Restore missing residues in antibody heavy or light chains using sequence language models. Learn more
Configure inputs to begin
Set options on the left, then click “Submit job” — or start from an example.
Heavy chain with missing residues
Light chain with missing residues
Restore missing residues in antibody heavy or light chains using sequence language models. Learn more
Configure inputs to begin
Set options on the left, then click “Submit job” — or start from an example.
Heavy chain with missing residues
Light chain with missing residues
AbLang is an antibody-specific language model developed at the Oxford Protein Informatics Group. It restores missing residues in antibody sequences, a common problem in B-cell receptor repertoire sequencing where over 40% of sequences in the Observed Antibody Space (OAS) database are missing their first 15 N-terminal amino acids.
The model uses a RoBERTa transformer architecture trained exclusively on antibody sequences from OAS. This specialization allows AbLang to outperform general protein language models like ESM-1b while matching the accuracy of IMGT germline-based restoration, without requiring any germline knowledge.
AbLang learns antibody-specific patterns through masked language modeling. During training, 1% to 25% of residues in each sequence are masked, and the model predicts the original amino acids from context. This training approach captures the statistical regularities of antibody sequences, enabling accurate restoration of missing positions.
Two separate models were trained:
Each model consists of two components: AbRep generates 768-dimensional embeddings from sequence context, and AbHead predicts amino acid likelihoods at each position. For restoration, the amino acid with highest likelihood at each masked position is selected.
On N-terminal restoration (first 15 positions):
| Chain | AbLang accuracy | IMGT germline accuracy | ESM-1b accuracy |
|---|---|---|---|
| Heavy | ~98% | ~98% | 64% |
| Light | ~96% | ~96% | 54% |
AbLang is also 7× faster than ESM-1b, processing 100 sequences in about 6.5 seconds versus 45 seconds.
Paste or upload uppercase antibody sequences, mark known missing residues with *, and select the heavy-chain or light-chain model. AbLang returns each submitted sequence beside its restored form, highlights replaced masks, preserves FASTA record names, and provides restored FASTA and detailed JSON files for download or downstream workflows.
| Field | Description |
|---|---|
Antibody sequences | FASTA or plain sequence text using uppercase standard amino-acid codes. Mark known missing residues with *. Each sequence can contain up to 158 residues; multi-sequence input should use FASTA. |
Missing residues can appear anywhere in the sequence, though the most common use case is N-terminal restoration. For example:
>heavy_chain_example
EVQLVESGGGLVQP**SLRLSCAASGFTF**SYAMSWVRQAPGKGLEWVSAI| Setting | Default | Description |
|---|---|---|
Chain type | Heavy chain | Selects the original heavy-chain or light-chain AbLang model. Heavy chains typically begin with EVQ or QVQ; light chains with DIQ or EIV. |
Restore terminal residues | Off | Uses IMGT numbering to infer missing residues at sequence ends before restoration. Enable this when the number of missing terminal residues is unknown; the restored sequence length may change. |
Results display every submitted sequence beside its restored version. Replacements for explicit * masks are highlighted when positions remain directly comparable.
| Output | Description |
|---|---|
| Restored sequence view | Preserved FASTA record name, submitted sequence, restored sequence, chain model, residue counts, and identity. |
ablang_restored_sequences.fasta | Primary restored FASTA output. Record names are preserved and the file can feed another sequence workflow step. |
ablang_restoration.json | Detailed records, summary values, settings context, and source and model provenance. |
AbLang works best for sequences that resemble those in its training data, antibody variable domains from the OAS database. Performance may degrade on highly unusual antibodies or non-human sequences not well-represented in OAS.
Terminal restoration depends on successful IMGT numbering. Antibody-like variable domains are the intended input; sequences that cannot be recognized by the numbering model may fail when Restore terminal residues is enabled.
Geometric deep learning model for predicting protein binding sites directly from 3D structure. Identifies where proteins interact with other proteins, antibodies, or disordered proteins with high accuracy, including for novel protein folds.
Predict protein aggregation nucleation propensity from amino acid sequences using the Lehner Lab CANYA neural network.
Isoelectric Point Calculator 2.0 - Predict protein/peptide isoelectric point (pI) using 18+ validated pKa scales, SVR models, and deep learning. Supports proteins, peptides, and comprehensive analysis.
Find open reading frames in DNA with the native NCBI ORFfinder search, translations, coordinates, and output files.
Predict protein solubility from amino acid sequence using the University of Manchester Protein-Sol method.
Predict metal and water binding sites in protein structures using 3D convolutional neural networks (AllMetal3D + Water3D).
Score protein mutations with evolutionary profiles from homologous sequences and inverse folding. EvoIF returns a dimensionless log-odds score for each submitted single or multi-site mutation.
Predict multiple protein developability properties from amino-acid sequences using a multitask ProstT5 adapter.
