CHAPTER OUTLIN 28.1 Overview of Complex and Quantitative Traits 28.2 Statistical Methods for Evaluating Quantitative Traits 28.3. Polygenic Inheritance 284 Identification of Genes That Control Quantitative Traits 285 Heritability 286 Selective Breeding Domesticated wheat, The color of wheat ranges fom a dark re to white, which isan example ofa complex or quantitative trait. COMPLEX AND QUANTITATIVE TRAITS In this chapter, we will examine complex tralts—characteris- tics that are determined by several genes and are significantly influenced by environmental factors. Most of the complex traits ‘that we will consider ate also called quantitative traits because they can be deseribed numerically. In humans, quantitative traits include height, the shape of the nose, and the rate at which we ‘metabolize food, to name a few examples. The field of genetics ‘that studies the mode of inheritance of complex and quantitative traits is called quantitative genetics. ‘Quantitative genetics is an important branch of genetics for several reasons, In agriculture, most ofthe key characteristics of interest o plant and animal breeders aze quantitative traits, These include traits such as weight, fruit size, resistance to disease, and the ability to withstand harsh environmental conditions. As we “will see, genetic techniques have improved our ability to develop strains of agriculturally important species with desirable quanti tative traits, In addition, many human diseases, such as asthma and diabetes, are viewed as complex traits because they are in- fluenced by several genes. Quantitative genetics is also impor- lant in the study of evolution. Many of the traits that allow a species to adapt to its environment are quantitative, Examples include the swift speed of the cheetah and the sturdiness of tree branches in windy climates. The importance of quantitative traits in the evolution of species will be discussed in Chapter 29, In this chapter, we will examine how genes and the envieon- ‘ment contribute tothe phenotypic expression of quantitative traits. ‘We will begin with an overview of quantitative traits and the use of statistical techniques to analyze them. We will then look at ex- perimental ways (0 identify quantitative wait loci (QTLS)— locations on chromosomes containing genes that affect the ‘outcome of quantitative traits. Advances in genetie mapping strat- ‘egies have enabled researchers to identify these genes. Last, we ‘will discuss heritability, whichis a measure ofthe amount of phe- ‘notypic variation in a population due to genetic variation, and consider various ways of calculating and modifying the genetic variation that affects phenotype. 28.1 OVERVIEW OF COMP! AND QUANTITATIVE TRAIT Learning Outcomes: 4. Ust examples of complex and quantitative tat, 2. Explain how quantitative tats may be described with a ‘frequency cistibtion. ‘When we compare characteristics among members of the same species, the differences may be complex or quantitative rather than 07708 CHAPTER 28 © COMPLEX AND QUANTITATIVE TRAITS eye eee TypeofTrat Examples ‘Anatomical Height weight number ofbistes in Drasophilaearlength welts Inco, and degre of pigmentation in fowers and skin Physiological Metabolic tas, speed of running and fight, ably to eats Withstand harsh temperatures, and milk production in mammals Behavioral Mating cals courtship tas, ably tolearm a maze, teats and abit to grow o move toward ght Diseases Predisposition toward hear disease, hypertension, cancer diabetes, eth, and ails qualitative. Humans, for example, have the same basic anatomical features (two eyes, two ears, and So on), but they differ in quantita- tive ways. People vary with regard to height, weight, the shape of {acial features, skin and hair pigmentation, and many other charac- teristics. As shown in Table 28.1, quantitative traits ean be eatego- rized as anatomical, physiological, or behavioral. In addition, ‘many human diseases are viewed as complex traits because they are influenced by many genes and environmental factors, Three of the leading causes of death worldwide—heart disease, cancer, and diabetes —are considered complex tats, Number 1 91529972467 4 ® Height inners 2422 Quantitative traits are described numerically. Height and ‘weight can be measured in centimeters (or inches) and kilograms (or pounds), respectively. The number of bristles on a fruit fly's, body can be counted, and metabolic rate can be assessed as the amount of glucose burned per minute. Behavioral traits can also be ‘quantified, A mating call can be evaluated with regard to its dura tion, sound level, and pattern. The ability to learn a maze can be described as the time and/or repetitions it takes to master the skill, Quantitative tats, such as height and weight, ae viewed as ‘continuous traits—teuts that do not fall into discrete categories ‘Some, such as bristle number in Drosophila, are meristic traits— traits that can be counted and expressed in whole numbers. Finally, certain diseases, such as diabetes, are viewed as complex traits, ‘even though the disease itself can also be considered qualitative either you have it or you don’, Even so, diseases have quantitative characteristics, such as their degree of severity. For example, some people with diabetes require low doses of insulin to prevent ad verse symptoms, whereas others require higher doses. For diabe- tes, heart disease, and cancer, the alleles of several different genes contribute to the likelihood that an individual will develop the disease, A certain threshold must be