LECTURE 10: EVALUATING GENOMIC VARIANTS

Question 1

Why do genetic testing? = 7

Answer 1

1 * Accurate molecular diagnosis
* facilitates access to appropriate treatment and services
* Ends “diagnostic odyssey”

2 * Disease prognosis
* severity of disease, disease course and survival

3 * Cascade screening (Identification of at-risk, asymptomatic family
members)
* surveillance and preventive measures
* informed reproductive choices

4 * Screening (neonatal, carrier testing)
* (pre-natal, pre-implantation, pre-symptommatic, and post-natal)
* early detection and treatment can prevent irreversible consequences
* informed reproductive choices

5 * Disease and treatment monitoring
* response to treatment, relapse
* drug sensitivity and resistance

6 * Enrolment into clinical trials
* Included on disease-specific patient registries

7 * Rationale therapeutic design for medical research
- Especially gene therapy

Question 2

Recap on a couple of definitions: 4

Answer 2

1 * Pathogenic variant: capable of causing disease.

2 * Benign variant: not harmful in effect (e.g. no negative change in phenotype).

3 * Be careful of “mutation” (variable usage) – strictly speaking these are ONLY pathogenic/disease-causing variants.

4 * Be careful of “polymorphism” (variant that is common in a normal population, usually considered >1%) – e.g. a benign variant.

Question 3

Diagnostic variant detection is targeted = 8

Answer 3

1 * In diagnostics, investigation of variants is usually “targeted”

2 * focus on a specific gene location, size or region, and/or type of variant:
3 * Exonic and splice-site sequences;
4 * Gains/losses larger than a predetermined minimum size

5 * Disease-associated:
6 * Recurrent mutations
7 * Genes, or regions within genes (e.g. “hot spots”, or panels of known associated genes)

8.* In research, the net is usually cast far wider, even as wide as whole exomes or whole genomes (becomes very time-consuming).

Question 4

Diagnostic variant detection is targeted

• These restrictions are chosen because: 3

Answer 4

1 * Ease of interpretation

2 * Proven clinical utility

3 * Limited resources (staff, consumables, equipment)

Question 5

Not all variants in an individual’s DNA are
disease-causing = 5

Answer 5

1 * For example:

2 * If we were to sequence an individual’s BRCA1 gene we might detect a list of several DNA variants.
3 * The task is then to determine which of those variants are disease causing…

4 * i.e. likely to affect gene expression or protein function in a detrimental manner

5 * Perhaps one will be. Perhaps none will be.

Question 6

Diagnostic lab performs genetic tests:

Whole gene sequencing (screening) VS
Genotyping(known variant detection)

Answer 6

Molecular interpretation

Does the variant have a functional
effect?

Example: The variant detected is rare,
predicted to affect protein function, and
shown to affect protein function in a
mouse model.

Clinical interpretation
Is the variant relevant to the clinical
the question being asked?

Example: The variant has been identified
in other patients with the same clinical
features, the zygosity and familial segregation fits with the inheritance pattern for this disease

Genotyping(known variant detection)

Molecular interpretation of the variant present or absent

Clinical interpretation Is the variant relevant to the clinical the question being asked?

Example: The result is consistent with
patient features and findings from
other diagnostic tests.

— Lab reports test result

Question 7

Steps in Determining Variant Pathogenicity = 12

Answer 7

1 * Is the variant known, or novel?

2 * Databases of disease-associated variants
3 * HGMD, ClinVar
4 * disease-specific or locus-specific databases

5 * Databases of “normal” variants
6 * dbSNP, ExAC, gnomAD, exome variant server
7 * >5 minor allele frequency (MAF) variants are immediately excluded

8 * Seen Locally?
9 * Maybe just not published, or a sequencing/aligning error?
10. * OMIM, Gene Reviews, published literature

11 * Critically appraise the evidence
12 * Do you agree with previous opinions? Do you think this is pathogenic
or not?

