Two-base substitutions at each of two nucleotides in the factor IX gene (F9), but not part of CpG dinucleotides, were recently reported in a small population sample collected in Mexico, a significant observation of recurrent sites ("hotspots") of mutation (P=0.00005). When these new data were combined with previously collected mutation data into two progressively larger and inclusive Latin American samples, additional mutations were observed at one recurrent site, nucleotide 17747, and an additional recurrent nucleotide was observed such that the recurrent nucleotides in these larger samples were also significant (P=0.0003 and 0.0003). In contrast, in three non-Latin American control samples, there was at most only one nucleotide that recurred only once, most likely a chance recurrence (P>/=0.5). When the significance of substitutions was analyzed at each recurrent nucleotide individually, nucleotide 17747 was shown to be a significant recurrent nucleotide by itself in all the Latin American population samples (P</=0.02). Furthermore, a standard statistical comparison of mutation frequencies in the previously collected data alone confirmed that the frequency of mutation at nucleotide 17747 is significantly higher in Latin Americans than in all other populations combined (P=0.01). Thus, nucleotide 17747 is a germline mutation hotspot in F9 specific to Latin American populations. This may be the first evidence for population-specific effects on germline mutation that causes human genetic disease. The significance of the observed recurrent sites was analyzed using new software called Hotspot Detector which is capable of detecting significant recurrent sites in small samples, extending the sensitivity of F9 as a human germline mutagen test. Hotspot Detector uses a Monte-Carlo simulation method that was validated by comparing its results with those from an exact probability formula derived from statistical theory.
Copyright 2000 Wiley-Liss, Inc.