BIOL 606 Home

Molecular Methods Of Identifying Paternity

Presenter : Colin Reynolds

Lecture 6, February 23, 1998
Focal Paper: P.A. Morin, J. Wallis, J.J.Moore and D.S. Woodruff . 1994 Molecular Ecology 3, 469-478.
The accurate assessment of paternity has, until recently been difficult if not impossible to achieve, and its scope largely of interest only in human paternity cases. With the advent of modern molecular techniques it is now possible to address questions of paternity in natural populations with much more certainty. The development of these techniques has an obvious impact on the fields of wildlife ecology, systematics, and evolution.

Parentage and therefore paternity is the smallest scale of evolution (at least the finest scale of heredity). Accurate assessment of paternity has shown to be a powerful tool for understanding questions in wildlife ecology including assessment of mating structure in bighorn sheep populations. Using molecular techniques it has been shown that 33-48% of offspring are fathered by coarsing males, a result that is not necessarily expected due to the fact that females are usually protected by large dominant males, and the increase in fitness to coarsing males is not immediately apparent. Such studies have shown that is possible to gain insight into the evolution of male specific characters and behaviours through the assessment of paternity.

Historically the molecular analysis of paternity was based on blood antigen systems. In 1900 Landsteiners described the first system for paternity analysis using the ABO blood antigen group proteins, which was followed by the use of other red blood cell antigens, isoenzymes, serum proteins, and most recently by the use of the highly polymorphic human leucocyte antigen (HLA). Some of these systems including the ABO blood system have the advantage that alleles are inherited in a Mendelian fashion and therefore potential fathers and be excluded on the basis that they do not share an allele in common with offspring. Although these systems laid the ground work for modern molecular methods of paternity analysis the loci studied were relatively invariable and require large pure DNA samples.

In order to address the problem of low variability in markers, multilocus DNA fingerprinting methods were developed, including minisatellites (in 1985), AFLP's, RAPD's, DAF, and AP-PCR. Of these methods minisatellites (DNA fingerprinting) was been employed most extensively. DNA fingerprinting with minisatellites involves isolation of a large amount of DNA, restriction digests, Southern blotting, and hybridization with a human minisatellite repeat probe. Although a large number of loci are examined simultaneously and therefore a high level of paternity exclusion can be obtained this method as well as many listed above have many disadvantages. Two of the main disadvantages include the fact that bands on the gel are not inherited in a Mendelian fashion, and all individuals in a paternity study must be compared on the same gel (greatly increasing the amount of DNA needed and the number of gels run).

The discovery of highly variable microsatellites (SSR's or STR's) has revolutionized paternity studies in natural populations. Microsatellites are repeated 2-6bp motifs which can exist in 5-15 allelic forms. Microsatellite genotyping employs PCR and therefore requires little DNA, alleles are co-dominant and inherited in a Mendelian fashion. Automated DNA sequencers or allelic ladders allow for comparison between gels. Although microsatellites must be cloned for new species, existing loci may work in related species. Disadvantages of microsatellites include allelic drop out (where chance variation in the amount of template DNA can cause heterozygotes to appear as homozygotes), and null alleles (mutations in primer binding sites).

When assessing paternity in a closed natural population such as the Gombe chimpanzees (Morin et al. 1994), all sampled potential males except one should be excluded. In contrast in an open population all sampled males may be excluded indicating the father was not sampled, or one male may be included but the probability of another potential father should be given. When mothers and offspring are both sampled, males can be excluded on the basis that they do not share the proposed paternal allele with the offspring, however when only offspring are sampled males can be excluded on the basis that they do not share either allele in the offspring. Thus probabilities of exclusion are high in studies where both mother and offspring are sampled.

Morin et al. found that with a suite of 8 loci paternity exclusion reached 98.9%. However they could only assign three offspring to fathers with absolute certainty. This result is likely constrained by the fact that many males (potential fathers of older offspring) had died due to disease, and DNA could not be isolated from many juveniles. Like bighorn sheep it is interesting to note that alpha males did not father at least one of the offspring (Faustino) yielding questions of the role of these males in chimp society. The study does however allow for the assignment of parentage to some offspring and will argument the well documented behavioral information on this study group.
BIOL 606 Home


Discussants : Renee Polziehn and Keith Jackson Rapporteur : Corey Davis

The discussion opened with the question of why should paternity be studied in a systematics and evolution coarse. Although parentage is the finest scale of heredity it is not clear that it shows the stream of evolution. Paternity studies do however allow the identification of mating structures and may allow for insight into the evolution of male mating strategies and behaviours. It was suggested that in order to study such things we do need a technique to establish the existing mating situation. More generally molecular markers may have the ability to estimate effective population size, subpopulation status, and gene flow between populations (introgression, migration etc.).

A hypothetical example of subpopulations of whitefish in a northern Alberta lake was then examined. The difference in allele frequencies at a suite of microsatellite loci between two populations may constitute separable subpopulations, and levels of gene flow between the two populations may also be estimated. It was notes with much certainty that the chance of two populations being different by chance at this undefined suite of loci would be less than 10-12.

The nature of and concerns about microsatellite markers were then examined. Firstly, microsatellites are defined as tandemly repeated short sequence motifs. They do not need to be variable in order to be define as microsatellites. It was stated that microsatellites are dispersed randomly about the genome of eukaryotes, and are selectively neutral. Concern about this neutrality was expressed by the fact that some tri-nucleotide repeats have been linked to human diseases. It was also stated that repeated regions may play some role in genome organization or as regulatory regions of genes.

The evolution of microsatellite loci was then examined. It was suggested that microsatellites may only have the resolving power to discriminate between populations (taxa) separated by ten thousand years of evolution. This statement is based in the facts that microsatellite loci have constrained allele sizes (neutrality?), longer repeats have higher mutation rates, longer repeats tend to get shorter, and shorter repeats tend to get longer. Phylogenies using differences in average allele size in bears have resulted in placement of black and brown bears in a clade to the exclusion of polar bears, which is in marked contrast to the accepted phylogeny. Modelling of microsatellite mutation rates or the use of species (stickleback) with known population ages may aid in defining the scope of usefulness of microsatellite markers.
Other concerns when using microsatellite markers include allelic drop out and null alleles. Null alleles can best be detected by examining loci in known pedigrees. Given that this is not always possible other ideas, such as individuals which one is unable to genotype, or deviations from Hardy-Weinberg expectations, may indicate null alleles. Allelic drop out occurs when template DNA solutions are of low concentrations low quality (as can often occur when using hair samples, which is often necessary as in Morin et al. 1994). Chance variation of which of the two copies, or both, is introduced to the PCR reaction can cause errors in genotyping. Studies have concluded that by repeating the typing seven times on homozygotes will yield 95% confidence interval on the genotype. Another study tested the error rates in typing on chimp hair. It was noted that type two errors were higher than expected, but that when using plucked hair type one errors were acceptable. It was thought that the same animals were types in both papers and that comparisons could be made to estimate the amount of mistyping in Morin et al. 1996. However the two papers use largely the same loci but not the same individuals.

The nature of microsatellite repeats in different higher taxa was examined. It was proposed that there may be some phylogenetic content in the fact that different groups have different common repeat motifs. The difference may be based on the mutation rates for different DNA polymerases in different lineages.

In conclusion it was felt that using microsatellites to study parentage is a worthy endeavour but that little was gained in studying the Gombe chimpanzees.

BIOL 606 Home