Estimation of the allele number in a natural population by the method of isofemale lines.

The number of alleles present in a natural population of unknown structure is estimated using a sequential sampling procedure applied to isofemale lines. Two questions are raised: how many individuals per isofemale line must be assayed and how many isofemale lines must be sampled to get an adequate sample to estimate the number of alleles, at a given risk, of the natural population? On the one hand, we show that when wild females are inseminated once, only two individuals per line are required. On the other hand, the number of isofemale lines that must be sampled depends on the risk chosen of losing an allele, on the number of alleles present in the population and on their drawing probabilities. When the population structure is known, an accurate answer can be provided. For an unknown population structure, one general sequential sampling previously described by J. Rouault and P. Capy is proposed to estimate the number of alleles in the population from data on isofemale lines.