Linkage information and bias in the presence of interference.

Although interference is an important biological phenomenon, linkage analyses often ignore it for computational efficiency. In this paper we examine the effects of ignoring interference on estimates of recombination fraction, and the power of statistical tests to detect interference. We show that for gametes from phase-known triple heterozygotes, no bias in the recombination fraction for adjacent loci results from the assumption of lack of interference even if interference is present. However, for samples from pairwise informative samples, there is bias introduced. An analytical form for this bias is obtained.