A Review of Statistical Methods for Genome Mapping

Framework maps of the human genome are an important staging post in the on-going effort to sequence the entire genome. The existence of high quality maps is also a prerequistite for studies attempting to determine the location of genes involved in common diseases. The basic experimental approaches to constructing both genetic and physical maps are briefly described as well as their respective uses. A variety of statistical approaches to map construction are outlined including parsimony, maximum likelihood and Bayesian methodologies. The mostly widely used of these, the method of maximum likelihood, is discussed in detail, particularly in the context of physical mapping using radiation hybrids. Finally, current statistical issues and problems in the field of genome mapping are described.