I’ve just posted a new paper on using the simple roots of E8 to reduce the particle count of the Lisi model from 240 to 8. See DynkinParticleReductions.pdf