Human Genetic Diversity Flashcards
Definition of a race and ethnic group
Race: A vague, unscientific term for a group of genetically related people who share certain physical characteristics
Ethnic group: Population of individuals around an assumption of common cultural origin - common nationality or cultural tradition
How can we know whether races exist?
- Can individuals be reliably allocated into valid genetic clusters, in which all members share more recent common ancestry than members of other clusters
- If so, do descriptors such as race or ethnicity capture any of the differences between such clusters
- Are these differences meaningful for health-related (or other functionally relevant) variation between groups
How is skin pigmentation varies in areas and how it doesn’t mean populations are closely related - give example
- Skin pigmentation in Africans, South Asians, Semang and New Guineans: New Gunieans have very dark skin but are more closely related to Asians than Africans
- This may be due to similar selection pressures in tropics - maintaining certain ancestral traits or similar traits arising by convergent evolution
How can you apportion human diversity?
Using Fst: measuring allele frequency at population level
- Fst is proportion of the variance in allele frequency variation distributed among subpopulations relative to the total population
- Fst = 0 : no differentiation, random mating between pops
- Fst = 1 : complete differentiation, no gene flow
What did Lewontin find in his human genetic variation study?
- That most of diversity is within local population = ~85%
- Between local population within continents = ~8%
- Between continents = ~6%
Conclusions: - Most variation lies between individuals within populations, so that racial classifications have little basis in genetics of the populations concerned
- From genetic standpoint - distant racial groups do not exist
How does the STRUCTURE analysis work?
Is a Bayesian framework: infer pop structure in a random sample, without assuming the populations a priori
- STRUCTURE - identifies subgroups within multi-locus datasets with distinctive allele frequencies
- Identifies clusters which minimise deviations from HW equilibrium
- Model based - tests the fit of the data to models of ‘K’ populations
What were the conclusions of Rosenberg et al and more recent studies?
- At K = 5, the structure identified corresponds to broad geographic regions
- Quite a lot of blurring at boundaries, and assignments of any one individual often not stable when the level of structure was changed
- Within continents, divisions do not usually correspond to predefined population classifications - except in Oceania and Americas
- Among population variance lower than estimated in classical studies (<2%) - lowest in Europe - 0.7%
Has STRUCTURE rediscovered race?
NO - that is not what these studies have shown:
- e.g., genes from two individuals from the same region will be on average only 4% more similar than individuals from other regions - smaller than earlier estimates
- So population geneticists are reluctant to reinstate the race concept, even when there is evidence for continental structure
- Rosenberg et al., suggest that world-level genetic boundaries do mostly correspond to major physical boundaries - Sahara; Himilayas; various oceans
- Results demonstrate that human variation is geographically structured as a result of human dispersals - not the same as reviving the race concept
How are genetic diversity and distance from East Africa related?
Negative correlation between geographic distance from East Africa and genetic diversity - greater distance = lower diversity
- Is indicative of repeated founder events and loss of diversity associated with that
- Liu et al., 2006, Prugnolle et al., 2005
What was shown about genetic variation in clines (contiuous) or clusters (discontinuous)?
- Continental clusters are robust - but within these there is evidence for clinal variation in level of structure
- Rosenberg et al., 2005
How is mtDNA structured globally?
- mtDNA tends to be structured into continental or sub-continental domains, but with rather less discontinuity from population to population within continents - reflecting the hierarchical series of founder effects of the out-of-Africa dispersals
- mtDNA ‘domains’ seem likely to correspond to the clusters identified by STRUCTURE
What paper showed admixture in human pops?
Hellenthal et al., 2014:
- Assessed timing of admixure events based on rates of decay of linkage disequilibrium among SNP haplotypes along
- So can use amount and extent of LD to see how much admixture had occurred
- Shallow decay curve: more recent admixture
- Steep decay curve: older admixture
- Modelling of curve shape allowed disentangling of complex population histories involving multiple admixture events at different times among multiple populations
- e.g., complex histories of mixing in Eastern Eurpoe and central Asia at diff timescales - corresponds well with historical record
What was shown about admixture in history of UK?
- 500,000 SNPs typed in 2039 ‘native’ UK people and 6209 Europeans - looked for evidence of clusters and structuring
- Found very weak structuring in UK -very small Fst value (0.003) - so very well admixed population
- Genetic contribution to southeastern England from Anlgo-Saxon migrations <50%, much of remainder dervied from pre-Roman
- Non-Saxon parts of UK - there are genetically differentiated subgroups rather than general ‘Celtic’ population
- Found higher amount of contribution from Scandinavia in Orkney - example of Vikings
Why is it challenging to look at ‘population history’ for humans?
- Problems disentangling cultural factors from genetic factors
- Populations are not homogeneous, and are fluid overtime
- Culture changes much more quickly than genetic structure
- Each locus is sexually reproducing species has an independent history - so may be able to talk about histories and homelands for alleles or haplotypes but not populations/individuals
What was found about skin color and ancestry in Brazil?
Parra et al., 2002:
- Found high overlap in African Ancestry scores among skin colour groups
- Therefore all individuals and regions include African, European and Native American ancestry, regardless of color
- So, in Brazil - ‘color’ is a poor predictor of african ancestry
