This course is about VARIATION: its causes, effects, and history.

Size: px

Start display at page:

Download "This course is about VARIATION: its causes, effects, and history."

Christina Terry
5 years ago
Views:

1 This course is about VARIATION: its causes, effects, and history. For thousands of years, western thought had accepted the Platonic view that an object s ultimate reality was its essence or ideal type. In biology, essentialism gave rise to the assumption that species are held together by their underlying, unchanging "types or ideal forms. On this view, individual variations are departures from the essence of a species; thus they are imperfections that make individuals less representative of the true nature of their species. Darwin destroyed essentialism in biology and replaced it with a radical new idea: variationism. Variationism is the view that species are united only by recent common ancestry. Thus every individual is equally representative of the species; the average phenotype is just a statistical abstraction, not the reflection of some higher, more pure or more ultimate reality. But Darwin was not a radical in the modern sense. He didn t set out to overturn essentialism. Facts and logic led him there. Anth/Biol 5221, 22 and 24 August 2018

that genes and Darwin were incompatible. R.A.

2 What causes variation, and why is some of it heritable (kids resemble parents?) Darwin didn t know. Mendel s discovery of genes (1865) was rediscovered in 1900, and most of the early geneticists concluded that genes and Darwin were incompatible. R.A. Fisher ( ) invented the analysis of variance (ANOVA) in 1918 to show that Darwin s ideas about the inheritance of variation were consistent with Mendel s genetics..

3 Transactions of the Royal Society of Edinburgh 52, (1918)

4 What causes variation, and why is some of it heritable (kids resemble parents?) Darwin didn t know. Mendel s discovery of genes (1865) was rediscovered in 1900, and most of the early geneticists concluded that genes and Darwin were incompatible. R.A. Fisher ( ) invented the analysis of variance (ANOVA) in 1918 to show that Darwin s ideas about the inheritance of variation were consistent with Mendel s genetics. This is one of the deepest, most general and most transformative ideas in the history of human thought - and oddly, most invisible! Fisher remains an obscure nerd celebrated by almost no one other than statisticians (right). Meanwhile, ANOVA has become the foundation of statistical thinking and practice in industry, government and medicine as well as in science.

The classic quantitative trait, to which we will return (172.3 172.3) 2 = 0 std. dev. = sqrt(variance) (182.4 172.

5 The classic quantitative trait, to which we will return ( ) 2 = 0 std. dev. = sqrt(variance) ( ) 2 = 102 sqrt(102) = 10.1 mean = average or expected height variance = average or expected squared deviation from the mean ( ) 2 = 470.9

$The central question answered by ANOVA: What fraction of the total variance is explained by or caused by sex?$

6 The female and male means differ bigly. Each sex has a smaller variance (s.d.) than the population as a whole. The central question answered by ANOVA: What fraction of the total variance is explained by or caused by sex? (In this case, by the difference between the female and male means?

$694 V(within) = 48.286 = 0.531*46.96 + 0.469*49.78 V(among) = 53.408 = 0.531*(165.6-172.5)^2 + 0.469*(180.2-172.5)^2 V(total) = 101.694 fraction explained by sex = 53.408/101.$

7 Heights of 252 women and 223 men in the Utah Genetic Reference Project 252 females : M = V = males : M = V = total : M = V = V(within) = = 0.531* *49.78 V(among) = = 0.531*( )^ *( )^2 V(total) = fraction explained by sex = / = 0.53

8 What about the effects of genes? (That is, of families?) Most of the individuals are siblings in 36 families with 1-12 sons and 1-12 daughters. 90% of the remaining variance is explained by genetic differences among the families. And 10% by effects of the environment (what remains after the effects of sex and of genes have been removed statistically). These people grew up in a very healthy and uniform environment (20 th -century Utah). In other times and places, the split tends to be 80/20 or even 70/30. For other traits, in most species, it may be anywhere from 80/20 to 20/80.

