Back

PIRCHE application versions 3 and 4 lead to equivalent T cell epitope mismatch scores in solid organ and stem cell transplantation modules

Matern, B. M.; Niemann, M.

2024-01-09 bioinformatics
10.1101/2024.01.09.574785 bioRxiv
Show abstract

Elevated PIRCHE scores between recipient and donor in organ and stem cell transplantation have been shown to correlate with increased risk of donor-specific HLA antibodies and graft-versus-host disease, respectively. With each revision of the PIRCHE application server, it is critical to completely evaluate the predicted scores, and compare with previous revisions. This manuscript compares the newly introduced PIRCHE version 4.2 with its predecessor version 3.3, which has been widely used in retrospective studies, using a virtual cohort of 10,000 transplant pairs. In the stem cell transplantation module, both versions yield identical results for 100% of the test population. In the solid organ transplantation module, 97% of the test population has identical PIRCHE scores in both versions. The deviating cases could be attributed to a refinement in the PIRCHE algorithms specification. For the 3% of cases with deviations, the determined magnitude of the difference is likely to be below the detection limit for clinical effects. We hereby confirm the equivalence in PIRCHE scores generated by the application server versions 3.3 and 4.2.

Matching journals

The top 9 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.