Question 8

First Principles = 10

Answer 8

1 * Generally assumed nonsense and frameshift variants are diseasecausing, provided that a loss of function/haploinsufficiency/dominant
negative effect is supported (might be recessive Vs dominant).

2 * Generally assumed synonymous variants and those not near genes are
not disease causing.

3 * Missense more difficult to predict, may have:
4 * no effect
5 * loss of function
6 * gain of function

7 * Both deletions and duplications can cause disease
8 * More likely if large
9 * Less likely if agenic (not in a gene) *
10. breakpoints may interrupt genes or create fusion genes

Question 9

sequence variants vs copy number variants

Answer 9

table…important on slide 15

Question 10

Revisit types of mutations = 6

Answer 10

1 * Mutation:
* change in DNA sequence (compared to normal sequence) that is disease-causing.

2 * Mode of inheritance
* Dominant, recessive, autosomal, X-linked

3 * Alleles/zygosity
* Heterozygous, homozygous, compound heterozygous, hemizygous

4 * Genomic location
* Coding, promoter/regulatory, splice-site

5 * Molecular change
* Substitution (synonymous, non-synonymous, missense, nonsense)
* Deletion/insertion, expansion/contraction

6 * Functional effect
* Loss or gain of function

Question 11

Types of mutations - location

Answer 11

diagram on slide 17

Question 12

Revisit types of mutations - Zygosity = 4

Answer 12

1 * Homozygous
* Same mutation present on both alleles

2 * Heterozygous
* Mutation present on one allele only

3 * Compound heterozygous
* Two different mutations present, one on each allele

4 * Hemizygous
* A mutation on one allele, and the other allele is absent

Question 13

Types of mutations – mode of inheritance = 2

Answer 13

1 * Recessive disease
* Requires 2 recessive alleles
* Homozygous or compound heterozygous state
* OR, requires 1 mutant allele only if other allele is absent (hemizygous)

2 * Dominant disease
* Requires 1 dominant allele
* Heterozygous state

Question 14

De novo and inherited mutations in autosomal dominant disease =4

Answer 14

1. If the variant is de novo  more likely to be disease-causing…
2. Variant detected in one parent = less
   likely to be disease-causing
3. Variant not detected in either parent = more likely to be causally linked
4. Severe congenital autosomal
   - dominant disorder:
   - Variant detection.

Question 15

Variants in trans or in cis

Answer 15

Autosomal recessive disease results from a loss of function mutation on both alleles

Question 16

Filtering (applying this to multiple variants) = 5

Answer 16

1. Starting variants
   - Remove variants observed
   - frequently in normal populations
2. Fewer
   - Remove synonymous variants
3. Fewer again
   - Remove variants not in coding
   - regions or at splice sites
4. Even fewer
   - Remove variants predicted to have
   little effect on protein function
5. Final set

Question 17

“In silico” prediction: 10

Answer 17

1 * Conservation across species

2 * Physicochemical difference in amino acid substitutions
3 * Grantham score (categorise amino acid substitutions by dissimilarity)

4 * ‘In silico’ prediction algorithms
5 * Effect of amino acid substitution on protein function
6 * SIFT, Polyphen2, MutationT@ster…

7 * Creation/destruction of splice sites
8 * NNSplice, MaxEntScan, Human splicing finder…

9 * Haploinsufficiency (loss of one allele sufficient to cause disease)

10 * Imprinting data (is maternal/paternal origin important)

Question 18

“In silico” prediction: SIFT and Polyphen2

Answer 18

diagram on slide 27

Question 19

“In silico” prediction: Conservation & amino acid group

Answer 19

Glycine
Special group: Side-chain consists of H
atom only; can fit a hydrophilic or
hydrophobic environment

Aspartic acid
Polar charged: Hydrophilic side chain tends to be fully charged, form ionic bonds and are often involved in chemical reactions.

diagram on slide 26

Question 20

Gene and disease specific data

Answer 20

table important on slide 28

Question 21

Functional data

what is it? range? goal? = 6

Answer 21

1 *Functional studies focus on the dynamic effects of changes at the DNA level on gene transcription, translation and protein-protein interactions.