Predict protein thermostability changes (ΔΔG) for point mutations using a graph neural network. Enables computational saturation mutagenesis screening to identify stabilizing mutations.
Faithful static-mode Aggrescan3D tool for per-residue aggregation propensity analysis from a single protein structure.
AbLang is an antibody-specific language model developed at the Oxford Protein Informatics Group. It restores missing residues in antibody sequences, a common problem in B-cell receptor repertoire sequencing where over 40% of sequences in the Observed Antibody Space (OAS) database are missing their first 15 N-terminal amino acids.
The model uses a RoBERTa transformer architecture trained exclusively on antibody sequences from OAS. This specialization allows AbLang to outperform general protein language models like ESM-1b while matching the accuracy of IMGT germline-based restoration, without requiring any germline knowledge.
AbLang learns antibody-specific patterns through masked language modeling. During training, 1% to 25% of residues in each sequence are masked, and the model predicts the original amino acids from context. This training approach captures the statistical regularities of antibody sequences, enabling accurate restoration of missing positions.
Two separate models were trained:
Each model consists of two components: AbRep generates 768-dimensional embeddings from sequence context, and AbHead predicts amino acid likelihoods at each position. For restoration, the amino acid with highest likelihood at each masked position is selected.
On N-terminal restoration (first 15 positions):
| Chain | AbLang accuracy | IMGT germline accuracy | ESM-1b accuracy |
|---|---|---|---|
| Heavy | ~98% | ~98% | 64% |
| Light | ~96% | ~96% | 54% |
AbLang is also 7× faster than ESM-1b, processing 100 sequences in about 6.5 seconds versus 45 seconds.
Paste or upload uppercase antibody sequences, mark known missing residues with *, and select the heavy-chain or light-chain model. AbLang returns each submitted sequence beside its restored form, highlights replaced masks, preserves FASTA record names, and provides restored FASTA and detailed JSON files for download or downstream workflows.
| Field | Description |
|---|---|
Antibody sequences | FASTA or plain sequence text using uppercase standard amino-acid codes. Mark known missing residues with *. Each sequence can contain up to 158 residues; multi-sequence input should use FASTA. |
Missing residues can appear anywhere in the sequence, though the most common use case is N-terminal restoration. For example:
>heavy_chain_example
EVQLVESGGGLVQP**SLRLSCAASGFTF**SYAMSWVRQAPGKGLEWVSAI| Setting | Default | Description |
|---|---|---|
Chain type | Heavy chain | Selects the original heavy-chain or light-chain AbLang model. Heavy chains typically begin with EVQ or QVQ; light chains with DIQ or EIV. |
Restore terminal residues | Off | Uses IMGT numbering to infer missing residues at sequence ends before restoration. Enable this when the number of missing terminal residues is unknown; the restored sequence length may change. |
Results display every submitted sequence beside its restored version. Replacements for explicit * masks are highlighted when positions remain directly comparable.
| Output | Description |
|---|---|
| Restored sequence view | Preserved FASTA record name, submitted sequence, restored sequence, chain model, residue counts, and identity. |
ablang_restored_sequences.fasta | Primary restored FASTA output. Record names are preserved and the file can feed another sequence workflow step. |
ablang_restoration.json | Detailed records, summary values, settings context, and source and model provenance. |
AbLang works best for sequences that resemble those in its training data, antibody variable domains from the OAS database. Performance may degrade on highly unusual antibodies or non-human sequences not well-represented in OAS.
Terminal restoration depends on successful IMGT numbering. Antibody-like variable domains are the intended input; sequences that cannot be recognized by the numbering model may fail when Restore terminal residues is enabled.
Geometric deep learning model for predicting protein binding sites directly from 3D structure. Identifies where proteins interact with other proteins, antibodies, or disordered proteins with high accuracy, including for novel protein folds.
Predict protein aggregation nucleation propensity from amino acid sequences using the Lehner Lab CANYA neural network.
Isoelectric Point Calculator 2.0 - Predict protein/peptide isoelectric point (pI) using 18+ validated pKa scales, SVR models, and deep learning. Supports proteins, peptides, and comprehensive analysis.
Find open reading frames in DNA with the native NCBI ORFfinder search, translations, coordinates, and output files.
Predict protein solubility from amino acid sequence using the University of Manchester Protein-Sol method.
Predict metal and water binding sites in protein structures using 3D convolutional neural networks (AllMetal3D + Water3D).
Score protein mutations with evolutionary profiles from homologous sequences and inverse folding. EvoIF returns a dimensionless log-odds score for each submitted single or multi-site mutation.
Predict multiple protein developability properties from amino-acid sequences using a multitask ProstT5 adapter.
Predict protein thermostability changes (ΔΔG) for point mutations using a graph neural network. Enables computational saturation mutagenesis screening to identify stabilizing mutations.
Faithful static-mode Aggrescan3D tool for per-residue aggregation propensity analysis from a single protein structure.