3/5/2018

Precision Medicine: A Guide to Genomics in Clinical Practice >

APPENDIX 6: Interpreting the Pathogenicity of Genetic

Variants
Video: Making Sense of Genetic Testing Results

Video: Making Sense of Genetic Testing Results

Play Video

INTRODUCTION
Next-generation sequencing gene panels for diagnostic or predisposition testing can be expected to
uncover genetic variants in every patient tested. However, not all of these variants are disease causing. It is
the role of the clinical testing lab to interpret and classify genetic variants as disease causing (pathogenic)
or benign. For a health care provider who may order these tests, it is important to understand the rapidly
evolving science and art of how these classifications are made.

Key Point

Even a general understanding of the methods used in variant interpretation will give health care providers
an appreciation for the limitations of currently available genomic medicine.

VARIANT CLASSIFICATION
There is currently no set standard for classifying variants, but groups like the ClinGen consortium
(www.clinicalgenome.org) are working to define standard procedures for classifying and curating sequence
variants. The most widely used framework for classifying variants in disease genes is based on an earlier set
of guidelines proposed by the American College of Medical Genetics and Genomics (ACMG)[1]. Because
classification o en occurs in the context of incomplete information, it is rarely definitive, but rather,
couched in terms of certainty. The five-tiered ACMG classification (pathogenic, likely pathogenic, uncertain
significance (VUS), likely benign, and benign) applies to moderately and highly penetrant disease genes.
Low-penetrance disease gene variants, such as those that come from Genome-Wide Association Studies
(GWAS), are almost always common variants and are categorized separately as risk factors. The new ACMG
guidelines, updated in 2015, provide a much more detailed framework for variant classification[2]. Not all
laboratories follow this guideline exactly, nor do they necessarily use the same vocabulary when reporting
results. However, most laboratories consider roughly the same evidence in making a determination about
variant pathogenicity.

1/5
3/5/2018

When evaluating a variant, two fundamental considerations are made: first, can the gene containing the
variant be plausibly linked to a disease, and second, does the variant a ect the function of the gene? All of
the following discussion aims at answering these deceptively simple questions.

EVIDENCE USED IN ASSESSMENT OF PATHOGENICITY

E ect of variant on protein structure or function

Any type of variant can be pathogenic, although the likelihood decreases substantially when considering
nonprotein-coding variants. Putative loss of function (LOF) variants (stop-gained, stop-lost, frameshi
insertion/deletion, splice site disruptor AG/GT) are the most deleterious and best candidates for pathogenic
variants. Missense variants can also be pathogenic, but it's di icult to know if any given missense variant
will impact the protein in such a way as to cause disease. Computational algorithms that estimate
deleteriousness based on features like conservation and location include programs like PhyloP, Polyphen,
SIFT, MutationTaster, and others. These algorithms can best be described as informative but imperfect.
Also, variants that occur in mutational hotspots within the gene where other known pathogenic variants
cluster, and are of the same type of variant previously associated with the disease, favor pathogenicity.

Allele frequency in others with and without the disease

In the absence of families, presence of the specific variant in multiple independent individuals a ected
with the same condition and absence in unrelated healthy persons supports pathogenicity. Conversely, if a
variant is thought to have high penetrance, identifying an una ected individual with the same variant
provides evidence against pathogenicity. Rare variants, those having a frequency in the population less
than expected based on the disease frequency, allelic heterogeneity, and mode of inheritance, are more
likely pathogenic. For example, the risk of breast cancer in the population is about 11%, but only 5–10% of
breast cancer is thought to be inherited. Moreover, there are hundreds of variants that can lead to breast
cancer. Therefore, any single variant underlying breast cancer should have an allele frequency less than
1%.

Human genetic linkage/cosegregation studies

Human genetic evidence is derived from family or population-based (unrelated individuals) studies, where
observed coinheritance or correlation of a variant with a disease phenotype is inferred from statistical
linkage (for family studies) or association (for population studies). This is the most direct and important
evidence for pathogenicity of a variant, but also the most di icult to obtain. Sometimes a variant can be
reclassified as this type of data accumulates.

In vivo/in vitro studies of variant function

Various laboratory in vivo, ex vivo, or in vitro experiments may support pathogenicity of a variant. The
evidence is strongest when the experiment involves an observable phenotype that is consistent with the
condition under investigation. Examples of these types of studies include the use of cell-based assays or
animal models where the variant can be shown to disrupt the gene or alter a relevant phenotype or where
repair of the variant can rescue a phenotype. Many metabolic enzymes have clinically available activity

2/5
3/5/2018

assays that can be used to assess the impact of a variant. Otherwise, most functional gene studies are
considered research.

Previous classification as pathogenic

Increasingly, known variants are deposited into locus-specific databases and large variant repositories like
the commercially available Human Gene Mutation Database (HGMD®)[3] and ClinVar[4], the largest publicly
available repository. ClinVar includes all variants from OMIM, some locus-specific databases, as well as
variants submitted by individual testing labs. Submitters to ClinVar o en provide a classification of the
variant, according to their interpretation. These classifications are mapped onto the ACMG framework
(pathogenic, likely pathogenic, VUS, likely benign, benign) for consistency. At present, approximately three-
quarters of ClinVar variants are preclassified, but some have multiple conflicting classifications. ClinVar
records may or may not contain the supporting evidence to back up their classifications, but sometimes
contain links to PubMed articles.

