Over the course of 3 years, 118 blood samples from patients referred for cytogenetic analysis were provided for somatic mutational analysis with the GPA assay. Some were referred for specific analyses related to known or suspected syndromes of deficiency in DNA repair syn­dromes [17-19, unpublished results] and many were diag­nosed with cytogenetic abnormalities [20, unpublished results]. However, 21 samples were from patients who had normal karyotypes and exhibited phenotypic abnormalities of unknown etiology. Eleven of these samples were hetero­zygous for the MN blood group. This being the ideal phe­notype for GPA somatic mutational analysis, they formed the basis of this study (Table 1).
Mutation frequencies for these patients (Table 1) were compared against those derived from a large population of new cord blood samples [4] and a small group of pediatric controls (n = 16, age = 10.4±3.1 years). Both sets were used because the pediatric controls, although the youngest available, were still significantly older than our patient population (p <0.001), and several studies have demon­strated an age-dependence in GPA Mf [21-23]. The total GPA Mf in these patients (8.5 ± 4.3 x 10^–6 ) was not signifi­cantly different from those of the newborns (Mf = 7.4 ± 6.5 x 10^–6 , p = 0.30) or the pediatric controls (Mf = 7.9 ±3.6 x 10^–6 , p = 0.83). These differences remained insignif­icant when the effect of age was taken into account (p  = 0.48, 0.45 for comparisons with the newborn and pediatric populations, respectively), as shown in Fig. 1, panel A.

Mutation frequencies in the GPA assay can be mecha­nistically characterized as those with loss of expression of one allele (“allele loss” or “N/O” mutants), and those with loss of expression of one allele and overexpression or du­plication of the remaining allele (“loss and duplication” or “N/N” mutants) [1,2]. The allele loss Mf in these patients, 2.9 ± 2.7 x 10^–6 , is significantly lower than that of the newborn population (3.9 ±2.1 x 10–6, p = 0.018) and lower than that of the pediatric population, 4.2± 2.7 x 10^–6 , although this difference does not quite reach statisti­cal significance (p = 0.055). Similar results are found if these comparisons are performed on age-adjusted data (p = 0.012, 0.081 for comparisons with the newborn and pedi­atric populations, respectively), as shown in Fig. 1, panel B.
Allele loss and duplication Mf for these patients was 5.6 ±2.5 x 10^–6, which is significantly higher than that of the newborn population (3.4 ± 6.1 x 10^–6 , p = 0.003). A similar with the N/N Mf of the pediatric controls (3.7 ± 1.4 x 10^–6 ) showed a trend in the same direction (p = 0.061). Age adjustment had little effect on these compari­sons (p = 0.003, 0.053 for comparisons with the newborn and pediatric populations, respectively), as shown in Fig. 1, panel C.