Back

From mice to humans: A multi-omic predictive framework for translational immunology

Prates-Syed, W. A.; Lira, A. A.; Cortes, N.; Silva, J. D.; Hamaguchi, B.; Carvalho, E.; Castillo-Chavez, A.; Duraes-Carvalho, R.; Cabral-Marques, O.; Sabino, E. C.; Krieger, J. E.; Hagan, T.; Cabral-Miranda, G.

2026-09-24 bioinformatics
10.64898/2026.09.18.752716 bioRxiv
Show abstract

Mice are key preclinical animal models in vaccine and immunological research, yet their predictive value for human immunity remains contested. Here, we evaluated the translatability of murine models across inactivated and subunit vaccination (influenza, hepatitis B), acute infection (S. aureus, E. coli), and injury (burns and trauma), integrating transcriptomic profiles with sequence evolution, cis-regulatory architecture, and functional annotation. Functional modules were more conserved between species than individual orthologous genes. Translational accuracy depended on stimulus intensity, as infections and injuries engaged conserved signatures, while single-dose vaccination diverged. We then built multilayer models to predict human expression rank and direction of change, and to classify shared leading-edge genes. Adding evolutionary and regulatory layers improved these predictions. We provide a step-by-step R Markdown notebook to apply the models to user data. The study code and datasets are available at https://github.com/wapsyed/mousetohuman_multilayer

Matching journals

The top 8 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.