reached in which the number ‘of disease-causing alleles results inthe development ofthe disease ‘These are referred to as threshold traits—traits that are inherited duc tothe contributions of many genes, Many Quantitative Traits Exhibit a Continuum of Phenotypic Variation ‘That Follows a Normal Distribution In Part I ofthis text, we discussed many traits that fall into discrete categories. For example, fruit flies might have white or red eyes, and pea plants may produce wrinkled or round seeds. The alleles thal govern these traits affect the phenotype in a qualitative way. In analyzing crosses involving these types of traits, each offspring ean be put into «particular phenotypic category, Such attributes are called diseontinuous traits. In contrast, quantitative traits show a con- ‘inwum of phenotypic variation within @ group of individuals, For such waits, it may be impossible 0 place organisms into a discrete phenotypic clas. For example, Figure 28.8 is a photograph show- ing the range of heights of 82 college students Though height is found at minimum and maximum values, the range of heights between these values is fhlsly continuous. FIGURE 28.1 Normal distribution of «quan. titative trait (a) The distribution of heights in 82 col lege students (b) A frequency distribution forthe heights of students shown in (2) ‘CONCEPT CHECK: Is eight a esconinuous (sere ra, ‘or doe olow a continaum?282 STATISTICAL METHODS FOR EVALUATING QUANTITATIVE TRAITS 709 How do geneticists describe traits that show a continuum of phenotypes? Because most quantitative waits do not naturally fall, nto a small number of diserete categories, an alternative way 10 describe them is with a frequency distribution. To construct frequency distribution, the trait is divided arbitrarily into a number ‘of convenient, discrete phenotypic categories. For example, in Figure 28.1, the range of heights is partitioned into 1-inch inter- vals. Then a graph is made that shows the number of individuals found in each of the categories. Figure 28.1b shows a frequency distribution forthe heights ‘of students pictured in Figure 28.1a, The measurement of height is plotted along the x-axis, and the number of individuals who ex- hibit that phenotype is plotted on the y-axis, The values along the -waxis are divided into the discrete I-ineh intervals that define the phenotypic categories, even though height is essentially continu- ‘ous within a group of individuals. For example, in Figure 28.12, 9 students were between 65.5 and 66.5 inches in height, which is plotted as the point (66 inches, 9 students) on the’ graph in Figure 28.1b. This type of analysis can be conducted on any group ‘of individuals who vary with regard to a quantitative wait, ‘The line in the frequency distribution depicts a normal distribution, a distribution for a large sample in which the trait of interest varies in a symmetrical way around an average value. The distribution of measurements of many biological characteristics is approximated by a symmetrical bell-shaped curve like that in Figure 28.1b, Normal distributions are common when the pheno- type is determined by the cumulative effect of many small inde- pendent Factors. 28.1 COMPREHENSION QUESTIONS: 41. Which ofthe following isan example of a quantitative trait? a. Height b. Rate of glucose metabolism © Abiityo learn 8 maze 4, Allofthe above are quantitative tals. 2. Saying that a quantitative tat follows a continuum means that a. the numerical value forthe tat increases with the age of the individual . environmental effects are adltve the phenotypes for the trait are continuous and do not fall Into discrete categories, 4, the trait continuously changes curing the Ife ofan individual 28.2 STATISTICAL METHODS FOR EVALUATING QUANTITATIVE TRAITS Learning Outcome: 4. Calculate the mean, variance, standard deviation, and correlation coefficient or quantitative tras, ane explain the meanings of these statistics {In the early eighteenth century in England, Francis Galton and his, student Kas! Pearson showed that many traits in humans and do mesticated animals are quantitative in nature. To understand the underlying genetic basis ofthese trail, they founded what hecame known as the biometric field of genetics, which involves the sta- tistical study of biological traits. During this period, Galton and Pearson developed various statistical tools for studying the varia- tion of quantitative traits within groups of individuals. Many of these tools are stil in use today. In this section, we will examine ‘how statistical tools are used to analyze the variation of quantita- tive traits within groups. Statistical Methods Are Used to Evaluate a ‘Frequency Distribution Quantitatively Statistical tools can be used to analyze @ normal distibution in a ‘number of ways. One measure you are probably familiar with is a parameter called the mean, which isthe sum of all the values in the group divided by the number of individuals inthe group. The ‘mean is computed using the following formula =x W where: F isthemean EX i the sum ofall the values in the group N isthe number of individuals in the group ‘Amore generalized form of this equation can be used: TX, X isthe mean B/,X; isthe sum of all the values in the group; each value inthe group is multiplied by its frequency (Zin the group N isthe number of individuals inthe group For example let's suppose a group of corn ears have the following lengths (rounded tothe nearest centimeter: 15, 14, 13, 14, 15, 16 16,17, 15, and 15. Then 4(15) +204) + 13 + 2006) +17 To F=t5em