L5, Genomics and Health II

Question 1

Trends in effect size/penetrance and allele frequency of gene variants:

Answer 1

• Highly penetrant mutations (large effect size) tend to have lower frequency (e.g. CF)
• Conversely, common variants typically have small effects (e.g. LMTK2 in Prostate cancer)
• Outliers include APOE4 (Alzheimer’s)

Question 2

GWAS: What is it? How is association established?

Answer 2

• Comparing two populations to see if frequency of genotypes is different
• SNP array -> generating genotypes for individuals with and without trait
• Higher incidence of a SNP allele in individuals with the trait compared to those without the trait = association (significance depends on degree of association and sample size)
• Nearby candidate genes can be identified and tested

Question 3

Why are SNPs useful in GWAS? What are tag and lead SNPs?

Answer 3

• SNP arrays use tag SNPs to represent SNPs nearby (haplotype), reducing the need to study every SNP
• LD clumping allows computational determination of index (lead SNP) within LD blocks -> reduce amount of information, can thus combine results from multiple studies

Question 4

Manhattan plots: What are they for?

Answer 4

• Used to visualise the p-values for investigated SNPs across the genome locations in GWAS
• Significant p-values show as peaks (uses -log^10(P) value)

Question 5

Limitations of GWAS:

Answer 5

• Cost millions
• Low predictive value (statistical association only with odds ratios less than 1.5)
• Around 90% of associated SNPs in non-coding DNA
• Difficult to study combinations of genes as disease cause -> Future prospect: polygenic risk scores
• Can’t correct for environmental factors (logistically difficult)

Question 6

Precision medicine:

Answer 6

• Therapy, instead of being generalised and responding to symptoms only, is tailored through DNA/RNA profiling
• Targeted therapy avoids the different effects of individuals to drugs
• Alternatively, DNA and RNA profiling can be used in determining how genetics influence drug response (pharmacogenetics)

Question 7

Application of transcriptomics to breast cancer treatment:

Answer 7

• Approximately 70% of all breast cancer patients present with ER+, HER2- breast cancer
• Adjuvant chemotherapy improves prognoses, but is exposing a large percentage needlessly to radiotherapy
• Oncotype DX (reverse transcriptase PCR-based), Mammaprint (micro-array based) are examples of transcriptomic methods for informing treatment via biomarker discovery

Question 8

Pharmacogenetics through genotyping arrays:

Answer 8

• Using DNA as inputs, custom microarrays and amplicon sequencing can be used to investigate gene variants for specific gene targets after PCR amplification -> more appropriate drug dosing
• e.g. CYP450 test (amplichip)

Question 9

Amplichip CYP450 test:

Answer 9

• Cytochrome P450 superfamily is a large and diverse group of enzymes that form the major system for metabolising lipids, hormones, toxins and drugs-> 12 of these are implicated in metabolism of most commonly used drugs
• Amplichip tests for CYP2D6 and CYP2C19 alleles -> ascertaining which metabolizer subgroup an individual belongs to

Question 10

Use of genetic testing for rare and inherited disorders and cancers:

Answer 10

• NHS currently tests a wide range of genetic disorders
• Both whole genome and whole exome techniques (among others)
• Useful to use a subset of genes in a panel - exon capture or exon amplification -> all possible alleles can be identified in one test (e.g. 450 gene panel for eye disorders)

Question 11

Filtering criteria for candidates to test for:

Answer 11

• Mutation must be likely to cause a change in gene expression or protein structure (e.g. nonsense, strong missense, splice site changes, frameshifts)
• Mutation should not be found in SNP databases or control genome sequences
• Same gene mutated in affected, unrelated individuals (and no unaffected individuals with the putative disease-causing genotype)

Question 12

What issues have caused prior clinical tests to fail?

Answer 12

• Causative variants in a gene not include in tested gene panel for amplicon sequencing / microarray
• Intron variants and exon deletions are difficult to detect with exome sequencing
• Non-coding RNA are not covered by exome sequencing
• Poor coverage in initial genome sequencing
• Novel triplet expansion disorder cannot be accurately estimated by short-read sequencing

Question 13

What might cause an NGS dead zone? How may they be recovered?

Answer 13

• Genes with high homology or repetitive regions
• Can sometimes be recovered using long-read NGS

Question 14

How can methylation status be used to detect cancers?

Answer 14

• Circulating cell-free DNA is released form dying cells; in healthy individuals, mainly released from hematopoietic cells whereas in cancers, high rate of apoptosis/necrosis
• Methylation patterns in cell-free DNA can indicate their tissue of origin (e.g. Galleri test -> enrichment through hybridisation capture and bisulfite DNA conversion and sequencing)