Multipoint radiation hybrid mapping: comparison of methods, sample size requirements, and optimal study characteristics.

Abstract

There are several statistical methods available for analyzing radiation hybrid (RH) data, but little is known about the ordering accuracy we can expect under common study conditions. Using analytic methods and computer simulation, we compared the ordering accuracy of three multipoint statistical methods: minimum breaks (MB), maximum likelihood (ML), and maximum posterior probability (PP). For 8, 12, and 16 markers and all combinations of numbers of hybrids, retention patterns, and marker spacings considered, the probabilities that the true order is identified as the best order were considerably higher with the ML and PP methods than with the MB method. ML and PP performed similarly, but PP tended to give slightly greater support for the best order than did ML. Our results can be used as guidelines for determining sample size requirements and optimal marker spacing for future RH mapping experiments. For equally spaced markers, intermarker spacing of 30 to 50 cR gave the highest probability of correctly ordering all the markers. For randomly spaced markers, 10-20 cR average intermarker spacing resulted in the highest proportion of markers being placed in a 1000:1 framework map. Assuming equal retention in the analysis when a centromeric model would be more appropriate did not affect the ability of the ML method to accurately order the markers, but did influence the distance estimates obtained.