In genetics, we are often interested in the amount of phenotypic ‘tiation in a group. As we will see later in this chapter and in (Chapter 29, variation lies atthe heart of breeding experiments and evolution. Without variation, selective breeding is impossible, and ‘natural selection cannot favor one phenotype over another. A com- ‘mon way to evaluate variation within a population is with a stais- tic called the variance, which is a measure of the variation around ‘the mean, It helps us appreciate how far a set of nurnbers is spread ‘out. Variance is determined by adding up the squared deviations710 (CHAPTER 28 © COMPLEX AND QUANTITATIVE TRAITS from the mean. In relatively small sample sizes, this sum is di- vided by N~ 1 to give an estimate of the variance. where Vy _— isthe variance X,— X isthe difference between each value and the mean Nis the number of observations For example, if we use the values given previously forthe lengths, ‘of eats of cor, the variance in length is calculated as follows’ Bi B= 4s - 15)? + 204-19) + (13 — 15) +206 — 15) +7 = 15)" BX NPSO+2+4 4244 Bi Qi- Hi= 120 BAX = Vq= 1.33 0m" ‘Although variance is a measure of the variation around the ‘mean, this statistic may be difficult to understand intuitively because the variance is computed from squared deviations. For ‘example, weight can be measured in grams; the corresponding variance is measured in grams squared, Even so, variances are centrally important to the analysis of quantitative traits because they ate additive under certain conditions. This means that the variances for different factors that contribute to a quantitative trait, such as genetic and environmental factors, can be added together to predict the total variance for that trait. Later, we will ‘examine how this property is useful in predicting the outcome of genetic crosses. To gain a moce intuitive grasp of variation, we can take the ‘square root of the variance. This statistic is called the standard deviation (SD). Again, using the example of the lengths of corn cars, the standard deviation is sp= WW, = Vis SD =1.15 em If the values in a population follow a normal distribution, it is ‘easier to appreciate the amount of variation by considering the standard deviation, Figure 28.2 illustrates the relationship be: ‘ween the standard deviation and the percentages of individuals ‘that deviate from the mean. Approximately 68% of all individuals, have values within one standard deviation from the mean, either in the positive or negative direction. About 95% are within two stan- dard deviations, and 99.7% are within three standard deviations. ‘When & quantitative characteristic follows a normal distribution, less than 0.3% of the individuals have values that are more or less ‘than three standard deviations from the meaa of the population. In ‘our corn example, three standard deviations equals 3.45 cm, Frequency 32 a 380 FIGURE 28.2 The relationship between the standard 4eviation and the proportions of individuals in a normal distribu- tion, For example, approximately 68% of the individuals in «population are between the mean and one standard deviation (I SD) above ot below ‘he mean, ‘CONCEPT CHECK: What percentage of individuals all mor than 2 SDs above ne mean? ‘Therefore, we expect approximately 0.39% of the ears of com have lengths less than 11.55 em or greater than 18.45 cm, assuming that, com ear length follows a normal distribution. GENETIC TIPS THe question: in a population of 100 ‘mal fruit flies, the mean abomen length s 2.0 rum and the standard (9) Wo x Figure 28 4a shows a situation in which seed weight is controlled bby one gene with light (wv) and heavy (W) alleles. heterozygous plant (Wu) is allowed to selffertilize, When seed weight is only Slightly influenced by environmental variation, as seen on the left, the ww, Ww, and WW genotypes result in well-defined phenotypic classes, When environmental variation has a greater effect on seed ‘weight as shown on the right, more phenotypic variation is found in seed weight for each genotype. The variation inthe frequency dlistibution on the right is much higher. Even so, most genotypes cean sill be eategorized into the three main classes By comparison, Figure 28,4b illustrates a situation in which seed weight is governed by three genes instead of one, each existing High environmencal eft FIGURE 28.4 How genotypes ‘and phenotypes may overlap for quanti- ‘ative traits. (a Situations in which seed weight is contlled by one gene, existing in ight (w) and heavy (19 alles. (b) Situ- ations in which seed weight is governed by thee genes instead of one, each existing in Tight and heavy alleles. (Note: The 1:2:1 and 16:15:20: 56:1 ratios were derived by using a Punnet square and assuming 8 ‘ross between individuals that are both rtcrorsgos for one thee citron Seed weight (HIE we MTW we 3 5 se te NCH Ws LI oe ee og oo i ° f Seed weight i i i ro wr wll’ wr WIV UW a WT wT Wr vr HUW ‘ . ‘s 20) s gees, respectively.) ‘Genesis The ait of genetics ocr reste genotype and phanctype depends on how many genes ae vowed and how much the envionment causes the phenotype ‘avy Inpar ah sang gene nfuences seed ‘weight Inthe graph on thee sige, envcon- mental variation does nt cause much var tion n seed weight. No overiap in seed weight is observed forthe wu Wx a HO geno ‘types. Inthe graph onthe ght se, environ mental variation nasa grate eect on seed weight Ins case, a ew individuals with the me genotype produce seeds having the same weight as seeds of few ndiadvls wth the Wa gonorye: and afew inddals