Computational optimization of hiPSC-derived red blood cell clone banks for alloimmunized patients in a saudi population

Abstract

Abstract Alloimmunization to red blood cell (RBC) antigens remains a major barrier to safe transfusion, particularly for chronically transfused patients, because compatible antigen-negative units are difficult to source from conventional donor inventories and demand for rare phenotypes is increasing. This study aimed to quantify the minimum size of an O RhD-negative (O-negative) human induced pluripotent stem cell (hiPSC)–derived RBC clone bank needed to cover alloimmunized transfusion recipients in a Saudi population and to provide a quantitative framework for prioritizing candidate clones based on real-world alloantibody frequencies and antigen-negativity constraints. Retrospective antibody-screening/identification data from 3,719 patients (185 screen-positive) and phenotyping data from 100 donors were analyzed; matching was restricted to 14 unique O-negative donor phenotypes as candidate clones. Antibody-identification text was normalized with a case-sensitive dictionary preserving paired antigens (e.g., C/c, E/e, K/k, S/s), and patient–clone compatibility required negativity for every antigen targeted by a patient’s alloantibodies. A strict set-cover optimization was solved using a greedy cumulative coverage heuristic and validated by exact binary integer linear programming (ILP) in R. Among 153 matchable positive-screen patients, 190 antibody-specificity occurrences were identified, dominated by anti-E (24.21%), anti-K (23.16%), anti-D (17.37%), and anti-c (7.89%). Across candidate clones, none was c-, e-, or k-negative; clone 57 provided the highest standalone coverage (83.66%). Greedy selection (57, 10, 9) achieved a coverage plateau at 134/153 (87.58%), and ILP confirmed three clones as the global minimum to attain the maximum achievable coverage; relative to all 185 screen-positive patients, coverage was 72.43%. Nineteen matchable patients were incompatible with all clones, largely driven by anti-c (15/19). These findings support a small, optimized hiPSC-RBC clone bank for broad coverage, while identifying anti-c–driven gaps that motivate expanded donor genotyping/phenotyping and targeted RHCE editing (e.g., CRISPR) alongside experimental work on differentiation, enucleation, antigen-expression stability, epigenetic regulation, immune safety, and cost-effectiveness. Code is publicly available (Zenodo: https://doi.org/10.5281/zenodo.22123689 ) under IRB approval (IRB-2017-03-79).

The authors' abstract, as published at the source. Scientific Reports, 2026 · DOI ↗