On the role of risk-associated genetic loci in modulating clinical course in multiple sclerosis