wth the Ha genotype produce seeds having the same weight as seeds rom ncviduas| wit re MY genotype. As shown in par. becomes even more aio estngush genotype based on phenotype when three fenes ore involved, The overlaps se minor ‘en enanmental vation doesnot couse much see¢ weight variation. However. when ‘vronmental variation has» greater effect ‘an phenaype, the overaps between geno ‘ypes and phenotypes are very pronounces andthe at appears to flow acontraum of ‘ A ‘ait, ‘CONCEPT CHECK: Esplin how gene num ber and environmental vison afc he coverlas between phenotypes and eiferent genotypes.24. IDENTIFICATION OF GENES THAT CONTROL QUANTITATIVE TRAITS 715 in light and heavy alleles. A cross between two heterozygotes is, ‘expected to produce seven genotypes in a 1:6:15:20:15:6:1 ratio. When the variation in enviconmental factors is low and/or plays a ‘minor roe inthe outcome ofthis trail as shown in the upper graph in Figure 28.4, nearly all individuals fall within a phenotypic class that corresponds to their genotype. When the environment has a teat effect on phenotype, as shown inthe lower graph, the situa- tion becomes more ambiguous. For example, individuals with one w allele and five W alleles have a phenotype that overlaps with that of individuals having six W alleles or two w alleles und four W alleles, ‘Therefore, it becomes difficult to categorize each genotype into a unique phenotypic class. Instead, the trait displays « continuum, ranging from light to heavy seed weight. 28.3 COMPREHENSION QUESTION 41. For many quantitative traits, genotypes and phenotypes tend to overlap because a the tat changes overtime, b. the tats polygenic. «environmental variation affects the tra 4, both band care true 28.4 IDENTIFICATION OF GENES THAT CONTROL QUANTITATIVE TRAITS Learning Outcomes: 4. Dotine quantitative trait focus. 2. Explain how quantitative trait loci are mapped along chromosomes using molecule markers. A goal of researchers working in the field of quantitative genetics is to idemtify the genes that ate associated with complex and quan. titative tails, This can be a challenging endeavor because such traits are usually polygenie. Inthe past few decades, the develop- ‘ment of many molecular approaches has made it easier for re- searchers to identify these genes. In this section, we will consider an approach to map the locations of genes that control quantitative traits based on the locations of molecular maskers. ‘The location on a chromosome that harbors one or more genes that affect the outcome of a quantitative tat is called a ‘quantitative trait locus (QTL). QTLs are chromosomal regions identified by genetic mapping, Because such mapping usually lo ‘cates a QTL toa relatively large chromosomal region, « QTL may ‘contain a single gene or it may contain two or more closely linked genes that affect a quantitative ait ‘To map genes, researchers usually determine their locations by identifying ther linkage to molecular markers, This approach is described in Chapter 25 (sce Figures 25.5, 25.6,25.7). The basis of QTL mapping is the association between genetically deter- mined phenotypes for quaatiative tats and molecular markers such as restriction fragment length polymorphisms (RFLPs), mic- rosatellites, and single nucleotide polymorphisms (SNPs). In this approach, a researcher identifies QTLs that are close to pasticular molecular markers whose locations along a chromosome are al- ready know. ‘The general strategy for QTL mapping is shown in Fig~ lure 28.5. Ths figure depicts two different strains of a diploid plant species with four chromosomes per set. The strains are highly in- bred, which means they are homozygous for most molecular mark- cers and genes. They differ in two important ways, First, the to strains differ with regard to many molecular markers; the sites of many markers along each chromosome are already known. These markers are designated 1A and 1B, 2A and 2B, and so forth. The ‘markers 1A and 1B mark the same chromosomal location in this species, namely, the upper tip of chromosome 1. However, the two markers are distinguishable in the two strains at the molecular level. For example, 1A might be a microsatellite that is 148 bp, ‘whereas 1B might be 212 bp. Second, the two strains differ in a quantitative trait of interest. In this example, the strain on the left produces large fruit, whereas the strain on the right produces small Fruit. The unknown genes affecting this trait are designated with a black or blue QTL. label. A black QT. indicates a ste that harbors ‘one or more alleles that promote large fruit. A blue QTTL is atthe ‘same site but carries alleles that promote small fruit. Prior to con- ‘ducting their crosses, researchers would not know the chromo- somal locations of the QTLs shown in this figure. The purpose of the experiment is to determine their location With these ideas in mind, the protocol shown in Figure 28.5 begins by crossing the two inbred strains to each other and then backcrossing the F) offspring to both parental strains, This pro- ‘duces a second generation with a great degree of variation. The ‘offspring from these backerosses are then characterized in two ways. Fits, they are examined for ther fruit size, and second, a cell sample from each individual is analyzed to determine which molecular markers are found in their chromosomes. The goal isto find an association between particular molecular markers and {rut size. For example, 2A is stoongly associated with large size, ‘whereas 2B is strongly associated with small