Advanced glossary on genetic epidemiology
- 1Institut Municipal d’Investigació Médica, Barcelona, Spain
- 2Departament de Ciéncies Experimentals i de la Salut, Universitat Pompeu Fabra, Barcelona, Spain
- Correspondence to: Dr N Malats, Carrer del Dr Aiguader 80, E-08003 Barcelona, Spain;
This is the last of a series of glossaries on terms used in genetic epidemiology published by the journal. This glossary covers the most advanced genetic terms, most of which are related to new study designs and laboratory techniques. It provides the reader with examples and references of real studies that applied each of the study designs defined in the glossary. This should help the reader grasp the subtleties of each of these strategies and will allow the reader to research the literature according to their interest.
Given the space constraints, we chose not to be exhaustive and be concise. Hence, we again encourage the interested reader to “explore” the classic bibliography on genetics and genetic epidemiology.3–6
AFFECTED SIB-PAIR APPROACH
Study design used to find genetic factors contributing to a complex trait. It tests for linkage by considering the proportion of shared alleles between affected sib-pairs at markers spaced over the whole genome or over a section of it. A null distribution of the expected relative frequencies of sibs sharing zero, one or two alleles at a marker can be derived and tested against the observed data. An excess of allele sharing at a marker may indicate the presence in its vicinity of a gene contributing to the disease. This method also permits testing of gene-environment interaction. This design was applied by Lachmeijer et al to assess the involvement of IL1B and IL1RN gene polymorphisms in causing pre-eclampsia.7 They collected 150 pairs of sisters that had suffered pre-eclampsia while pregnant and typed two polymorphisms at IL1B and one at IL1RN. Unfortunately, the degree of allele sharing among sisters did not suggest that those genes were involved in pre-eclampsia.
ALLELE SPECIFIC AMPLIFICATION
Polymerase chain reaction based methods for detecting disease causing mutations that consist in amplifying specifically one or both alleles by using specific primers in one or two independent reactions. If two allele specific primers are used in a single reaction, additional chemistry is needed to determine which primer produced the amplification.
Comparison of the frequency of alleles in candidate genes between unrelated affected and unaffected individuals. The alleles analysed may be thought to contribute to the disease or be in linkage disequilibrium with any such causative variation. It can provide sufficient power to distinguish slight variations in disease risks being more sensitive than linkage methods when the genes of interest contribute to disease susceptibility but are neither necessary nor sufficient to cause disease. The methodology it uses is the same as used by epidemiological studies (cohort and case-control design). For instance, Perikac-Vance et al had mapped a gene conferring susceptibility to late onset Alzheimer’s disease at chromosome 19q13.28; as apoliprotein E is found bound to the amyloid plaques characteristic of Alzheimer’s disease and is also found in that genome region, it became a candidate gene for Alzheimer’s disease. This was confirmed by Strittmatter et al by typing variants of the ApoE gene in 30 affected individuals and in 91 presumably healthy controls.9 They found that the frequency of the APOE-∈4 allele was significantly higher in the patients than in the controls, which showed that this allele confers susceptibility to Alzheimer’s disease.
CASE ONLY DESIGN
Approach to screen gene-environment interactions under the assumption of independence between exposure and genotype in the population. This design does not require control subjects. Therefore, sample sizes will be less than half than those required in case-control studies and the estimated odds ratios will not suffer from potential biases related to control selection. Cases are distributed in a 2×2 table according to their genetic and environmental exposure status. To further explore the differences between a case only and a case-control design we suggest the reader looks at the study by Bai et al that compared both approaches to assess gene-environmental interaction on the disease liability.10
CASE-PARENTAL CONTROL DESIGN
Design based on the TDT test, which compares the relative frequencies of transmitted and non-transmitted alleles from parents to their affected offspring. It prevents the confounding effects of population stratification and permits testing of gene-environmental interactions by stratifying cases according to their environmental exposure status. For example, in a seminal paper, Spielman et al compared the genotypes at the insulin gene of juvenile diabetics and their parents and found that heterozygous parents transmitted to their affected children class 1 more often that other classes of alleles, and therefore concluded that susceptibility to juvenile diabetes is linked to the insulin gene.11
Any strategy that permits finding the chromosomal location of one or more genes, often related to a disease. See affected sib-pair approach, case-parental control design, and linkage analysis.
