Interpretable deep learning to uncover the molecular binding patterns determining TCR–epitope interactions

Abstract

Background: The recognition of an epitope by a T-cell receptor (TCR) is crucial for eliminating pathogens and establishing immunological memory. Prediction of the binding of any TCR–epitope pair is still a challenging task, especially with novel epitopes, because the underlying patterns that drive the recognition are still largely unknown to both domain experts and machine learning models. Results: The binding of a TCR and epitope sequence can only occur when amino acids from both sequences are in close contact with each other. We analyze the distance between interacting molecules of the TCR and epitope sequences and compare this to the amino acids that are important for TCR–epitope prediction models. Important residues are determined by using interpretable deep learning techniques or, more specifically, feature attribution extraction methods, on two state-of-the-art TCR–epitope prediction models: ImRex and TITAN. Highlighting feature attributions on the molecular complex reveals additional insights to the domain expert about why the prediction was made and can offer novel insights into the factors that determine TCR affinity on a molecular level. We also show which residues of the TCR and epitope sequences determine binding prediction for ImRex and TITAN and use those to explain model performance. Conclusions: Extracting feature attributions is a useful way to verify your model and data for challenging problems where small hard-to-detect problems can accumulate to inaccurate results.

Can we predict T cell specificity with digital biology and machine learning?

“…Weber et al 12 achieved an average of 62 ± 6% ROC-AUC for TITAN, compared with 50% for ImRex on a reference data set of unseen epitopes from VDJdb and COVID-19 data sets. Values of 56 ± 5% and 55 ± 3% were reported for TITAN and ImRex, respectively, in a subsequent paper from the Meysman group 45 . Other groups have published unseen epitope ROC-AUC values ranging from 47% to 97%; however, many of these values are reported on different data sets (Table 1), lack confidence estimates following validation [46][47][48][49] and have not been consistently reproducible in independent evaluations 50 .…”

TEINet: a deep learning framework for prediction of TCR-epitope binding specificity

“…Trained with such a skewed dataset, TEINet was driven to make predictions based on the epitope sequences without the participation of TCRs, as discussed in Dens et al . [32]. That is, when the input pairs consist of frequent epitopes, the model tends to predict “1s”, and conversely, it is likely to predict “0s” when encountering pairs with infrequent epitopes.…”

Quantitative approaches for decoding the specificity of the human T cell repertoire

“…The lack of standardization for these factors complicates TCR-pMHC binding prediction and comparative benchmarks. Subsequent evaluations by Meysman’s group underscored the unpredictability of unseen epitope predictions, reinforcing the call for advanced models and rigorous, standardized evaluation protocols ( 65 ).…”