9 The families disaggregated

10 Summary about phenotypic variation Every quantitative phenotype you can think of varies. Often the distributions are roughly normal. If some of this variation is heritable, then evolution by natural selection is inevitable (Darwin s world-changing insight). Fisher invented ANOVA to show that darwinian evolution of quantitative traits is compatible with mendelian genetics. (If many genetic loci make small, independent contributions, and so does the environment). This year (October 1, 2018) is the paper s 100 th anniversary. It + Darwin changed how we think about variation. Now the variance can be partitioned into contributions associated with factors that explain the total. Often (but not always) we can interpret the factors as causes. For example, genes and environment. ( Sex?)

11 Why do some slides have this leaf as the background? Because my other course studies aspens at Silver Lake!

12 Tree 9-8 belongs to Clone 9, and Tree 13-9 belongs to Clone 4-6

13 c8 c9 c2a c1 c12 c5a c4-6 c7 c3 c14

15 Genomes and their variation Informal Pop Quiz! How big is the haploid human genome? 3 million base pairs (Mbp) 30 Mbp 300 Mbp 3 billion bp (Gbp) 30 Gbp How many functional genes does it contain? 20,000 40,000 60,000 80, ,000

16 Genomes and their variation What fraction of nucleotide positions are heterozygous in a typical outbred individual? 1/2000 (0.0005) 1/1000 (0.001) 1/500 (0.002) 1/200 (0.005) 1/100 (0.01) Is this more or less heterozygosity than is typical for mammals? More Less Is the heterozygosity within an individual the same thing as the average genetic difference between members of the population? Yes, the same number, for the same reason. No, not the same thing at all.

Genomes and their variation Phenotypic variation arises from genetic and environmental variation. Both are usually major contributors, and each influences the other.

17 Genomes and their variation Phenotypic variation arises from genetic and environmental variation. Both are usually major contributors, and each influences the other. The genetic variation is encoded by DNA sequence variation. What does DNA sequence variation look like? How can we best describe and analyze it? How is it created and maintained? How do genomes evolve? What can we infer about those processes, and about population history, from quantitative analysis of present-day variaton?

18 How does a chimp differ from a human? Better hair! Fewer self-doubts! 40,000 amino-acid substitutions! (~20,000 on each line of descent) Assuming 20,000 proteins, the difference is around 2/protein, or roughly 1/250 amino acids. ~18,000,000 nucleotide substitutions! Given 3x10 9 base pairs per haploid genome, that s 1/167 base pairs, or 0.6% different (99.4% the same). How did these nt and aa substitutions (and other genomic changes) occur on the chimp and human lineages, while billions of others did not occur? And why did they take 5 myr?

Substitution is a population process An evolutionary substitution occurs when: (1) Mutation creates a new allele. (Initially there is only one copy of this allele in the population.

) Whether the new mutation is lost from the population, or eventually fixed, is determined during the polymorphic phase (2), which may last a long time.

19 Substitution is a population process An evolutionary substitution occurs when: (1) Mutation creates a new allele. (Initially there is only one copy of this allele in the population.) (2) The new allele increases in frequency (briefly increasing genetic variation or polymorphism ). (3) The old alleles are lost. (The population then contains only descendants of the new allele.) Whether the new mutation is lost from the population, or eventually fixed, is determined during the polymorphic phase (2), which may last a long time. Thus present-day variation and long-term evolution are connected through the population processes that change allele frequencies. The most important of these processes or forces are genetic drift and natural selection. During the next few weeks we will develop a general quantitative framework for thinking about genetic variation at the DNA sequence level, and about rates of allele frequency change under the combined influences of drift and selection. Photo credits: Curtis Orchard (upper), Philip Cunningham (lower)

20 The most-studied molecular polymorphism: alcohol dehydrogenase (Adh) in Drosophila melanogaster + AAG FF AAG ACG FS SS Lysine Threonine slow fast (S) (F) - From FM Johnson & C Denniston (1964) Nature 204,

22 Adh is polymorphic in D. melanogaster populations all over the world. What is the allele frequency of adh F in Miami? p(f) = (1)(1/110) + (½)(24/110) + (0)(85/110) or = 1/ /110 = 13/110 = p(f) = 2/ /220 = 26/220 = 13/110 = What is the heterozygosity? H = 24/110 = What is the expected heterozygosity? E(H) = 2pq = 2(0.118)(0.882) = 0.208

E(heterozygosity) is the natural way to measure genetic variation. Why? Because it s the expected variance of allele numbers per fly! How many Fast alleles does a typical Miami fly have?

