with v.end cannot overwrite an alignment answer. Against observed nucleotide boundaries it is exact on 92.9 % of junctions where v.end is 71.8 %, so where v.end is -1 this carries a better answer than nothing.

j_start_inferred

J start inferred

As v.end.inferred, for j.start.

cdr3_original

CDR3 as submitted

The CDR3 as the reference publication reported it, before repair.

fix_needed

Fix needed

Whether the repaired CDR3 differs from the submitted one.

fix_good

Fix good

Whether the CDR3 could be placed on both germline segments.

v_fix_type

V fix type

How the V side was repaired: NoFixNeeded, FixAdd, FixTrim, FixReplace, TruncatedGermline (the boundary is real but a lower bound, because IMGT ships that allele’s germline record stopping inside the anchor region), or a Failed* reason.

j_fix_type

J fix type

How the J side was repaired.

v_canonical

V anchor canonical

Whether the CDR3 begins with Cys104, which a TCR junction does by definition. False marks a record kept for what it reports and filterable by anyone who wants only canonical junctions.

j_canonical

J anchor canonical

Whether the CDR3 ends with Phe118 or Trp118, which a TCR junction does by definition. False marks a record kept and filterable, not a variant.

v_canonical_submitted

V anchor canonical as submitted

Whether the submitted CDR3 began with Cys104, before any repair.

j_canonical_submitted

J anchor canonical as submitted

Whether the submitted CDR3 ended with Phe118 or Trp118, before any repair.

cdr3_one_cysteine

Single cysteine

Whether the CDR3 carries no cysteine after the Cys104 it opens with.

v_segm_submitted

V as submitted

The V call as the reference publication reported it, before IMGT harmonisation and allele disambiguation. Compare with v.segm to see what the build decided.

j_segm_submitted

J as submitted

The J call as the reference publication reported it, before harmonisation.

d_segm_submitted

D as submitted

The D call as the reference publication reported it, before harmonisation. Beta only; the alpha locus has no D.

v_segm_arda

V called by the markup engine

The V allele the markup engine aligned the junction against, which is evidence about the sequence rather than about what was reported. It is NOT what ships: v.segm carries the curated call. Empty where the engine named no allele.

j_segm_arda

J called by the markup engine

The J allele the markup engine aligned against. Not what ships; see v.segm.arda.

TCR_hash

TCR hash

SHA256 hash of TCR structure used for structure visualization.

epitopes#

Column

Title

antigen_epitope

Epitope

antigen_species

Epitope species

antigen_gene

Epitope gene

epitope_length

Epitope length

mhc_class

MHC class

records

Records

chains

Chains

references

References

clonotypes

Clonotypes

proteome_peptide

Proteome peptide

proteome_substitution

Substitution

restriction#

Column

Title

antigen_epitope

Epitope

antigen_species

Epitope species

mhc_a

MHC A

mhc_b

MHC B

mhc_class

MHC class

mhc_a_status

MHC A status

mhc_b_status

MHC B status

records

Records

references

References

alleles_reported

Alleles reported

mhc_a_top

Top-scoring allele

mhc_a_rank

Curated allele rank

mhc_a_percentile

Curated allele percentile

promiscuity

Promiscuity

mhcmatch_version

mhcmatch version

evidence#

Column

Title

record_id

Record id

gene

Gene

evidence_id

Evidence id

evidence_type

Evidence type

evidence_source

Evidence source

evidence_value

Evidence value

evidence_score

Evidence score

first_seen_release

First seen release

The motif files#

⚠ These two files are parsed positionally by vdjdb-web, which hands Tablesaw a fixed column-type array with no header check. An inserted, removed or reordered column mistypes or shifts every column after it, without an error. Column order is a contract.

cluster_members.txt#

Column

Title

species

Species

antigen.epitope

Epitope

antigen.gene

Epitope gene

antigen.species

Epitope species

mhc.a

MHC A

mhc.b

MHC B

mhc.class

MHC class

gene

Gene

cdr3aa

CDR3

x

x

y

y

cid

Cluster id

csz

Cluster size

v.segm

V

j.segm

J

v.end

V end

j.start

J start

v.segm.repr

V representative

j.segm.repr

J representative

motif_pwms.txt#

Column

Title

species

Species

antigen.epitope

Epitope

gene

Gene

aa

Residue

pos

Position

len

Length

v.segm.repr

V representative

j.segm.repr

J representative

cid

Cluster id

csz

Cluster size

count

Count

count.bg

Background count

total.bg

Background total

count.bg.i

Imputed background count

total.bg.i

Imputed background total

need.impute

Imputed

freq

Frequency

freq.bg

Background frequency

I

Information

I.norm

Normalised information

height.I

Letter height

height.I.norm

Normalised letter height

antigen.gene

Epitope gene

antigen.species

Epitope species

mhc.a

MHC A

mhc.b

MHC B

mhc.class

MHC class