Evolutionary perspective of genetic variants, and pleiotropy.

Question 1

What is fecundity?

Answer 1

The reproductive success of an individual.

Question 2

How does fecundity relate to the natural selection of rare vs. common genetic variants (Uher, 2009).

Answer 2

The proposition is that low penetrant SNP variants become common in the population because they are (individually) less damaging to fecundity than high penetrant variants… which are, as a result, rare in the population (Uher, 2009).

Question 3

How was the fecundity prediction tested in SCZ,

males 75%, women 50%)(Rees et al., 2011

Answer 3

o Idea tested directly in the case of schizophrenia; affected individuals did indeed show a marked decrease in fecundity, especially in males with a 75% reduction from an average, women a 50% reduction, reasons for this are very interesting (Rees et al., 2011). Due to it being the women who carries and chooses to have the child, not the man.

Question 4

How does the fecundity argument predict high penetrance of de novo variants (i.e. not present in either parent’s somatic cells). What does it predict about the age and stability of SNPs?

Answer 4

 damaging variants are rare because they markedly increase risk for psychopathology - in turn markedly reducing fecundity - in turn reducing the chances of the genetic variant becoming established in the population
 because of the strong negative selection pressure tending to eliminate them, any rare damaging variant is likely to have entered the population relatively recently
 pending removal by natural selection rare variants can have a large effect on risk for disorder, this is consistent with the finding that de-novo rare variants (not present in the parents but newly arisen) show particularly high penetrance for disorder - since they are unfiltered by natural selection. They’re “newly minted”, not in the somatic cells of the parents, but they are in the children- not subject to selection pressures.
 in order to maintain some degree of frequency in the population against strong negative selection, the damaging variants must be continually replenished (otherwise they would disappear) – some evidence for this idea from analysis of rate of de-novo CNVs identified in schizophrenia.
 these same selection arguments would suggest that the common (individually) low penetrant SNP variants are likely to be in comparative terms much older (quite ancient possibly) than rare variants, and a more stable presence in the population.

Question 5

Define “missing heritability”, SCZ example: c.60-80% -> c.30%.

Answer 5

 The apparent mismatch between heritability estimates, and the actual DNA differences (odds ratio) between well and ill people in population. At present, cumulative increase in odds-ratio (i.e., cumulative increased risk) stemming from genetic risk variants present in the general population amounts to less than that which should be there from heritability estimates, e.g., for highly heritable schizophrenia (60-80% heritability estimate) all the current known risk variants amount to about 30% of the heritability.

Question 6

Explanations of missing heritability

• flawed method = inflation
• poor tagging by single GWAS marker, or independent causal variants at same locus (one louder than others, hiding other contributions; Gusev et al., 2014), thus Owen (2014) technological limitations.
• Winner’s curse (mostly GWAS),
• Mixture
• Pleiotropy

Answer 6

 It could be that the heritability estimates are inflated (e.g., by flaw in DZ/MZ twin designs).
 It could be that the genetic liability is there at the molecular level, we just have not found it yet, or that the known GWAS loci is not fully quantified by any single marker. This maybe happen if the true causal variant is poorly tagged by any single GWAS marker, or if multiple independent causal variants exist at the locus (in which case the variance would be explained by the most-significant marker, hiding other local contributions that would explain at least some of the “missing heritability” (Gusev et al., 2014).. in this regard, Owen (2014) argues that most of the missing genetic signal is merely hidden by the limitations of current genotyping technology and size of samples. The power of studies (sample size) and technologies are getting better.
 Propensity for the “Winner’s Curse”
• The Winner’s Curse in genetic association studies appears as an upward bias in the estimated effect of a newly identified allele on disease risk when the study design lacks sufficient statistical power. This manifests mostly in GWAS in which 300,000-1,000,000 single-nucleotide polymorphisms are tested.
 It could be a mixture of the above and/or mechanisms that contribute to heritable risk but are not picked up in standard genotyping assays; epigenetic changes.
 Or pleiotropy.

Question 7

Define pleiotropy.

Answer 7

The phenomenon of one gene variant affecting more than one phenotypic characteristic.

Question 8

What are the implications of pleiotropy for clinical disorder classification & diagnosis?

• blurred boundaries
• X-disorder genetic effects, e.g. generalised epilepsy. reflect poor knowledge of underlying biology.
• Multiplication of diagnostic categories –> sub-cats.
• Cutoff in treatment points difficult to establish,

Answer 8

o Many of the same genetic variants confer risk across multiple diagnostic categories; holds for both common and rare variants (Sullivan et al., 2013). Significant overlap for schizophrenia in many disorders, esp. bipolar at the molecular level. Interest in symptom overlap. Extent to which anorexia nervosa and OCD overlap molecularly is very interesting in terms of the psychopathological manifestations of both disorders.
o Implication for clinical disorder classification & diagnosis.
 Well-known at the symptom level that there is a lot of overlap and blurred boundaries between disorders.
 Cross-disorder genetic effects are widespread, extending to generalised epilepsy (increased risk for epilepsy is a common co-morbidity with psychiatric disorder). It is becoming increasingly clear that diagnostic categories used in the clinic (guided by DSM/ICD) map poorly on to the underlying biology.
 Diagnostic categories within manuals have been multiplying, due largely to recognition of overlap between conditions at the level of symptoms.
 Leading to increasing sub-categories of condition (and progressively bigger and bigger diagnostic manuals).
 Increasing recognition of symptom overlap is consistent with overlap at the biological level.
 This progressive increase in recognition of symptom complexity is reflective of the biology. Difficult for clinicians to establish cut off points for treatment and patient reassurance, but this is not a faithful reflection of the biological underpinnings of these disorders.
 People initially thought that the biological research is too immature to include biological markers in psychiatric guidelines such as the DSM, BUT debate about need for a radical (and highly controversial) re-appraisal of how disorders are classified and need for new approaches that recognise the continuous nature of psychopathology; and the need for diagnosis and treatment to be informed by biology as embodied in the Research Domain Criteria (RDoC) initiative led by NIMH in the USA.

Question 9

Pleiotropy: anorexia OCD ovrlap.

Answer 9

Extent to which anorexia nervosa and OCD overlap molecularly is very interesting in terms of the psychopathological manifestations of both disorders.