TCR-TRANSLATE: Conditional Generation of Real Antigen Specific T-cell Receptor Sequences
Bibliographic record
Abstract
The paradoxical nature of T-cell receptor (TCR) specificity, which requires both precise recognition and adequate coverage of antigenic peptide-MHCs (pMHCs), poses a fundamental challenge in immunology. Efforts at modeling this complex many-to-many mapping have been greatly impeded by a severe lack of experimental data. To address this, we present TCR-TRANSLATE, a novel framework that adapts low-resource machine translation techniques to the TCR:pMHC specificity domain. Here, we explore sequence-to-sequence (seq2seq) modeling with various training strategies, including semi-synthetic data augmentation and multi-task objectives to generate antigen specific TCR sequences for a given target of interest. We benchmark twelve model variants derived from the BART and T5 model architectures on a target-rich validation set of well-studied pMHCs, finding an optimal model, TCRT5, that generated validated antigen-specific CDR3β sequences for previously unseen antigens. While current limitations include a narrow validation set and a focus on the CDR3β loop, our approach demonstrates the potential of seq2seq models in rapidly generating antigen-specific TCR repertoires, offering a promising avenue for increasing throughput in precision immunotherapies. Our findings highlight both the capabilities and limitations of sequence-based conditional TCR design, emphasizing the need for experimental validation to bridge the gaps between predictions, metrics, and functional capacity.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".