size. By comparison, ‘9A and 9B are not associated with large of seul size, because a QTL affecting this trait is not found on this chromosome. Also, ‘markers L4A and 14B, which ae fairly far away from a QTL, are not strongly associated with any particular QU. Markers that are ‘on the same chromosome but far away from a QTL are often sepa rated from a QTL during meiosis in the F heterozygote due to crossing over. Only closely linked markers are strongly associated witha particular QTL. Overall, QTL mapping involves the analysis ofa large num- ber of markers and offspring. The data are analyzed by computer programs that can statistically associate the phenotype (€2, fruit size) with particular markers. Mazkers found throughout the ge- ‘nome ofa species provide a way to identify the locations of several different genes that possess allelic differences that may affect the ‘outcome of a quantitative trait As an carly example of QTL mapping, in 1988, Andrew Paterson and his colleagues examined quantitative trait inheri- tance in tomato plants. They studied a domestic strain of tomato716 (CHAPTER 28 © COMPLEX AND QUANTITATIVE TRAITS Not: The locations of GTLs ae not known athe sar af this experiment Strain prosucing large fruit Strain producing smi a@ Qu 2@ Qe on onsen ge on oneg Qce we So Th omy yon 2 He sol) [oo om on TAD ne cn Qua ce a wad Grn weld Gos sale Wes0 sae Son wad oe sod se ood eo m2 sald (en veld ee sale gon eS Sew cold (oc a we ou we be od de wed due gg | om ‘ort am SAS gs WAL ion tea Lisa 20 Joon” so Gon wag [fron sof Yiso “coe on. on on en, OTL 2 GT promoting ergo size ut {OTL promoting smallsze tt Cross the 2 stains to produce Fs otsprng on onsaQ goo WS 2 nT Ue AQ gaot wa Qre sale (dae sale das male (ave all Wes wale Gre ane dan wal ue ald (oo or on sae se oad (ioe sais (ise 20a (aoe on- ‘on 7 Backers he Fy ofepring jo bath parental Sins, Crtsping trom the boccerosses contan ferent combnatone ‘of parental eomosomes. For many aisaring tam {ho backeronses determine ‘russe nd molecular mater composiion 2A, 5A A. ana ton She Ths suggest ran associated with arge tut dala fe cose tothase markers, FIGURE 28.5 The general strategy for QTL mapping via molecular markers. ‘Two differen inbred strains have four chromosomes per se. “The stain onthe loft produces large ft, and the strain om the right prodices sll feu. The goal of this mapping strates isto locate the unknown genes affecting this ait, which are designated with a QTL label. A black QTL indicates a site carrying one or moze alleles that promote large fruit, and a blue QTL caries alleles that promote small rut. The two strains differ with epard to many molecular markers designated 1A and 1B,2A and 2B, and 0 forth, The two strains are crossed, and ten the F offspring ae backcrosed o both parental strains. Many offspring from the backcrot.s are then ex amined for tei fruit size and Io determine which molecular mazkers are found in their chromosomes, The dita are analyzed by computer programs that, an statistically associate the phenotype (eg frit size) with particular markers, Markers found throughout the genome ofthis species provide a way to locate several different genes that may affect the outcome ofa single quantitative trait In this cas, the analysis predits four QTLs promoting heavier ‘uit weight that are linked to regions ofthe chromosomes with the following markers: 2A, 5A, IIA, and 194, ‘CONCEPT CHECK: What te the two ways hat stains AandB der?and @ South American green-fruited variety. These two strains differed in their RFLPs, and they also exhibited dramatic differ ‘ences in three agriculturally important characteristics: fruit mass, soluble solids content, and fruit pH, The researchers crossed the two strains and then backerossed the offspring tothe domestic to mato, The researchers then examined 237 plants with regard to 70, known RELP markers. In ition, 5-20 tomatoes from each plant were analyzed with regard fo fruit mass, soluble solids content, and fruit pH. Using this approach, the researchers were able 10 map genes contributing much of the variation in these traits to particular sites along the tomato chromosomes. They identified six loci causing variation in fruit mass, four affecting soluble sol ids content, and five with effects on fruit pH. More recently, the DNA sequence of the entire genome of many species has been determined. In such cases, the mapping of QTLs to a defined chromosomal region may allow researchers 10 analyze the DNA sequence in that region and to identify one oF more genes that influence & quantitative trait of interest 28.4 COMPREHENSION QUESTIONS 1 Agia here one or more genes affecting a ‘quantitative vat are a. ste in a cel located b. site in a chromosome, located «. site in a cell, expressed 4. stein a chromosome, expressed 2. Tomap GILs, strains ae crossed that fer with regard to a. a quantitative tr. b, molecular marke. © a quantitative trait and molecular markers 4, a quantitative trait anda discontinuous trait eens Learning Outcomes: 41. Explain the relationship among phenotypic variance, genetic variance, and environmental variance using an equation. 