Hunting for genetic risk variants for psychopathology. Flashcards
Outline process, findings and limitations of linkage mapping.
Simplest case scenario. Analyse co-segregation of DNA markers w/ presence of disorder in question. Effective in revealing single gene causes of rare, fully penetrant neurological disorders (e.g. HD) and rare sub-forms of disorder (AD, PD, MS, FTD…) However, no single gene is fully penetrant/Mendelian in its role in major psychiatric disorders has been detected to date.
Case-control outline, pros and limitations.
Trawling for gene risk in a bigger pool, not looking in families but in populations (looking for significant enrichment of a particular genetic variant in cases than in controls? Some kind of genotyping analysis + tests of association (Fishers EP, Chi square, etc.) + odds ratio (increased risk conferred by variant.
- Useful bc. relatively inexpensive type of epidemiological study that can be conducted by small team/individual researchers in single faciliteis.
- However, observational, so less control. May be more difficult to establish timeline of exposure to disease outcome compared to, say, prospective cohort studies. Also, difficult obtaining reliable information about background, exposure status, etc.
Candidate-led approach, form of case-control study.
Based on a priori knowledge of biological functional impact on trait/disease in question, ergo limited by such reliance, results in info bottleneck, esp. for complex disorders –> hypothesis driven. . Largely unsuccessful, failure to replicate: small sample-sizes, underpowered, sampling/genotyping errors/sample ascertainment (population stratification: genetic makeup varies between ethnic/geographic populations. If case & control pops not well matched for ethnicity or geographic origin, then false positive associations can occurs because of confounds from population stratification. Also inefficient for rare exposures.
GWAS: Hypo-free, brute force, like other studies (linkage disequilibrium)of alleles at 2 assoc.(close) loci. Systematic interrogation, so independent of a priori. Requirements. “Tag SNP” haplotypes.
Unbiased/hypothesis-free brute force approach (screen variation across whole genome). Based on linkage disequlibrium (like other assoc. exps), i.e. a non-random stat assoc. of alleles @ 2 or more loci, which is characteristically assoc. w/ short physical distance genetic markers. GWAS systematically interrogates entire genome, making experiment independent of a prior hypothesis, this is v. important for psychiatric disorders while knowledge of pathophysiology (thus ability to choose HQ candidate genes) v. limited. Requires huge samples, and generous sharing of data between organisations, high density maps of geetic variability between individuals in human pops, vast majority in form of SNPs, & Tag SNPs (impossible to have every SNP variant arrayed; instead index SNPS capitalise on genetic linkage in haplotypes, i.e. stretches of DNAs physically linked together, so inherited together) necessary to capture most of SNP genetic variability between individuals to be arrayed on to DNA chips, allowing high throughput comparisons of allelic frequencies in cases & controls.
Implications of GWAS findings? (Genetic liability, polygeny)
Highly likely we all carry some genetic liability for psychopathology in the form of multiple, common risk SNPs that, singly, are not very damaging but when accumulated could provide significant risk – known as our “Polygenic risk” (sometimes “common risk” is used as a shorthand term).
Categorising risk variants: SNPs vs. CNVs (psychopathology, <1 in 500; de novo, disorder examples)
Categorised into: SNPs (low-pen common; complex genetic architecture, difficult for GWAS), low-pen, rare (question of utility in finding these, difficult to detect), CNVs (high-penetrance, rare; simple genetic architecture, psychopathology-associated CNVs often inherited, but can spontaneously arise de novo, mainly during meiosis. These tend to be most damaging- evolutionary exp. None are fully penetrant. Then there are common, high pen variants which are atypical of common diseases due to natural selection. Exception is APOE, since AD arises later in life after reproduction.
Limitations of GWAS:
- Effect sizes of genetic variants (3% variation in height).
- Not entirely agnostic
- Mostly suitable for common (5% freq.), low-moderate risk variants Rare variants w/ frequencies lying between limits of deleterious mutations and polymorphic variations (i.e. 0.1%-15) not examined (Autism, Burmeister, McInnis, Zollner, 2008). Candidate led approaches also not suitable, but identifying rare variants on a whole-genome basis previously not feasible, but then 1000 genome project (example of WGS, just one application of NGS).
o Important to recognise that GWASs do not represent a panacea for the challenges of gene identification in behavioural genetics. Existing GWASs suggest that the effect of any single genetic variant on complex phenotypes is likely to be small, much smaller than originally anticipated. For example, a GWAS of adult height found significant associations with 20 different genetic variants, collectively accounting for only 3% of phenotypic variation in height. GWASs of behavioural traits will require massively large samples.
o GWA Studies are not entirely “agnostic”. Labelling them as hypothesis-free may disregard the fact that the output of any biological experiment is primarily determined by the extent to which the hypotheses tested truly hold. GWAS are based on a priori hypotheses, dictated by the design of genotyping platforms or the analysis methodologies.
o Also, the GWAS approach is mostly suitable to the detection of common variants conferring low/moderate risks. Most GWAS analyse SNPs with minor allele frequencies of >5%, Rare variants with frequencies lying between the limits of deleterious mutations and polymorphic variations (i.e. 0.1-1%) are not examined by current GWAS (due to sample size and sequencing constraints). There is growing evidence that the genetic diathesis for some behavioural disorders, most notably autism, appears to be largely attributable to the effects of multiple rare genetic mutations (Burmeister, McInnis, & Zollner, 2008). Rare variants are more likely to be detected by extensive resequencing of carefully selected candidate genes in relatively large numbers of carefully chosen cases, together with a thorough analysis of the functional effects of any suspected variants (Bodmer & Bonilla, 2008). However, candidate-led approaches are probably not sufficient to identify important rare variants and sequencing on a whole-genome basis is required, this was not feasible in the past, but technical advances, such as the 1000 genome project (The 1000 Genomes Project Consortium, 2015; which provided an unparalleled insight into human variation on a population level, a compendium of in-depth data on variation across 2504 individuals from 26 cross-continental populations, including data on rare variants. This is example of WGS, just one of the applications of NGS.
NG/ultra-high throughput sequencing: key difference from DNA chips. Early exome sequencing (Iossifov et al., 2014; 2500 simplex families ASD, de novo missense/disrupting mutations and CNVs c.30% cases). Use of NGS.
o Next generation sequencing (NGS) also known as ‘deep’/ultra-high throughput sequencing – offers significantly better resolution than ‘DNA chip’ technology, therefore good for revealing rare variants; key difference NGS does not rely on arraying index-SNPs as in DNA chip methods, instead can directly screen the genome.
o Early work using NGS (mainly for cost reasons) was limited to sequencing the coding regions of genes (known as exome sequencing), now NGS much cheaper so association studies can employ ‘whole genome sequencing’ – where we can directly screen the entire genome ‘nucleotide base by nucleotide base’ for variants associated with psychopathology.
Exome sequencing still provided great results. Iossifov et al. (2014) sequenced more than 2500 simplex families, each with a child who was diagnosed with ASD, and they discovered de novo missense mutations, de novo gene-disrupting mutations and CNVs in ~30% of all cases.
o NGS detecting further rare variants of high penetrance for psychopathology:
Additional rarer, smaller CNVs
SNV (single nucleotide variants – i.e. point mutations, not to be confused with SNPs which are common and of low penetrance for disorder)
InDels (insertions or deletions of very short sequences)
