Linkage Analysis

While most methods are now focused on identifying the precise molecular cause of disease, indirect tests can be quite useful, particularly for mapping causes of a disease in a family. For example, testing for markers such as single nucleotide polymorphisms (SNPs) or short tandem repeats (STRs, which are 2–5 base long repetitive elements with varying numbers of repeats) that are found throughout the genome can be extremely useful as a way to identify likely causal genes, particularly in diseases where multiple possible causal genes have been implicated. For example, in hereditary spherocytosis a number of genes including ANK1 , SPTB , SPTA , SLC4A1 , and EBP42 are implicated in the disease. Many of these genes are quite large and while sequencing a panel of genes is certainly possible, in a large family, the use of either SNP-based methods or other markers such as STRs can identify the likely causal locus to help focus targeted sequencing efforts. These methods are commonly used for diagnostic mapping in resource-poor settings where whole-genome sequencing (WGS) methods may not be available and these methods can also be extremely useful in other settings. For example, if a family is being followed with a known disease, but no coding mutations are identified on targeted sequencing, these approaches can help validate that there is linkage to a specific gene and they may assist in the efforts to identify mutations in regulatory regions of the implicated gene. Specifically, segregation of markers that are in linkage with the causal mutation should only be found in affected family members and would suggest that the causal mutation is located nearby. These methods are also commonly used as an initial screen in families where possible cancer predisposition syndromes may exist and can help focus in-depth analysis on certain regions of the genome. Even in cases where whole-exome or -genome sequencing is performed, linkage can provide an excellent indirect approach to focus on marker genes that segregate appropriately in individuals who have a particular disease.

Fluorescent In Situ Hybridization

Fluorescent in situ hybridization (FISH) was developed in the 1980s and uses fluorescently labeled DNA probes to query whether entire chromosomes or parts of a chromosome may be duplicated or deleted in cells. The fluorescently labeled DNA probes are complementary to the region of interest on a chromosome and therefore specifically hybridize only to this region and not to others. FISH is commonly used to assess for gain or loss of chromosomes or large parts of chromosomes in patients with hematologic malignancies. Typically, a number of cells from the bone marrow are tested for the presence of such chromosomal aberrations, which can have important roles both in terms of disease diagnosis and prognosis. FISH has a benefit in that it is a cytogenetic method and therefore individual cells are assessed rather than a population of cells in aggregate, which is the case for other methods that examine copy number variation, including array comparative genomic hybridization (CGH) or genome-wide SNP microarrays. In addition, FISH remains the best clinically available method to detect classic cytogenetic changes that are diagnostic and implicated in a number of pediatric cancers, such as translocations that are frequently seen in leukemia and certain solid tumors.

Array CGH

FISH lacks the sensitivity to detect smaller chromosomal deletions or duplications, which often represent important DNA copy number variations found in disease. Array CGH takes advantage of microarrays that have oligonucleotide probes at varying densities to detect differences in DNA copy number by comparing a sample genome with a reference sample (or group of reference samples) and examining whether there is an increase or decrease in signals at a particular genomic region in comparison with the control, which would be indicative of duplications or deletions, respectively. This method has significant sensitivity to detect DNA copy number changes, particularly smaller changes, in a variety of different samples. This can be applied to congenital disorders, where copy number changes can cause disease when present in the germline. In addition, in acquired hematologic or other malignancies, there can be acquired copy number changes. This increased sensitivity to detect smaller copy number changes has led to an increased detection of copy number changes in genomes of unclear significance. A number of resources are cataloging such changes in humans, although nonuniform deposition of deletion information into such databases makes the ongoing interpretation of either germline or acquired somatic copy number changes difficult in some cases. It is likely that as these databases grow with more phenotype information available, there will be increased insight into whether a deletion or duplication may be pathogenic.