WEIGHING THE EVIDENCE

These lines of evidence, and others, may be combined or weighed di erently by individual testing
laboratories. Consequently, it is possible that di erent labs will o er contradictory classifications of
variants. One study found a low rate of concordance when examining potentially pathogenic variants in
two arrhythmia genes with a two-lab concordance rate of 33% and three-lab concordance rate of only 10%
[5]. Similarly, major discrepancies in classification are found when comparing publicly accessible variant

databases[6].

A variant that lacks an evidence base to declare it as pathogenic or benign will fall into a gray zone—a VUS.
VUS rates vary from lab to lab. O en, laboratories that have run the most tests are able to classify variants
more easily. VUS rates are also lower for genes that are generally well studied and higher for those less well
studied.

Because of the degree of uncertainty around variant pathogenicity in many cases, variants found upon
testing, without a definitive classification, should be reevaluated periodically. Over time, as more evidence
accumulates, classifications should be more definitive.

COMMUNICATING WITH THE PATIENT

It is irresponsible, in our opinion, to express confidence in having achieved a genetic diagnosis or doubt a
diagnosis beyond what is warranted by available evidence. Ascribing pathogenicity to a variant may cause
the patient to stop seeking a diagnosis, may lead to major reproductive decisions, and may suggest
invasive screening or other medical (mis)adventures. Ignoring a variant may delay the patient's diagnosis.
Still, a core facet of the practice of medicine is the obligation to make clinical decisions, and paralysis in the
face of uncertain evidence is hardly preferable to over- or under-action. Rather, any clinician who utilizes
genetic testing must be prepared to:

Consider the likelihood of identifying uncertain variants, and the potential health implications thereof,
before ordering a test.
3/5
3/5/2018

Counsel patients on the possibility and ramifications of VUS before ordering a test.

Invest the time needed to fully understand the results from the laboratory and how they were derived.

Communicate with specialists, including genetic counselors, as needed.

Include and educate the patient in understanding what is and is not known about a variant and the
involved gene.

Weigh the risks and benefits of acting or not acting on a variant.

Request that the laboratory reanalyze the variant a er su icient time has passed, or educate the patient
that reanalysis is possible should they desire it in the future.

RESOURCES
Locus specific database (LSDB) list (http://grenada.lumc.nl/LSDB_list/lsdbs)

List of LSDBs based on various online resources and direct submissions, maintained by Leiden University
Medical Center.

Human Gene Mutation Database (HGMD) (http://www.hgmd.org)

The public version of HGMD is freely available to registered users from academic institutions/nonprofit
organizations while the subscription version (HGMD Professional) is available to academic, clinical, and
commercial users under license via BIOBASE GmbH.

ClinVar (www.ncbi.nlm.nih.gov/clinvar/)

A freely available searchable database from the NCBI of variants and their clinical classifications. The
database includes germline and somatic variants of any size, type, or genomic location. Interpretations are
submitted by clinical testing laboratories, research laboratories, locus-specific databases, OMIM®,
GeneReviews™, UniProt, expert panels, and practice guidelines.

The Significance of Unknown Significance: A Teachable Moment

(http://www.medscape.com/viewarticle/865546?src=par_jmbm_stm_mscpedt&faf=1)

A discussion of a lawsuit related to variant interpretation that highlights salient methods and pitfalls in
interpreting genetic test results.

REFERENCES

1. Richards  CS, Bale  S, Bellissimo  DB,  et al. ACMG recommendations for standards for interpretation and
reporting of sequence variations: revisions 2007. Genet Med . 2008;10:294–300.  [PubMed: 18414213]

2. Richards  S, Aziz  N, Bale  S,  et al. Standards and guidelines for the interpretation of sequence variants: a
joint consensus recommendation of the American College of Medical Genetics and Genomics and the
Association for Molecular Pathology. Genet Med . 2015;17:405–424.  [PubMed: 25741868]
4/5
3/5/2018

3. Stenson  PD, Mort  M, Ball  EV, Shaw  K, Phillips  A, Cooper  DN. The Human Gene Mutation Database:
building a comprehensive mutation repository for clinical and molecular genetics, diagnostic testing and
personalized genomic medicine. Hum Genet . 2014;133:1–9.  [PubMed: 24077912]

4. Landrum  MJ, Lee  JM, Benson  M,  et al. ClinVar: public archive of interpretations of clinically relevant
variants. Nucleic Acids Res . 2016;44:D862–D868.  [PubMed: 26582918]

5. Van Driest  SL, Wells  QS, Stallings  S,  et al. Association of arrhythmia-related genetic variants with
phenotypes documented in electronic medical records. JAMA . 2016;315:47–57.  [PubMed: 26746457]
[JAMA and JAMA Network Journals Full Text]

6. Vail  PJ, Morris  B, van Kan  A,  et al. Comparison of locus-specific databases for BRCA1 and BRCA2
variants reveals disparity in variant classification within and among databases. J Community Genet . 2015;
6:351–359.  [PubMed: 25782689]

McGraw Hill
Copyright © McGraw-Hill Global Education Holdings, LLC.
All rights reserved.
Your IP address is 200.3.149.5

Access Provided by: Pontificia Universidad Javeriana

Silverchair

5/5