2. Describe how interactions and associations between geno- ‘ype and environmental factors may affect phenotypic variance 43. Detine hertabity, and distinguish between broad-sense heritability and narrow-sense heritability 4. Calcuate narrowsense heritability using correlation coefcients. ‘As we have just seen, recently developed approaches in molecular _mapping have enabled researchers to identify the genes that contrib- ute (0 a quantitative trait The other key factor that affects the phe- notypic outcome of quantitative waits is the environment. All tits ‘of organisms are influenced by both genetics and the environment, and this kind of interaction is particularly pertinent in the study of 285 HERITABILITY 717 quantitative traits. Researchers want to understand how variation, both genetic and environmental, affects the phenotypic results. In this section, we will examine how geneticists analyze the genetic and environmental components that affect quantlative traits, Both Genetic Variance and Environmental Variance May Contribute to Phenotypic Variance Earlier, we examined the amount of phenotypic variation within a ‘group by calculating the variance, Geneticists partion quantita- tive trait variation into components that are attributable tothe fol- lowing different causes Genetic variance (Ve) Environmental variance (Vi) Variance due to interactions between genetic and environ- mental factors (Voxe) ‘Variance due to associations between genetic and environ- ental factors (Vow) Let's begin by considering a simple situation in which Vo and Vi are the only factors that determine phenotypic variance, and the ‘genetic and environmental factors are independent ofeach other. If so, then the total variance for a tat in a group of individuals is Vp= Vet Ve where Vo is the total phenotypic variance Vg. is the relative amount of variance due to genetic variation Vo is the relative amount of variance due to ‘environmental variation Why is this equation useful? The partitioning of variance into genetic and environmental components allows us to estimate their relative importance in influencing the phenotypic variance ‘within a group. If Vq is very high and Vp is very low, genetics, plays the greater role in promoting phenotypic variation within a ‘group. Alternatively, if Ve is low and Vj. is high, environmental factors underlie much of the phenotypic variation. As described later inthis chapter, a livestock breeder might want to apply selec- tive breeding if Vg for an important quantitative tat is high. In this way, the characteristics of the herd may be improved, Alterna tively, if Ve is negligible, it would make more sense to investigate ‘and manipulate the environmental causes of phenotypic variation, With experimental and domesticated species, one possible way to determine Vq and Vis by comparing the variation in traits between genetically identical and genetically disparate groups, For ‘example, researchers have used the practice of inbreeding to de~ ‘velop genetically homogeneous strains of mice. Inbeceding in mice involves many generations of brother-sister matings, which ‘eventually produces strains that are monomorphic—carry the same allele—for all oftheir genes. Within such an inbred strain of mice, Vo equals zero, Therefore all phenotypic variance is due to Vp. When studying quantitative traits such as weight, an experi- ‘meater might want to know the genetic and environmental vaiance718 (CHAPTER 28 © COMPLEX AND QUANTITATIVE TRAITS fora different, genetically heterogeneous group of mice. To do so, the genetically homogeneous and heterogeneous mice could be raised under the same environmental conditions and theie weights ‘measured (in grams). The phenotypic variance for weight could then be calculated as described earlier. Let's suppose we obtained the following results: Vp= 15 g° for the group of genetically homogeneous mice Vp = 22 g° for the group of genetically heterogeneous mice In the case of the homogeneous mice, Vp = Vp, because Vo ‘equals 0. Therefore, Ve equals 15 g”. To estimate Vg for the het- cerogencous group of mice, we could assume that Vy is the same for them as it is for the homogeneous mice, because the two ‘groups were raised in identical environments. This assumption allows us to calculate the genetic variance in weight forthe het- «erogeneous mic. Vo=Vot Ve gra Ve+ IS Vo=T8" ‘This result tells us that some of the phenotypic variance in the genetically heterogeneous group is due to the environment (namely, 15 g°), and some (7 ¢°) is due to genetic variation in alleles that affect weight. Phenotypic Variance May Also Be Influenced by Interactions and Associations Between Genotype and the Environment “Thus far, we have considered the simple situation in which genetic ‘variance and environmental variance axe independent of each other and affect the phenotypic variance in an additive way, As another ‘example, let's suppose that three genotypes, HH, Hh and hh, affect Ihcight, producing tall, intermediate, and short plants, respectively. Greater sunlight makes the plants grow taller regardless of theit _Benotypes. In this case, our assumption that Vp= Vq + Vz would be reasonably vali, However, le’s consider a different environmental factor such as minerals in the soil. As a hypothetical example, let's ‘suppose the HY allele encodes a fully functional protein involved ‘with mineral uptake from the soil, whereas the h allele carries a ‘mutation that causes a decrease inthe affinity ofthis protein for certain minerals. In this ease, the Hh and hh plants are shorter because they do not lake up enough minerals to support maxi- ‘mal growth, whereas the HHT plants are not limited by mineral uptake. According to this hypothetical seenario, adding miner- als to the soil enables the defective protein to transport more ‘minerals into the roots, thereby enhancing the growth rate of hh plants by a large amount and the Hh plants by a smaller amount, because the minerals