Methods and apparatus for use in complex genetics classification are disclosed. In one illustrative example, a method is performed using data from a sample population which includes counts of individuals in the sample population associated with both a particular diploid haplotype pair and a particular genetic trait class. A correspondence to analysis (COA) on the data from the sample population is performed (at least in part) so as to determine a first plurality of n-dimensional coordinates for each diploid haplotype pair and a second plurality of n-dimensional coordinates for each genetic trait class. A linear or quadratic classification analysis is then performed based on the first and the second pluralities of n-dimensional coordinates and diploid haplotype pairs of the at least two genes from an individual sample. Advantageously, the individual sample can be accurately classified into one of the genetic trait classes using the method.