Genome-Wide or Focused SNP Arrays

Microarrays provide the opportunity to genotype SNPs in a large-scale and potentially genome-wide manner. These approaches can be used for several applications. Similar to array CGH, these methods can be used to detect copy number variation that is either found in the germline or that is acquired. The resolution of the deletions detected using such approaches can be as good or in some cases, depending upon SNP or probe density, better than the resolution achieved with array CGH methods. In many cases, SNP arrays are used in place of array CGH in many clinical labs to detect copy number changes routinely. Indeed, the use of these SNP arrays is particularly widespread in the diagnostic evaluation of hematologic malignancies. Another application of such arrays is to genotype common SNPs in the genome, either for linkage mapping, as discussed above, or for identification of a common variation that confers risk of having certain diseases. While this has proven useful in some diseases, it is important to bear in mind that most such associations are probabilistic and not deterministic of acquiring disease. The clinical utility of this application is not clear, although a number of direct-to-consumer services offer such genotyping and will report relative risk information to individuals who request such services.

As more disease-associated mutations are being identified, there have been efforts to develop focused SNP arrays for specific phenotypes or diseases. In the future, such approaches may have clinical utility. However this application may be surpassed by large-scale genome sequencing as it becomes affordable (discussed below). One limitation of this approach is that there is continuous discovery of new causal alleles and genes in many diseases, which limits the clinical utility of such arrays.

Multiplex Ligation-Dependent Probe Amplification

Multiplex ligation-dependent probe amplification (MLPA) is a molecular approach that involves annealing of two adjacent oligonucleotides to a segment of genomic DNA followed by quantitative polymerase chain reaction (PCR) amplification to characterize copy number or other changes in the DNA. A series of MLPA probes can together screen and map deletions that occur in a particular region. In contrast to array CGH or SNP arrays, this approach is best applied to detect DNA copy number alterations in specific focused regions and this approach can allow such alterations to be finely mapped. For example, MLPA is commonly used to map deletions that commonly occur in the α-globin gene locus in cases of α-thalassemia. Traditionally, this mapping was done using Southern blotting (to determine a specific DNA sequence), but this is now rarely done for clinical purposes and in most instances MLPA is used for such applications in clinical labs. When results from MLPA are reported, it is important to bear in mind that the resolution will depend upon the number of probes used and, in general, precise deletion coordinates will not be defined using MLPA alone. Often to better map deletion sites, PCR-based Sanger sequencing approaches are used to map breakpoints once general coordinates have been defined using MLPA (discussed below).

Targeted PCR Analysis

Often a single mutation confers significant disease risk or occurs in the majority of cases of a disease. In these cases, PCR approaches can be used for amplification and separation of different alleles. A number of approaches to separate different alleles using PCR have been developed and since these methods are largely specific to individual platforms, the details of specific methods will not be covered here. These approaches are commonly used for detection of mutations, such as factor V Leiden that confers an increased risk of venous thrombosis and is found in several percent of the general population. In certain hematologic malignancies, such as myeloproliferative diseases, there are common mutations such as JAK2 V617F that occurs in many cases and focused genotyping of this variant is often performed as a clinical test. These tests can be done at low cost and relatively rapidly because of their focused nature and output of the presence/absence of a single mutation. However, using this approach it is impossible to detect relevant variants that have not been genotyped in a gene of interest.

Focused Sanger and Gene Panel Sequencing

Traditional DNA sequencing has relied upon the chain terminator method developed by Frederick Sanger, where a chain-terminating dideoxynucleotide is coupled to a fluorescent dye and this can allow sequencing in a single reaction by detecting terminated DNA fragments of various sizes. A chromatogram obtained after capillary-based separation displays the sequentially elongated fragments that each end in a specific fluorescent terminating dideoxynucleotide, allowing identification of the sequence of DNA. This approach is often applied to a series of reactions using PCR that cover a gene or in some instances a panel of genes implicated in disease. A limitation is that a single reaction can only assess a single sequence of several hundred bases and therefore multiple reactions will need to be run for most genes. It should be kept in mind that while Sanger sequencing can be very sensitive to detect point mutations, copy number or structural changes in genes will often be missed using this approach. Therefore, targeted sequencing of a disease gene can often be complemented using array CGH or SNP array-based approaches to look for deletions that may be implicated in a subset of cases of a particular disease.

