I read Yates and colleagues' recent paper [1] on polymerase chain reaction (PCR)-based diagnosis of thiopurine S-methyltransferase (TPM) deficiency and the accompanying editorial [2] with great interest. The authors have shown that the presence of any of three missense mutations can be detected with a PCR-based assay. They have further demonstrated that genotype is highly correlated with measured enzyme activity. However, Yates and colleagues did not address several points worthy of comment. They did not explicitly cite studies showing that the sequence variants studied segregate with diminished enzyme activity. Such linkage studies would be particularly interesting in their patient who had diminished enzyme activity and no identified TPM mutation. They provide no population-based evidence on the presence of other silent variants in addition to T474C or on the presence of conservative substitutions that lack clinical significance. Finally, they do not address the smaller-than-expected proportion of homozygotes with TPM deficiency. Do these persons have a phenotype that interferes with viability in the absence of treatment with thiopurine drugs?

I recognize that because mutational analysis of TPM is still in progress, answers to my questions may not be known. The authors are to be congratulated on transferring technology from the research to the clinical laboratory to study this important therapeutic problem. I am confident that they appreciate that additional questions raised by their report do not represent criticism but rather a desire to learn more.