System, method, and computer software product for genotype determination using probe array data

A method for genotyping a plurality of polymorphisms in a nucleic acid sample, the method comprising:

• hybridizing a nucleic acid sample with a plurality of allele-specific perfect-match probes provided in an array of perfect-match probes for a plurality of target sequences which the array is designed to genotype, wherein, for substantially all of the plurality of target sequences, the array is without corresponding mismatch probes;
• summarizing intensity values associated with the hybridizing to obtain a plurality of signal values, the plurality of signal values comprising a respective signal value for each respective allele of the plurality of polymorphisms;
• transforming the respective signal value for each allele into a space that includes a contrast dimension to obtain a plurality of transformed signal values;
• evaluating likelihood of all plausible divisions of the transformed signal values into seed genotypes to identify most likely plausible divisions;
• averaging over the most likely plausible divisions to derive a plurality of seed genotype clusters;
• and
• genotyping the plurality of polymorphisms, wherein genotyping comprises comparing the transformed signal values with a set of typical values for each genotype, wherein the set of typical values comprises prior values, wherein the prior values comprise estimates of genotype cluster center locations and genotype cluster center variances of the plurality of seed genotype clusters determined from clustering properties of the transformed signal values;
• wherein the steps of summarizing, transforming, evaluating, averaging, and genotyping are performed on a computer, and wherein the computer comprises a computer processor.