23 E(heterozygosity) is the natural way to measure genetic variation. Why? Because it s the expected variance of allele numbers per fly! How many Fast alleles does a typical Miami fly have? mean(#f) = (2)(1/110) + (1)(24/110) + (0)(85/110) = = 2p var(#f) = ( ) 2 (1/110) + ( ) 2 (24/110) + ( ) 2 (85/110) = And what did we expect (given the observed allele frequencies)? E[var(#F)] = (2-2p) 2 (p 2 ) + (1-2p) 2 (2pq) + (0-2p) 2 (q 2 ) = 2pq = 2(0.118)(0.882) = This is exactly the same as the expected heterozygosity (2pq)! And var(#f) = var(#s), as long as we are considering just two alleles. The variance and the heterozygosity are both highest when p = q = ½.

24 Exercise: Calculate the frequencies of the three alleles in this sample of alkaline phosphatase genotypes from humans. Then calculate the expected genotype frequencies, which are shown in the last column. What is the expected heterozygosity? At what allele frequencies would it be highest?

that were surveyed at several to many loci.

25 Enzyme polymorphisms are abundant at many loci and in all species. These histograms show the average enzyme heterozygosities of individuals in hundreds of species that were surveyed at several to many loci. A typical species is polymorphic at somewhere between one third and one half of its loci. A typical individual is heterozygous at 4-15% of its loci. That s a lot of genetic variation!

26 But there s much more at the DNA level! Marty Kreitman sequenced 11 alleles of Adh for his Ph.D. thesis project in the early 1980s. The Fast alleles are slightly more similar to each other, on average, than the Slow alleles. But all 11 alleles differ from each other, at least by an insertion/deletion difference. On average they differ by about 15 nucleotide differences (0.6%). The average per-base-pair nucleotide difference or heterozygosity is called π. For these Adh sequences, π =

27 How should we represent genetic variation? There s no best way. It depends on the question! K slow T fast

28 All of the polymorphic sites in color Slow (lys) T G C A Fast (thr)

29 Implication: The genetic differences between species are evolutionary substitutions caused by the fixation of polymorphisms within species. In other words, the differences between species are just like the differences within them, only more so. If this is true, then to understand how evolution works, we need to understand the processes that generate and shape genetic polymorphism. K The big three: mutation, drift and selection. (Also recombination, migration, etc.)

30 And the differences within species are just like those between them, only less so! SNPs in the lactase gene region on chromosome 2, sampled from Utahns of European ancestry. The consensus (majority) nucleotides are shown at the top ( cons ). The first 26 chromosomes have exactly this sequence (as do 60 others which were omitted to save space). The others are shown as their differences from the consensus. These 101 variable sites are embedded in a region of roughly 140,000 base pairs. cons aaggaggcgacattccgcttcaggcattcctatctaaacagaccaacgtaagggtacaatgcctaacccagacgtttcaactctggctgttattcctcgat t t c g gg.a.ca.ag.g.gt G G G G G G..a.gt...t...gac.c.tgtct...a..gc..t.t..c 38...c...ccgga...gat..at..gg..c...tc.gGaaa.g..ccttt...tg...c...t.t...g...t g...g..c...ccgga...gat..at..gg..c...tc.gGaaa.g..ccttt...tg...c...t.t...g...t gg.a..at.gt.c.t..tcc...agtag.t.cat..g...t..ttccgg..a.gt...t...gac.c.tgtct c.t..tcc...agtag.t.cat..g...t.gttccgG..a.gt...t...gac.c.tgtct...a..gc..t.t..c 42...c.t..tcc...agtag.t.cat..g...t.gttccgG..a.gt...t...gac.c.tgtct...a..gc..t.t..c 43..aa..at.gt.c.t..tcc...agtag.t.cat..g...t.g.tc.gG..a.gt...t...gac.c.tgtct...g...t.t..c 44..aa..at.gt.c.t..tcc...agtag.t.cat..g...t..ttc.gG..acgt...t...gac.c.tgtct...a..gc..t.t..c 45..aa..at.gt.c.t..tcc...agtag.t.cat..g...t.gttc.gG..a.gt...t...gac.c.tgtct...a..gc..t.t..c 46...g...g..c...ccgga...gat..at..gg..c...tc.gGaaa.g..ccttt...tg...cg.gt.t..ctata.ccg.c..ctcg. 47 gg.a..at.gt.c.t..tcc...agtag.t.cat..g...t.gttccgg..a.gt...t...gac.c.tgtct...a..gc..t.t..c 48 gg.a..at.gt.c.t..tcc...agtag.t.cat..g...t.gttccgg..a.gt...t...gac.c.tgtct...a..gc..t.t..c 49 gg.a..at.gt.c.t..tcc...agtag.t.cat..g...t.gttccgg..a.gt...t...gac.c.tgtct...a..gc..t.t..c 50...g..c..tatccgga...g.tc.atcgg.tc.g.tg.tc.gG..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 51 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ct...a..gc..t.t..c 52 gg.a.ca.ag.g.gtta.ccgga...g.t..atc.g.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 53 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 54 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 55 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 56 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 57 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggtg..cg.gt.t..ctata.ccg.c..ctcg. 58 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.ca.ctcg. 59 gg.a.ca.ag.g.gtta.ccgga...g.tc.atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 60 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggtg..cg.gt.t..ctata.ccg.c..ctcg.