Fraction of the total phenotypic variation in a population that is caused by genetic differences between individuals: genetic variance/total variance. The genetic variance is the part of the total variance that is caused by allelic variations at whatever loci influence the trait. The total variance is the amount of variation in phenotype in a defined population. It only applies to a population on which observations are made and cannot be extended to other populations that have different allele frequencies or environments. Therefore, it cannot be used to explain differences between populations. Lichtenstein et al applied this strategy to assess the effects of heritable and environmental factors in cancers at various sites on the basis of the twin registries from Finland, Sweden, and Denmark.12
Strategy for gene mapping by testing for linkage between markers and phenotypes using families. In classic linkage analysis the transmission model is fixed (possibly with parameter values obtained from segregation analysis) and the likelihoods (LOD scores) of the disease and marker data are compared under the null hypothesis of no linkage and the alternative hypothesis of linkage. Non-parametric linkage analysis avoids fixing an explicit mode of inheritance of the disease. Free application programs for human genetic linkage analysis are listed, classified, and available for downloading (http://linkage.rockefeller.edu/). For instance, using data from 39 families containing individuals affected with cystic fibrosis, Tsui et al found that the inheritance of alleles at the D0CRI-917 polymorphism seemed to be linked to cystic fibrosis.13 Later on, and guided by this discovery, Kerem et al found that cystic fibrosis was caused by mutations in the CFTR gene, which is close to the D0CRI-917 polymorphism.14
A condition in which alleles at two loci or genes are found together in a population at a greater frequency than that predicted simply by the product of their individual allele frequencies. Alleles at markers near disease causing genes tend to be in linkage disequilibrium in the affected individuals. This is particularly the case in isolated, homogeneous populations, in which it can be assumed that most affected individuals carry the same mutation. Thus, Hastbacka et al found that diastrophic dysplasia, a rare disease almost confined to Finland, mapped to the genome region 5q32–q33.1 by observing that, in patients, alleles at the polymorphims in that region were in close linkage disequilibrium with each other.15
A statistical estimate, obtained in linkage analysis, which indicates whether alleles at two loci are inherited together more often than expected and are thus likely to be placed near each other on a chromosome. A LOD score is the ratio of two probabilities: (1) the probability of the observed inheritance of a trait (usually a disease) and alleles at a marker in a pedigree if they were linked given a inheritance model for the trait and a recombination probability between marker and disease, and (2) the probability of the observed inheritance of a trait and marker in a pedigree under the assumption that they are not linked. A LOD score is the logarithm of the ratio of those two probabilities. LOD scores can be added across pedigrees, and are usually taken to indicate significant linkage if they are above three. The recombination fraction that gives the highest LOD score from a marker of known genomic location can be used to map a gene.
A novel method of studying large numbers of genes simultaneously by automating and miniaturising a hybridisation detection system. The method uses a robot to precisely apply tiny droplets containing DNA to glass slides. Researchers then attach fluorescent labels to DNA from the cell they are studying. The labelled probes are allowed to bind to complementary DNA strands on the slides. The slides are put into a scanning microscope that can measure the brightness of each fluorescent dot; brightness reveals how much of a specific DNA fragment is present.
MULTIFACTORIAL THRESHOLD (MFT) MODELS OF INHERITANCE
Models that assume the joint effect of multiple genes and environmental exposures in determining the liability of an individual to present the trait of interest. A threshold is assumed under which the subject would not present the trait and above it would.
POLYMERASE CHAIN REACTION (PCR)
A procedure for obtaining a large number of copies of a particular segment of DNA. The principle depends on the requirement by DNA polymerase of a primer with a 3′ end to which nucleotides can be added. Two such synthetic primers define a segment that is replicated in a thermal cycle of denaturation, reannealing (reformation of complementary primer-DNA structure), and replication. Each cycle, which takes two to three minutes, doubles the amount of DNA between the primer boundaries. Thirty cycles would yield 230 copies. PCR has made it possible to characterise extremely small amounts of DNA.
RESTRICTION FRAGMENT LENGTH POLYMORPHISM (RFLP)
Genetic variation at the site where a restriction enzyme cuts a piece of DNA. Such variations affect the ability of the restriction enzyme to cut, and therefore, produce different fragment sizes. Most RFLPs are single base pair changes in the 4–6 bp target sequence of the restriction enzyme. Vice versa, many single nucleotide polymorphisms (SNPs) are RFLPs and can be detected with this technique.
Analysis of the inheritance ratios of offspring from a particular parental cross to test for conformity with Mendelian theory. Either genotypes or phenotypes can be the object of segregation analysis.
Determining the exact order of the base pairs in a segment of DNA by biochemical methods. Semiautomated biochemical methods are available for sequencing, which are based in the sequential incorporation of fluorescently labelled nucleotides.
SINGLE STRAND CONFORMATION POLYMORPHISM (SSCP)
Fast and simple technique widely used for mutation detection in various diseases. Basically, a fragment of interest is amplified by PCR, followed by electrophoresis in non-denaturing gel. The mutant DNA is separated from the normal due to the difference in mobility in electrophoresis, which is believed to be caused by the conformational change of the single stranded mutant DNA. Usually the DNA fragment size is restricted to less than 200 bp as the sensitivity of PCR-SSCP decreases with fragment size.
WHOLE GENOME SCAN
Linkage analysis in which markers placed at regular intervals and covering the whole genome are typed. It is often the first approach when no genetic information is available about a particular phenotype. For instance, Stefansson et al found that neuregulin-1 is a candidate gene for schizophrenia after typing 950 microsatellite markers covering the whole genome in 110 Icelandic patients for whom they had reconstructed their genealogical relationships.16
Both authors have contributed equally to the manuscript.