Current challenges for unseen-epitope TCR interaction prediction and a new perspective derived from image classification

The prediction of epitope recognition by T-cell receptors (TCRs) has seen many advancements in recent years, with several methods now available that can predict recognition for a specific set of epitopes. However, the generic case of evaluating all possible TCR-epitope pairs remains challenging, mainly due to the high diversity of the interacting sequences and the limited amount of currently available training data. In this work, we provide an overview of the current state of this unsolved problem. First, we examine appropriate validation strategies to accurately assess the generalization performance of generic TCR-epitope recognition models when applied to both seen and unseen epitopes. In addition, we present a novel feature representation approach, which we call ImRex (interaction map recognition). This approach is based on the pairwise combination of physicochemical properties of the individual amino acids in the CDR3 and epitope sequences, which provides a convolutional neural network with the combined representation of both sequences. Lastly, we highlight various challenges that are specific to TCR-epitope data and that can adversely affect model performance. These include the issue of selecting negative data, the imbalanced epitope distribution of curated TCR-epitope datasets and the potential exchangeability of TCR alpha and beta chains. Our results indicate that while extrapolation to unseen epitopes remains a difficult challenge, ImRex makes this feasible for a subset of epitopes that are not too dissimilar from the training data. We show that appropriate feature engineering methods and rigorous benchmark standards are required to create and validate TCR-epitope predictive models.

The pulmonary vasculature in lethal COVID-19 and idiopathic pulmonary fibrosis at single cell resolution

Aims

SARS-CoV-2 infection causes COVID-19, which in severe cases evokes life-threatening acute respiratory distress syndrome (ARDS). Transcriptome signatures and the functional relevance of non-vascular cell types (e.g. immune and epithelial cells) in COVID-19 are becoming increasingly evident. However, despite its known contribution to vascular inflammation, recruitment/invasion of immune cells, vascular leakage and perturbed hemostasis in the lungs of severe COVID-19 patients, an in-depth interrogation of the endothelial cell (EC) compartment in lethal COVID-19 is lacking. Moreover, progressive fibrotic lung disease represents one of the complications of COVID-19 pneumonia and ARDS. Analogous features between idiopathic pulmonary fibrosis (IPF) and COVID-19 suggest partial similarities in their pathophysiology, yet, a head-to-head comparison of pulmonary cell transcriptomes between both conditions has not been implemented to date.

Methods and Results

We performed single nucleus RNA-seq (snRNA-seq) on frozen lungs from 7 deceased COVID-19 patients, 6 IPF explant lungs and 12 controls. The vascular fraction, comprising 38,794 nuclei, could be subclustered into 14 distinct EC subtypes. Non-vascular cell types, comprising 137,746 nuclei, were subclustered and used for EC-interactome analyses. Pulmonary ECs of deceased COVID-19 patients showed an enrichment of genes involved in cellular stress, as well as signatures suggestive of dampened immunomodulation and impaired vessel wall integrity. In addition, increased abundance of a population of systemic capillary and venous ECs was identified in COVID-19 and IPF. COVID-19 systemic ECs closely resembled their IPF counterparts, and a set of 30 genes was found congruently enriched in systemic ECs across studies. Receptor-ligand interaction analysis of ECs with non-vascular cell types in the pulmonary micro-environment revealed numerous previously unknown interactions specifically enriched/depleted in COVID-19 and/or IPF.

Conclusions

This study uncovered novel insights into the abundance, expression patterns and interactomes of EC subtypes in COVID-19 and IPF, relevant for future investigations into the progression and treatment of both lethal conditions.

Translational perspective

While assessing clinical and molecular characteristics of severe and lethal COVID-19 cases, the vasculature’s undeniable role in disease progression has been widely acknowledged. COVID-19 lung pathology moreover shares certain clinical features with late-stage IPF – yet an in-depth interrogation and direct comparison of the endothelium at single-cell level in both conditions is still lacking. By comparing the transcriptomes of ECs from lungs of deceased COVID-19 patients to those from IPF explant and control lungs, we gathered key insights the heterogeneous composition and potential roles of ECs in both lethal diseases, which may serve as a foundation for development of novel therapeutics.