are more easly taken up by the plants (Figure 28.6). The height of HH plants is not affected by min- cral supplementation, When the environmental effects on phe notype differ according to genotype, this phenomenon is called ‘2 genotype-environment interaction. Variation due to interac- tions between genetic and environmental factors is termed Vow, as noted earlier Genotype Standard_—_—Soll supplemented ‘sl ‘with minerals No ‘tect oe Sill fenuronmestal ‘tect Hn Large ‘tect FIGURE 28.6 A schematic itustratio ofa genotype- ‘environment interaetion. When grown in standard sol, the thre {genotypes HH, Hh, and hk show tll, atrmediae, and shoet height, respectively. When the sis supplemented with minerals, large effect is seen onthe nk genotype and a smaller fet on the Hh genotype. The FIT genotype is unaffected by the environmental change, Interactions between genetic and environmental factors are common. As an example, Table 28.3 shows results from a study conducted in 2000 by Cristina Vieira, Trudy Mackay, and col- leagues in which they investigated the genotype-environment inter- action for QTLs affecting life span in Drosophila melanogaster. The Temperature __Male__—_‘Female__—_Male_—_—_—Female Standard 36 395 375 288 High 363 a8 22 288 ow ms 483 458. m9 “Longevity was esse inte moan number fay falas Aan Bere insted ss. songs cle Oregon a 2 respecvey The stant Fig, eu erpeatiecontans were 25, 29, and 14 respectedata in the table compare the life spans in days of male and female ‘ies from two different stains of D. melanogaster raised at differ ent temperatures. Because males and females differ in their sex chromosomes and gene expression patter, they can be viewed as having different genotypes. The effects of environmental changes depended greatly on the strain and the sex of the flies. Under stan- dard culture conditions, te females of strain A had the longest life span, whereas females of stain B had the shortest. In strain , high, temperature increased the longevity of males and decreased the Tongevity of females. In contrast, under hotter conditions, the lon- gevity of males of strain B was dramatically reduced, whereas f males of this same strain were not significantly affected. Lower growth temperature also had different effets in these 1wo stains Although low temperature increased the longevity of both strains, the effects were most dramatic in the males of strain A and the females of strain B. Taken together, these result illustrate the potential complexity of the effects of genotype-environmental interaction on a quantitative trait such as life span Another issue confronting geneticists is that genotypes may not be randomly distributed in all possible environments. When. certain genotypes are preferentially found in particular environ- ments, this phenomenon is called a genotype-environment association (Vi..z). When such an association occurs, the effects ‘of genotype and environment are not independent of each other, and the association needs to be considered when determining the effects of genetic and environmental variance on the total pheno. typic variance. Genotype-environment associations are very common in human genetics, since members of families tend 10 have more similar environments than do members of the popu tion as a whole. One way to evaluate this type of effect is to com- pare individuals who have different genetic relationships, such as identical versus fraternal twins, We will examine this approach, later in this section. Another strategy that geneticists might fol- low is to analyze siblings that have been adopted by different parenis at birth, Their environmental conditions tend to be more disparate, and this may help to minimize the effects of genotype- environment association, Heritability Is the Relative Amount of Phenotypic Variance That Is Due to Genetic Variance Another way to view variance is to focus our attention on the ge- netic contribution to phenotypic variance. The term heritability refers to the amount of phenotypic variance within a specific group of individuals raised ina particular environment that is due to genetic variance, Both genes and the environment are essential to produce the traits of an organism, Even s0, variation of trait in a population may be due entirely to environmental variation, entirely to genetic variation, oF more commonly to a combination, of the two. If all of the phenotypic variance in a group is due to ge- netie variance, the heritability will have a value of I. Ifall of the phenotypic variation is due to environmental effects, the herita- bility will equal 0. For most groups of oxganisms, the heritabil- ity for a given trait lies between these two extremes. For example, both genes and diet affect the size an individual will, 285 HERITABILITY 719 attain, In @ given population, some individuals inherit alleles ‘that tend to make them larger, and a proper diet also promotes larger size. Other individuals inherit alleles that make them small, and an inadequate diet may contribute to small size ‘Taken together, both geneties and the environment affect the amount of phenotypie variation fora trait such as size, If we assume that environment and genetics are independent and the only two factors affecting phenotype, then ig? = Vole Ing? is the heritability in the broad sense Vos is the variance due to genetics Vp isthe total phenotypic variance, which equals Va +Ve The heritability defined here, Hy’, called broad-sense heritability takes into account different types of genetic variation ‘that may affect