The effect of sample size on the estimation of the frequency of DNA-profiles in RFLP-analysis

Restriction fragment length polymorphism analysis (RFLP-analysis) was carried out on blood samples from 616 unrelated Danish Caucasians. DNA was restricted with Hinf I and analyzed with the single locus variable number tandem repeats (VNTR) probes MS1, MS31, MS43a and YNH24. The effect of the sample size on the estimates of the frequencies of DNA-profiles was investigated using reference samples of 50-1200 bands for each probe. The effects of using the upper confidence limit and a minimum default allele frequency in the calculation of the frequency of the DNA-profile were investigated. It is concluded that very small samples, e.g. 50 individuals, may be used as reference samples in the estimation of the frequency of a DNA-profile. The use of an upper confidence limit was without significant effect, whereas the use of a minimum default allele frequency prevented large underestimates.