31 Infant mammals all need to digest lactose, but adults normally don t. got LCT-13910*T?

32 LCT region and putative lactase persistence mutations intron 13 Ingram et al. (2009) Human Genetics 124:

33 Homozygous intervals of chromosome 2 in 90 Utahns LCT-13190*T 2.4 Mbp 19 genes

34 How should we talk about (describe) this variation? aaggaggcgacattccgcttcaggcattcctatctaaacagaccaacgtaagggtacaatgcctaacccagacgtttcaactctggctgttattcctcgat t c g gg.a.ca.ag.g.gt G G..a.gt...t...gac.c.tgtct...a..gc..t.t..c 38...c...ccgga...gat..at..gg..c...tc.gGaaa.g..ccttt...tg...c...t.t...g...t g...g..c...ccgga...gat..at..gg..c...tc.gGaaa.g..ccttt...tg...c...t.t...g...t gg.a..at.gt.c.t..tcc...agtag.t.cat..g...t..ttccgg..a.gt...t...gac.c.tgtct c.t..tcc...agtag.t.cat..g...t.gttccgG..a.gt...t...gac.c.tgtct...a..gc..t.t..c 43..aa..at.gt.c.t..tcc...agtag.t.cat..g...t.g.tc.gG..a.gt...t...gac.c.tgtct...g...t.t..c 44..aa..at.gt.c.t..tcc...agtag.t.cat..g...t..ttc.gG..acgt...t...gac.c.tgtct...a..gc..t.t..c 46...g...g..c...ccgga...gat..at..gg..c...tc.gGaaa.g..ccttt...tg...cg.gt.t..ctata.ccg.c..ctcg. 47 gg.a..at.gt.c.t..tcc...agtag.t.cat..g...t.gttccgg..a.gt...t...gac.c.tgtct...a..gc..t.t..c 50...g..c..tatccgga...g.tc.atcgg.tc.g.tg.tc.gG..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 51 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ct...a..gc..t.t..c 52 gg.a.ca.ag.g.gtta.ccgga...g.t..atc.g.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.c..ctcg. 57 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggtg..cg.gt.t..ctata.ccg.c..ctcg. 58 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggt...cg.gt.t..ctata.ccg.ca.ctcg. 60 gg.a.ca.ag.g.gtta.ccgga...g.t..atcgg.tc.g.tg.tc.gg..a.g.g...tg...ggtg..cg.gt.t..ctata.ccg.c..ctcg. The A in the central position (T on the other strand) is the mutation that causes the lactase gene to remain on in adulthood, conferring tolerance to the consumption of lactose (milk sugar).

