Gene Conservation Laboratory
Statistics Program for Analyzing Mixtures (SPAM) Software

SPAM FAQs Page

SPAM software logo

Proper Handling of Null Alleles

Loci with null alleles can be used in a SPAM baseline, but they must be treated as phenotypes (see page 20 - 21 of the SPAM 3.2 User's Guide).

  1. Use the observed genotypes and the null allele model to derive maximum likelihood estimates of the allele frequencies. NOTE: GENEPOP can do this.
  2. Assuming Hardy-Weinberg, calculate the expected probability for each possible genotype.
  3. Calculate the expected probability for each 'observable genotype'. For example, if the locus has 5 alleles,{A, B, C, D, E}, and E is a null allele, then the possible genotypes are:

    AA AB AC AD AE BB BC BD BE ...

    and the observable genotypes are

    AA'={AA,AE} AB' AC' AD' BB'={BB,BE} BC' BD' CC'={CC,CE} CD' DD'={DD,DE}.

    The probability of the observable genotype AA' = probability of the true genotype AA + probability of the true genotype AE.
  4. Define the locus as a PHENOTYPE in the *Characters section of the SPAM control file, with 10 possible characters (number of observable genotypes). For each population in the baseline, enter the relative frequency of the observable genotypes.

Why not just use observed frequencies?

Because you may not observe all possible genotypes, hence leading to structural zeros in the baseline.