M-regression, false discovery rates and outlier detection with application to genetic association studies

Robust multiple linear regression methods are valuable tools when underlying classicalassumptions are not completely fulfilled. In this setting, robust methods ensure that theanalysis is not significantly disturbed by any outlying observation. However, knowledge ofthese observations may be important to assess the underlying mechanisms of the data.Therefore, a robust outlier test is discussed, together with an adequate false discoveryrate correction measure, to be used in the context of multiple linear regression withcategorical explanatory variables. The methodology focuses on genetic association studiesof quantitative traits, though it has much broader applications. The method is alsocompared to a benchmark rule from the literature and its good performance is validatedby a simulation study and a real data example from a candidate gene study.