2 *Range from:
3 * cDNA studies to examine the effects of DNA changes on transcripts, to
4 * complex animal models (eg mice, zebrafish) where DNA or transcript level changes are introduced and in vivo phenotypic changes are measured

5. *Goal is to understand the relationship between genetic variants and phenotype: *
   
   6. typically in the realm of research laboratories, although some diagnostic laboratories use cDNA work for the investigation of splicing variants

Question 22

Family studies

Answer 22

*Attempt to demonstrate:
* segregation with disease
* inheritance pattern
* de novo variants
* in cis and in trans

Question 23

Other data - Multiple variants in a single codon

Answer 23

Example: DNAJB6 gene associated with myofibrillar myopathy (slowly progressive weakness associated with myofibril degeneration on EM).

Functional studies demonstrated one variant to result in reduced function, therefore assume that the other variants that cause the same amino acid change do too.

Question 24

Other data - Variant hot-spots

Answer 24

• Example: SOS1 gene and Noonan syndrome
• Disease-causing variants have nonrandom distribution, occurring in specific functional domains.

Question 25

Sequence variant classification = 3

Answer 25

1 * Most laboratories adhere to a 5 tier classification system

2 * There are published guidelines to assist classification,
e.g. AMCG =
American College of Medical Genetics & Genomics

3 * Local knowledge can also alter the confidence to call variants as significant

Question 26

Describing sequence and copy number
changes:

Answer 26

1 * Changes are described, using internationally recognised nomenclature, with reference to the known sequence.

2 * Human Genome Variation Society (HGVS) nomenclature
* Recommended descriptions for changes at DNA, RNA and protein level
* See http://varnomen.hgvs.org/

3 * International System for Cytogenetic Nomenclature (ISCN)
4 * Recommended descriptions for changes at chromosome and microarray level

5 * Nomenclature designed to be meaningful and unequivocal

6 * Must mention version of the reference sequence

7 * Understood and recognised internationally across laboratories and
countries

8. A report using recognised nomenclature and correct referencing will be interpretable by any other lab in the world and will continue to be interpretable in years to come

Question 27

Describing DNA sequence level changes
cont…

• Sequence changes are described according to: 2

Answer 27

• Sequence changes are described according to:

1 * change at the DNA coding level (denoted c.)

2 * change at the amino acid or protein level (denoted p.)

Question 28

RNA codon to amino acid conversion

Answer 28

codon chart on slide 36

Question 29

Example Case 1

Answer 29

1 * 30 yo female presents to during the second trimester of her
first pregnancy. On ultrasound examination, the foetus has
features consistent with a lethal skeletal dysplasia.
2. Medical termination of pregnancy is performed

3 * Following postmortem and consultation with a geneticist, genetic testing is requested to confirm a diagnosis of thanatophoric dysplasia (osteochondrodysplasia with short
limbs, narrow thorax and normal trunk length)

4 * This disease is usually autosomal dominant, due to gain of
function mutations in the FGFR3 gene.

5 * The result is:
* Hg19, FGFR3: c.[1118A>G];[=] p.[(Tyr372Cys)];[=]

6 * What is the mutation?
o TAT > TGT (tyrosine  cysteine)

7 * How could this mutation be described?
o Nonsense, missense or frame-shift?
o Homozygous or heterozygous?

Question 30

Example Case 2:

Answer 30

1 * A couple presents in ED with their 1 week old son, who presents with vomiting, dehydration and shock.

2 * There is a history of other male babies presenting with similar features in this family.

3 * Following a genetics consultation, genetic testing is requested to confirm a diagnosis of X-linked adrenal hypoplasia congenita.

4 * Usually caused by loss of function mutations in NROB1 gene.

5 * The result that comes back:
Hg19, NR0B1: c.[192C>A];[0] p.[(Tyr63*)];[0]

6.* What is the mutation?
* TAC > TAA (tyrosine  STOP)

7 * What type of mutation is this (missense, nonsense or frameshift)?

8 * Why does it look homozygous?