35 The vocabulary (-ies) of homologs: a muddle, beware! vocabulary system: Position on chromosome locus locus locus Protein- or RNA-coding locus gene gene gene 1 of 2 or more sequence variants at locus allele allele allele Physical instance of the DNA at a locus gene allele gene copy Gillespie favors system 2, but we think system 3 is clearer because it distinguishes all four aspects of gene-iness Orthologous genes occupy the same locus in different species (i.e., they are separated by a speciation event). Paralogous alleles occupy different loci in the same or different species (i.e., they are separated by a gene-duplication event). Alleles, orthologs and paralogs are all homologs, because they descend from a common ancestral sequence.

36 How MUCH human variation is there, and how is it distributed? How different are your maternally and paternally inherited genomes? (π)(3x10 9 ) = (0.0008)(3x10 9 ) = 2,400,000 bp het That s more than 10% of the difference between chimps and humans! And we re much less variable than most species! Only a small fraction of these differences cause phenotypic differences, because only a small fraction are in genes. But that s still a very large number of functionally meaningful differences! xome bp het bp examined pi (x10-4 ) 1 71,483 92,639, , ,060, ,190 81,359, ,922 74,162, ,344 77,924, ,864 72,380, ,010 68,527, ,477 57,476, ,329 50,834, ,040 52,184, ,477 56,680, ,607 51,160, ,250 43,915, ,083 47,425, ,847 31,682, ,994 27,736, ,247 27,124, ,711 30,357, ,499 15,060, ,726 31,795, ,415 50,367, ,469 20,478, X 23,818 50,809, Y 348 2,304, Total 920,752 1,225,448,

Big picture from genome sequencing (e. g., 1000 Genomes Project) Individual nucleotide heterozygosities (π) are 0.0005 0.001 (1.5 3 M/genome) Higher in Africa, lower in Eurasia.

37 Big picture from genome sequencing (e. g., 1000 Genomes Project) Individual nucleotide heterozygosities (π) are (1.5 3 M/genome) Higher in Africa, lower in Eurasia. Some cause amino-acid polymorphisms in proteins (10-20 K/genome, ~0.5/protein!) Many insertion-deletion polymorphisms ( indels of 1-50 bp, ~550 K/genome) Many loss-of-function (LOF) mutations (~150/person, ~20 in known disease genes) Many copy-number variants ( CNVs of 2 kb 2 Mb, ~200/genome) But allele-frequency differences between populations are modest: For typical loci, ~85% of the world-wide variation is within local populations. Only 10-15% is accounted for (added) by differences between major groups. Long-standing question: Does this mean the races are imaginary?

39 Genetic distances among 467 individuals: 10 SNPs

40 Genetic distances among 467 individuals: 100 SNPs

41 Genetic distances among 467 individuals: 1000 SNPs Europeans E.Asians Africans Indians

42 Genetic distances among 467 individuals: 10,000 SNPs E.Asians Europeans Africans Indians

43 Genetic distances among 467 individuals: 261,000 SNPs CEU Europeans CHB E.Asians JPT Indians Africans YRI

44 Summary about genomic variation Heterozygosity (expected) is the natural measure of genetic variation. It can be described at many levels, for example: individual nucleotide or amino-acid sites whole genes or chromosome segments (sequence, or presence/absence) Per-site heterozygosities are low: (p for humans). But staggering per genome: ( x 3x10 9 = 2.4x10 6 per person). Most genomic variation is not in genes, so probably meaningless. Even so, the part with phenotypic effects is absolutely enormous. Per-site, most human variation (>85%) occurs within local populations. Less than 15% is explained by differences among continental races. And there are no diagnostic (fixed) differences among races. However, allelic states are not independent among loci. So per-genome, long-separated populations may cohere statistically.

GENETICS - CLUTCH CH.1 INTRODUCTION TO GENETICS.

GENETICS - CLUTCH CH.1 INTRODUCTION TO GENETICS. !! www.clutchprep.com CONCEPT: HISTORY OF GENETICS The earliest use of genetics was through of plants and animals (8000-1000 B.C.) Selective breeding (artificial selection) is the process of breeding organisms