Click on image to view larger version.
Fig. 3. Mismatch distribution of 141 cacao accessions in the USDA cacao collection maintained at Mayaguez, Puerto Rico. Computer program MM-DIST (Kalinowski et al., 2006) was used for computation of mismatch distribution. The data set from the single genotyping was used to compute the probability that two individuals differ at k loci. Both empirical and expected distributions showed that the full siblings will likely differ by at least three out of 15 loci in their multilocus SSR profiles. The unrelated individuals will likely differ by at least five out of 15 loci. The error mismatch distributions are therefore unlikely to overlap with distributions of true genotypic difference (McKelvey and Schwartz, 2004; Kalinowski et al., 2006).
