Two methods for measuring the non-randomness of chromosome abnormalities

Applying conventional statistical techniques to cytogenetic data usually faces the problem of working with small numbers and many classes. We describe two techniques, one based on a binomial test procedure, the other on Monte Carlo simulations, aimed at studying the non-randomness of chromosomal aberrations. Both techniques were used to study the distribution of breakpoints involved in variant Philadelphia translocations in chronic myeloid leukaemia. The results showed that 28 bands were non-randomly rearranged (P less than or equal to 0.05). Furthermore, the probabilities calculated from the binomial test procedure were close to those calculated from the Monte Carlo simulations.