Accepted_test

Establishing the relationship between SOD1 protein mutations and ALS patient lifespan using molecular modeling and graph neural networks

by Venzel A.S. | Ivanisenko V.A. | Institute of Cytology and Genetics, SB RAS, Novosibirsk, Russia | Institute of Cytology and Genetics, SB RAS, Novosibirsk, Russia

Abstract ID: 642

Event: BGRS-abstracts

Sections: [Sym 3] Section “Structural biology of proteins nucleic acids and membranes”

The study addresses mutations in the SOD1 gene, encoding the enzyme superoxide dismutase 1, which are a primary cause of hereditary amyotrophic lateral sclerosis (ALS), an incurable neurodegenerative disorder. Misfolding of the SOD1 protein due to these mutations may be the molecular basis of ALS. These mutations can disrupt the hydrogen bond network within the protein, leading to mutant conformations that differ from the wild-type structure. The investigation focuses on the changes in hydrogen bonds in mutant variants and their relationship to protein conformation and properties. Graph neural networks (GNNs) are employed to analyze the hydrogen bond networks of mutants, characterize SOD1 mutant structures, and identify quantitative relationships between hydrogen bonds and protein properties, such as the lifespan of ALS patients. The resulting method correlates protein structure with properties by representing the protein structure as a hydrogen bond graph, where nodes represent amino acids and edges denote the stability of hydrogen bonds.