Double recombinants are poor surrogate markers for genotyping errors
by
Melanie Bahlo
The Walter and Eliza Hall Institute of Medical Research

Errors in genotyping data can arise as a result of pedigree errors, genotyping errors and mutations. These errors have been shown to significantly reduce evidence for linkage, even when only present in small numbers. Mendelian error checking is the usage of mendelian inheritance rules to deduce inconsistencies in genotyping data. Sibpair studies are a common and popular way of finding linkage for complex human diseases. Often it is not possible to get parents of these sibpairs for genotyping due to the late age of onset of the disease being studied or the decision has been made to save on genotyping costs by not typing parents. In these cases it is not possible to carry out mendelian error checking. Unusually closely spaced recombination events are a surrogate marker for data inconsistencies. Likelihood methods exist (Douglas et al. 2000) that test if closely spaced double recombination events are in fact more likely due to errors, rather than chance. Other methods use more heuristic approaches, such as checking for multiple recombinants by using the haplotype output from GENEHUNTER. We show that double recombinants are poor surrogate markers for genotyping errors. Any method based on this principle performs poorly for markers spaced at genome wide scan typical distances, with a high false positive and false negative rate. It can also produce false positive linkage results. Obviously likelihood based methods are superior to those based on heuristic approaches and offer the only safe way of detecting errors using double recombinants. These methods should not be viewed as viable substitutes for mendelian error checking. Douglas JA, Boehnke M, Lange K (2000). A multipoint method for detecting genotyping errors and mutations in sibling-pair linkage data. American Journal of Human Genetics 66:1287-97

Date received: August 28, 2001