03 Genetic Variation in Clinic Flashcards
What three ways can we use genetic variation to predict/quantify risk?
Create polygenic risk scores from SNPs across a genome.
Can look for SNPs associated with increased LDL cholesterol leading to FH, who were negative on other tests.
When a patient has a history of breast cancer, but no one mutation, could then look at several SNPs affecting risk.
How can we use genetic variation to predict the response to therapy? Give an example
Pharmacogenomics. Use best drugs, and prevent side effects and adverse reactions. Vemurafenib is a very favourable treatment to metastatic melanoma patients with a BRAF 600 variant.
What’s the third use of genetic variation clinically?
Diagnosis rare, high-impact mendelian disease
How mnay will suffer from a rare disease in the UK?
1 in 17. That’s 1.3 million.
How many different rare diseases have been described?
6000
What proportion of rare diseases are genetic?
80%
What proportion of people with a rare disease die before their 5th birthday?
30%
Who is the reference human genome made up from currently?
13 individuals from buffalo, New York. 66% is 1 male.
Genome Build Hg38 came out in 2013. How many times a year is it updated?
4 times a year
What was the difference between Hg37 and Hg38?
Centromeric regions were added and some gaps were filled in
Who tends to use Hg37 and who tends to use Hg38 leading to some confusion?
The Clinic often uses Hg37. But Variation is mapped against Hg38.
Name these types of genetic variation in increasing size:
S_Vs
small in____s
S___ T____ R____
R__________ insertions
C___N____V___
L____ S_____ R_____
A_____
SNVs
small indels
Short Tandem Repeats
Retroelement Insertions
Copy Number Variants
Large structural rearrangments
Aneuploidy
What things can disrupt genes, their function, and their expression?
SNVs, indels, CNVs, damage, positional changes, epigenetic changes
What are the results of SNVs that are insertions or deletions?
A frame shift, that usually leads to a premature termination codon, leading to either a truncated protein, or nonsense mediated decay.
Why might a substitution SNV be synonymous?
Because the genetic code is degenerate
How can a substitution SNV in a protein coding region be damaging?
By affecting splicing
What are missense and nonsense?
A substitution SNV that leads to either a different amino acid, or a stop codon
Explain Inversion, Unbalanced and Balanced (reciprocal) Translocation in terms of larger chromosomal changes
A chunk of DNA flips.
A chunk of DNA moves elsewhere.
Two chunks of DNA swap places.
What effect can a deletion CNV of a gene have on the other alelle?
It can lead to the manifestation of a disease from a recessive variant on the other chromosome
What can be affected by a chunk of DNA being moved or deleted?
The breakpoints are critical, it can affect the reading frame, regulatory elements, exons, introns etc.
How much of our genomes are the same between people?
99.9%
How many changes are there between individuals in their genomes?
3 million
What are morbid genes?
All genes whose mutant form are causally associated with human disease
What is the OMIM resource for?
A database of gene variant-phenotype relationships. 6,200 phenotypes attributed to alterations in 3,900 genes. Tells you mode of inheritance etc.
What is the PanelApp resource for?
A resource of disease-gene associations that has been reviewed by experts as to the pathogenicity of variation in certain genes. 5589 genes and genomic entities on it.
What is the Red, Amber, Green ranking on Panel App?
Green genes have enough evidence that they should be use on panels in clinical testing. Amber its suggested that you await new evidence. Red genes have no evidence of a disease link at this time.
What factors do you need to consider when you ask the question, what genomic test should I do?
Well it’s dependent on the disease mechanism, the phenotype, the inheritance type and so on.
What are the different scales of how much of the genome you could test?
A single variant, exon or gene.
A gene panel - phenotypically heterogenous disease.
Only Morbid genes, 5000-6000 genes (clinical exome) when the phenotype is very vague, like prenatal cases, so you aren’t sure of the panel.
Complete exome.
Whole Genome.
What is the purpose of the national test directory?
To standardise genetic testing and provide equity of testing. Each UK nation needs to provide tests on the directory for their citizens if they meet the eligibility criteria. Prevents unneeded tests.
Whats a downside of the national test directory?
Some single gene testing now has to be a panel, this can increase incidental findings and take more time.
What are the 2 general questions you ask when you find a variant in a clinical test?
Does this variant disrupt function? And is that disruption of function sufficient to cause the disease seen in the patient.
Where can you find direct literature evidence for a variant classification?
HGMD, NCBI, ClinVar
Whats a downside of ClinVar
it tells you if a variant was identified before, but not the patient’s phenotype. It needs to be consistent to be used.
What are you looking for when looking for functional evidence?
Model organisms or in vitro evidence. But it needs to be robust, preferably replicated.
What evidence for variant classification can we generate in house?
Do a family study, see if the variant segregates with the disease or not, or if the variant is de novo.
Why does variant frequency matter?
Because higher frequency variants do not cause mendelian disease. It’s rare alleles that cause disease.
What are hypomorphic alleles?
Variants that cause reduced gene function, but not complete loss. Can manifest phenotypically, but lesser than a disease.
What can happen when you have two hypomorphic alleles?
When heterozygous, hypomorphic alleles can cause mild disease
How can some pathogenic variants occur in the healthy population?
Due to age of onset, and incomplete penetrance
What are the populations like in these CNV variation databases?
DGV
DbVar
WTCC
SgD
DGV - Unaffected caucasians. (DGV gold is better curated for quality)
DbVar - A mix of healthy and clinically relevant CNV variation.
WTCCC - Unaffected controls.
SgD - A limited number of unaffected controls.
GnomAD started as ExAC browser with how many exomes from control populations?
61,000
How many exomes and genomes are in gnomAD as of v3?
125,000 exomes and 76,000 genomes
Who are specifically excluded from gnomAD?
Cohorts from severe pediatric disease
Which databases can you use for CNV analysis to ask ‘has this CNV been seen numerous times in patients with a consistent clinical phenotype’?
DECIPHER, ECARUCA, ISCA, Unique, NHSE WGS
Which databases can you use for SNV analysis to ask ‘has this SNV been seen in numerous times in patients with a consistent clinical phenotype’?
100k project, NHS WGS, DDD, LSDBs/HGMD, ClinVar(Although can’t be sure of phenotype)
What are 2 less well used websites for SNV analysis:
G___M_____ for genes without a well published phenotype.
R________ to tackle unsolved cases
Gene Matcher
RDSolver
What questions can you ask to assess structural variants with theoretical predictions?
Is it balanced or unbalanced?
Can a deleted gene cause haploinsufficiency or a duplicated gene cause triplosensitivity?
Is a duplication in tandem or put elsewhere?
Has a gene moved away from its regulatory elements?
Where specifically are the breakpoints?
What questions can you ask to assess the effect of an SNV from a theoretical prediction?
What’s normal variation at this site?
Will it affect splicing?
Is this area preserved across species or related genes?
Is it in a conserved proximal control element (GC box, CAAT box, TATA box)?
Will it change the amino acid or create a stop codon?
What’s some of the filtering of all the NGS data that is produced?
Limit to genes suggested by the HPO terms for the patient phenotype.
What SNVs are high impact?
Transcript ablation (affecting start or stop codons), splice variants, stop gains, frameshifts, stop or start loss, transcript amplification.
What SNVs are moderate impact?
Inframe indels, missense variants, protein altering variants
What SNVs are low impact?
snyonymous variants, variants in introns