Methods and apparatus for complex gentics classification based on correspondence anlysis and linear/quadratic analysis

摘要

Software methods for identifying associations between genetic information and particular genetic traits are described. A candidate single nucleotide polymorphism (SNP) combination is selected from a plurality of candidate SNP combinations for a gene associated with (or suspected to be associated with) a genetic trait. Haplotype data associated with this candidate SNP combination are read for a plurality of individuals and grouped into a positive-responding group and a negative-responding group based on whether a predetermined trait criteria for an individual is met. A statistical analysis on the grouped haplotype data is performed to obtain a statistical measurement associated with the candidate SNP combination. The acts of selecting, reading, grouping, and performing are repeated as necessary to identify the candidate SNP combination having the optimal statistical measurement (if one exists). In one approach, all possible SNP combinations are selected and statistically analyzed. In another approach, a directed search based on results of previous statistical analysis of SNP combinations is performed until the optimal statistical measurement is obtained. In addition, the number of SNP combinations selected and analyzed may be reduced based on a simultaneous testing procedure.