Whole-Genome or -Exome Sequencing

While Sanger sequencing methods were once the primary method to map DNA sequences, the development of high-throughput next-generation sequencing (NGS) platforms over the past several years has allowed rapid and low-cost sequencing of large portions of the genome. NGS takes advantage of various technologies to sequence millions to billions of DNA strands in parallel to yield substantially more throughput than Sanger sequencing approaches. Moreover, NGS approaches bypass fragment-cloning steps that were necessary for genome sequencing using traditional Sanger sequencing approaches. In NGS, DNA is broken into short fragments, a subset of fragments may be enriched (such as with exome sequencing where sequences that encode regions overlapping with exons are enriched), and the sequences of all the fragments are then read on NGS platforms. The details of these approaches vary depending upon the technology or platform used, but in general all NGS methods use similar principles.

The use of NGS platforms to sequence the entire genome or exome of a patient with a particular disorder for clinical purposes has only begun to emerge and these technologies are currently being primarily used in the research setting. Studies are beginning to address the diagnostic yield of these approaches. One important consideration in using these approaches is that since the entire genome (or a substantial portion) is sequenced, potentially pathogenic variants may be identified that may not contribute to the disease that initially motivated the WGS. At the current time, there are varying opinions on the types of circumstances in which such incidental findings can be reported to patients and studies are exploring the impact of delivering such information to patients. There has been substantial debate in the community regarding the delivery and broad impact of the information derived from NGS on patients, physicians, and society. Undoubtedly, this is an area that will evolve considerably in the coming years.

Currently WGS or whole-exome sequencing (WES) are ordered in the clinical setting for detection of rare variants in patients with a phenotype that is suspected to be due to a single-gene disorder, after known single-gene candidates have either been eliminated or when a multigene testing approach is prohibitively expensive. Before such tests are ordered, it is important that clinicians gather a thorough family history, fully evaluate a patient’s phenotype, and obtain appropriate informed consent. There is little doubt that as WGS and WES are routinely employed in clinical settings, specific guidelines for when it is best to use these tests for patients with particular groups of diseases will emerge.

It is important for clinicians to be aware of the significant limitations of WGS and WES. NGS approaches can currently only sequence nonrepetitive regions of the genome and some diseases occur in repetitive DNA (i.e., fragile X syndrome) and thus diagnoses will be missed using such approaches. In addition, NGS approaches cannot currently detect most copy number variants, insertion–deletion variants, or chromosomal translocations. It is likely that with improvements in technology to allow for longer reads in NGS and with better computational methods, these types of variants could be detected using these approaches in the future.

Since numerous potential causal variants may be identified using such broad-based sequencing approaches, the American College of Medical Geneticists has developed recommendations regarding categories to which variants should be assigned and these are helpful for clinicians to be aware of when interpreting results from WES or WGS (although they apply to other sequencing approaches as well).

• •
  

  Disease-causing: A sequence variant that has previously been reported as a cause of a disorder

  
  
• •
  

  Likely disease-causing: A sequence variant that has not been previously reported, but is of a type that would be expected to cause disease

  
  
• •
  

  Possibly disease-causing: A sequence variant that has not been previously reported and is of the type that may or may not cause a particular disorder

  
  
• •
  

  Likely not disease-causing: Sequence variation that has not been reported and is not likely of the type that would cause a particular disorder

  
  
• •
  

  Not disease-causing: Sequence variation has previously been reported and is a recognized neutral variation

  
  
• •
  

  Variant of unknown clinical significance: Sequence variation is not known or expected to be causative of disease, but is found to be associated with a clinical presentation.

While these categories are useful when results are reported, they are largely dependent upon databases of prior variants. As more sequencing data are being reported, it is important to bear in mind that some variants previously thought to be pathogenic are being reclassified as benign. It is very likely that the categorization for a particular variant may evolve over time and therefore it is useful to evaluate the prior reports of any genetic variants identified in such studies on an individual basis. In some cases, it is important to bear in mind that in many single-gene disorders, variable penetrance or expressivity (where patients with a particular mutation may or may not have the disease or may have varying severities of the disease) may have a significant and underestimated impact.