Computational and Machine Learning Approaches in Solid Organ and Hematopoietic Stem Cell Transplantation

Allograft rejection and transplant failure remain major challenges in both solid organ and hematopoietic stem cell transplantation (HSCT). Conventional donor–recipient matching based on HLA allele-level comparisons only partially reflects the molecular determinants of immune recognition. Subtle variations at the epitope level can critically influence alloimmune and graft-versus-host responses. Molecular or epitope-level matching provides a more precise assessment of compatibility by identifying the specific amino acid residues targeted by donor-specific antibodies and T cells. These molecular insights pave the way toward personalized transplantation medicine, where immunological risk can be quantified and immunosuppression or conditioning regimens optimized. Incorporating computational and machine learning (ML) models into molecular matching enhances the prediction of rejection or GVHD and supports more equitable, data-driven donor allocation strategies.The goal of this Research Topic is to advance computational and machine learning (ML) approaches for molecular matching across both solid organ transplantation and HSCT. Despite remarkable progress in HLA typing, immunogenetic databases, and data analytics, existing allele-based matching systems remain limited in their predictive capacity. Recent computational advances enable multidimensional integration of donor–recipient data, epitope mismatches, immunogenicity, antibody reactivity, and clinical outcomes, facilitating individualized risk prediction.We particularly welcome studies exploring whether models developed in solid organ transplantation can inform HSCT donor selection, and vice versa. The overlap in immunogenetic mechanisms makes ML-based tools highly relevant across both fields. Notably, groups such as ours are developing AI-guided tools for optimized donor selection in HSCT—illustrating the translational potential of computational immunology in precision transplant medicine. This Topic aims to encourage cross-disciplinary collaboration to validate and deploy interpretable, clinically applicable ML models that improve donor matching, minimize rejection or GVHD, and optimize long-term outcomes.This Research Topic welcomes original research, reviews, and perspectives advancing computational and ML models for molecular (epitope-level) matching in transplantation, including both solid organ and HSCT settings. We encourage submissions on:• Computational pipelines for mismatch quantification and immunogenicity prediction• ML-based prediction of allograft rejection or GVHD• Integration of NGS-based HLA typing with molecular models• Interpretable ML and AI-guided decision-support tools for donor selection• Ethical, regulatory, and allocation challenges in AI-driven transplantationBy bridging computational immunology, data science, and clinical transplantation, this Research Topic aims to accelerate the adoption of precision-matching tools to enhance graft survival, reduce immune complications, and improve donor allocation equity.Topic Editor Kelley M.K. Hitchman serves on the speakers bureau of One Lambda (a Thermo Fisher Company). All other Topic Editors declare no competing interests with regards to the Research Topic subject. Topic editors: Ahmed Mostafa, Karen Sherwood, Kelley M.K. Hitchman Source: https://www.frontiersin.org/research-topics/75362/computational-and-machine-learning-approaches-in-solid-organ-and-hematopoietic-stem-cell-transplantation