POLYMORPHIC REPEATS IN HUMAN GENES

The invention provides computational methods and compositions for identifying polymorphic repeats in genes. Candidate polymorphic repeats are identified by detecting tandem repeats in a target coding sequence, scoring the repeats for polymorphic probability, and generating a dataset correlating the repeats with polymorphic probability. Actual polymorphic repeat are identified by further detecting the candidate polymorphic repeat in each of a population of different coding sequences, and determining whether the candidate polymorphic repeat is polymorphic in the population. Computationally derived polymorphic repeats are validated with phenotypic variations and these correlates are used to detect the presence or absence of such phenotypic variation in test genes. Variances at polymorphic repeats are identified by detecting in a test gene or coding region the presence or absence of variance at a disclosed unconventional polymorphic repeat.