In general, the fitness increment of any one allele depends in a complicated way on many other alleles; but, because of the way that the science of population genetics was developed, evolutionary scientists tend to think of epistasis as the exception to the rule. In the first models of natural selection devised in the early 20th century, each gene was considered to make its own characteristic contribution to fitness, against an average background of other genes. In introductory college courses, population genetics is still taught this way.
Examples of tightly linked genes having epistatic effects on fitness are found in supergenes and the human major histocompatibility complex genes. The effect can occur directly at the genomic level, where one gene could code for a protein preventing transcription of the other gene. Alternatively, the effect can occur at the phenotypic level. For example, the gene causing albinism would hide the gene controlling color of a person's hair. In another example, a gene coding for a widow's peak would be hidden by a gene causing baldness. Fitness epistasis (where the affected trait is fitness) is one cause of linkage disequilibrium.
Studying genetic interactions can reveal gene function, the nature of the mutations, functional redundancy, and protein interactions. Because protein complexes are responsible for most biological functions, genetic interactions are a powerful tool.
Two-locus epistatic interactions can be either synergistic (negative) or antagonistic (positive). In the example of a haploid organism with genotypes (at two loci) AB, Ab, aB or ab, we can think of the following trait values where higher values suggest greater expression of the characteristic (the exact values are simply given as examples):
|No epistasis (additive across loci)||2||1||1||0|
Hence, we can classify thus:
|Trait values||Type of epistasis|
|AB = Ab + aB - ab||No epistasis, additive inheritance|
|AB > Ab + aB - ab||Synergistic epistasis|
|AB < Ab + aB - ab||Antagonistic epistasis|
Understanding whether the majority of genetic interactions are synergistic or antagonistic will help solve such problems as the evolution of sex.
However, the evidence for this hypothesis has not always been straightforward and the model proposed by Kondrashov has often been criticized for assuming mutation parameters far from real world observations. For example, see