Variant Analysis & Annotation

Question 1

What are the issues with large ‘healthy’ population databases such as Exac and Gnomad?

Answer 1

• Likely includes individuals with late onset disorders

- Some populations more represented than other (i.e. Exac mainly European populations)

Question 2

What is a constraint score?

Answer 2

Constraint score indicates selection pressure, does not work as well for recessive disorders.

Question 3

What information is needed to assess a variants pathogenicity?

Answer 3

Interpretation of a variant for use in clinical decision making requires comprehensive knowledge of

• the patient’s phenotype
• mode of inheritance for the disease gene
• mutational mechanism (e.g. haploinsufficiency, dominant negative)
• protein structure/function
• the strength of the gene-disease relationship

With the exception of the patient’s phenotype data, most of this information can be obtained from the published
literature/databases by a clinical scientist who can also collate the required population data and in silico predictions of variant effect.

Filter variants AF <0.01 in population frequency databases (e.g. gnomAD and 100 genomes project)&raquo_space; Prioritise based on functionality HP= stop loss/gain changes, splice site sequences and non-synonymous AA changes… LP= Synonymous amino acid changes&raquo_space; Prioritise variants in genes that have been previously associated with disease phenotype of the referred patient (OMIM)&raquo_space; Select variants and review using PubMed and results of in-silico tool analysis

Question 4

What steps could be taken to try and elucidate the pathogenicity of a class 3 variant?

Answer 4

These might include further genetic or non-genetic tests, clinical investigations and/or co-segregation
testing. Contacting other laboratories to see if they have evidence to change classification.

RNA studies? variant could have created a cryptic splice site

Question 5

What are some examples of functional tests which can be used as evidence for PS3?

Answer 5

Biochemical- Enzyme level within range that is diagnostic for the disease e.g.
low/absent alpha-galactosidase levels in Fabry’s disease

Muscle Biopsy- Evidence of a specific pathology, defined by immunohistochemistry or direct morphological assessment, associated with a single genetic aetiology

Drug Response - Improved glycaemic response in patients with HNF1A/4A MODY treated with sulphonylurea tablets

Messenger RNA analysis - Aberrant splicing shown by reverse transcriptase-PCR and characterisation of product(s) by Sanger sequencing

Microsatellite analysis of tumour tissue - Somatic loss of heterozygosity at the MEN1 locus in a patient with a
germline novel variant

Neuroimaging - Findings consistent with a single specific genetic aetiology (e.g. a CASK mutation)

Question 6

What annotation can be used to filter variants?

Answer 6

• Remove common SNPs. MAF >5% in gnomad.
• remove 5’ and 3’ UTR regions
• remove non–splice related intronic variants
• remove synonymous variants-unless first or last in exon
• Protect all within HGMD and OMIM
• Phenotypic Prioritisation? Exomiser
• Gene panels (BED)

Question 7

What quality aspects can be used to filter variants?

Answer 7

• QUAL (Phred-scaled quality score for the assertion made in ALT)
• DP (Depth of coverage)
  Any flags applied to the VCF e.g. LowQual etc