The reactivity graph of the IgM IgOme
Ongoing
Representing high-throughput antibody binding as an undirected graph of cross-reactivities, then reading its geometry.
Project · Ongoing
Problem
Microarray reactivity data are high-dimensional and noisy. How do we represent them so that cross-reactivity patterns among the mimotope probes become visible and comparable across samples?
Method
Measure reactivity correlations across patient samples. Use ROC curves to determine a correlation-level cutoff that predicts cross-reactivity. Build an undirected graph of predicted cross-reactivities and analyze its community structure using spectral embedding1,2.
Result & open questions
Highly connected graphs with clusters of mimotopes that represent sequence motifs. The clusters can be used to define a lower-dimensional representation of the IgM IgOme.
Related: Reading a reactivity graph · Graph representations
References
1.
Ferdinandov, D. et al. Reactivity graph yields interpretable IgM repertoire signatures as potential tumor biomarkers. International Journal of Molecular Sciences 24, 2597 (2023).
2.
Pashova-Dimova, S. et al. Changes in the public IgM repertoire and its idiotypic connectivity in alzheimer’s disease and frontotemporal dementia. J Neuroimmunol 409, 578775 (2025).