Tomohiko Ishihara

This paper shows that enhancing TARDBP exitron splicing reduces TDP-43 aggregation and neurodegeneration in disease models using ASOs.

The paper develops MOVA, a machine learning method using AlphaFold2 structural data to predict pathogenicity of ALS-related gene variants, showing good accuracy for TARDBP and FUS.

The paper shows that age-related DNA demethylation in the TARDBP 3'UTR reduces alternative splicing and increases TARDBP mRNA expression in the human motor cortex, linking this to earlier ALS onset.