Our approach to a genetic analysis of G6PD deficiency in the Saudi population was to identify haplotypes of G6PD using a combination of DNA sequencing of exons 3, 4, 5, 6, 7, 9, 10, 11 and 12 and introns 11 and 13, utilising physical linkage to define G6PD haplotypes in male patients, and PHASE analysis to reconstruct haplotypes in female patients. Definition of haplotype phase in female patients who are compound heterozygotes (for mutations or polymorphisms) can be addressed successfully using PHASE reconstruction of haplotypes, when a suitably significant number of patients are available, as was the case in our study. Haplotype analysis addresses not only the distribution of pathogenic mutations, but their linkage in cis with other pathogenic mutations and polymorphisms, and thus permits linkage disequilibrium analysis for any given gene. In this communication, we report the results of a 6-locus haplotypic analysis of the G6PD gene in Saudi patients suffering from G6PD deficiency. 6-locus haplotyping identified 13 haplotypes in males (1 normal, 12 pathogenic) and 8 in females (1 normal, 7 pathogenic). Extending the haplotype analysis loci distal to exon 9 this resulted in only one further pathogenic mutation being identified (Anant, Arg463His). However it generates a significant amount of information about polymorphic variants within the gene and increased the total number of haplotypes observed. Analysis of pairwise linkage disequilibrium revealed significant linkage patterns between specific pairs of loci.
To date, only a few previously published reports of G6PD phenotype-genotype correlations have identified genotypes that represent combinations of more than one non-conservative mutation. Indeed, in a review by Beutler & Vulliamy [15–17] only 10 of 140 non-conservative amino acid mutations are combinations of more than one mutation. The present study has identified 5 novel, pathogenic G6PD haplotypes (Jeddah A through E), all of which are represented by combinations of 2 or 3 non-conservative amino acid substitutions (Table 2). Even allowing for the fact that these combinations are rare in the Saudi population and may be rare in other populations, this provides a significant insight into the cumulative mutation process in the G6PD gene.
We examined the effect of G6PD haplotype on levels of expression of G6PD in order to identify haplotypes which conferred the greatest clinical burden to patients. Our data supports the hypothesis that G6PD haplotypes representing two or more non-conservative amino acid mutations confer a greater reduction in G6PD expression than in haplotypes which only represent a single non-conservative amino acid mutation. This is perhaps unsurprising since the structure of the G6PD enzyme is more abnormal in the former cases.
Identification of the novel Jeddah A, B, C, D and E haplotypes could prove extremely important, since even carriers of two or more non-conservative amino acid mutations appear to have clinically significant G6PD deficiency.
The advantages of a haplotypic analysis of the G6PD gene is beyond doubt. We elected to pursue this approach to provide a wider viewpoint of G6PD genetics than a study focussed on the detection of single mutations. Detection of a given haplotype in males in our study provided confidence that identification of that haplotype in females represented confirmed linkage, not a computational artifact of the PHASE analysis.
Despite a significant effort in sequencing the G6PD gene in our patient cohort, ‘normal’ G6PD haplotypes were identified in 43 hemizygous male patients and in 26 female patients who possessed two apparently ‘normal’ X chromosomes. This may indicate the presence of a pathogenic mutation situated outside the regions covered by our sequencing strategy which may affect G6PD expression, but nonetheless our data is consistent with other studies  in which no G6PD mutations have been detected.
The predominant G6PD haplotypes in our patient cohort represent previously reported Aures (p.Ile48Thr) and Mediterranean (p.Ser188Phe) haplotypes. However, this report represents the first complete haplotypic analysis of the G6PD gene in patients from Saudi Arabia. We describe 5 previously unreported G6PD haplotypes which are characterised by novel combinations of extant mutations. These novel combinations pose interesting questions as to the existence of, and interplay between, putative mutation hotspots within the G6PD gene.