the phenotype. As we have seen throughout this text, genes ean affect phenotypes in various ways. As described ‘earlier, the Nilsson-Ehle experiment showed thatthe alleles deter- rining hull color in wheat affect the phenotype in an additive way. A heterozygote shows a phenotype that is intermediate between the respective homozygotes. Altematively alleles affecting other waits may show a dominanvecessive relationship. In this case, the alleles are not strictly additive, because the heterozygote has phenotype close to, of pethaps the same as, the homozygote con- taining two copies of the dominant allele, For example, Mendel discovered that both PP and Fp pea plants have purple flowers. In addition, another complicating factor is epistais (described in ‘Chapter 4, in whieh the alleles for one gene can mask the pheno- typie expression of the alleles of another gene. To account for these differences, geneticists usually subdivide Ve into these three different genetic categories: Ve= Vat Vo + Vi wher Vq is the variance due to the additive effets of alleles Vp. is the variance due to the effects of alleles that follow a dominantrecessive pattern of inheritance V; isthe variance due to the effects of alleles that inter= actin an epistatic manner In analyzing quantitative traits, geneticists may focus on Vq and ignore the contributions of Vp and V. They do this for scientific as well as practical reasons. For some quantitative teats, the additive effects of alleles may play a primary role in the phenotypic outcome. In addition, when the alleles behave additively, we can predict the outcomes of crosses based on the ‘quantitative characteristics of the parents. The heritability of a trait due tothe additive effects of alleles is called narrow-sense heritability: fa = VV For many quantitative tits, the value of Vy may be relatively large compared with Vp and Vj In such eases, th determination of the narrow-sense heritability provides an estimate of the broad: sense heritability0 (CHAPTER 28 © COMPLEX AND QUANTITATIVE TRAITS How can the narrow-sense heritability be determined? In this chapter, we will consider two common ways. As discussed later, one way fo calculate the narrow-sense heritability involves selective breeding practices, which are done with agricultural species. A second common strategy for determining narrow- sense heritability involves measurement of a quantitative trait among groups of genetically related individuals. For example, agriculturally important traits, such as egg weight in poultry, ‘ean be analyzed in this way. To caleulate the heritability, a r¢- searcher determines the observed egg weights between indivi uals whose genetic relationships are known, such as a mother and her female offspring. These data can then be used to com- pute a correlation coefficient between the parent and offspring, using the methods described earlier. The narrow-sense herita Dilty is then calculated as I? rang where ray i8 the observed phenotypic correlation coefficient between related individuals egy is the expected correlation coefficient based on the known genetic relationship In our example, rai the observed phenotypic correlation co- cfficient between parent and offspring. In actual research studies, the ‘observed phenotypic correlation coefficient for egg weights between ‘mothers and daughters bas been found to be about 0.3 (although this varies among stains) The expected correlation coeficint, ray. based on the known genetic relationship. A parent and child share 0% of thee genetic matral, 0 ra equal 0.50. Thus, bo = taba 0.30.50 =0.60 (Note: For siblings, ray = 0.50; for identical wins, ray = 1.05 and for an aunt-niece telationship, ray = 0.25.) According to this calculation, about 60% of the phenotypic variance in egg ‘weight is duc to additive genetic variance; the other 40% is due to the environment, This calculation assumes that Vy and V; are negligible. When calculating heritabilities from correlation coeffi- cients, keep in mind that such a computation also assumes that {genetics and the environment are independent variables, How- cer, this is not always the case. The environments of parents, and offspring are often more similar to cach other than are the environments of unrelated individuals. As mentioned earlier, there are several ways (© minimize this confounding Factor, First, in human studies, researchers may analyze the heritabili- ties from correlation coefficients between adopted children and their biological parents. Alternatively, they ean examine a vari- ty of relationships (aunt-niece, identical twins versus fraternal ‘wins, and so on) and see if the heritability values are roughly the same in all cases. This approach was applied in the study that is described neat. Eto 28A ‘The Heritability of Dermal Ridge Count in Human Fingerprints Is Very High Fingerprints are inherited asa quantitative trait. Ithas long been {known that identical wins have fingerprints that are very i lar, whereas fraternal (wins show considerably less similarity Galton was the first rescarcher to study fingerprint patterns, but this tat became more amenable to genetic studies inthe 1920s, ‘when Kristine Bonnevie, a Norwegian geneticist, developed & ‘method for counting the number of ridges within a human fingerprint. ‘As shown in Figure 28.7, human fingerprints can be categorized as having an arch, loop, or whorl, or a combination ‘of these patterns. The primary difference among these patterns is the number of triple junctions, each known as atriradius (Fig- ‘ure 28.7b and c). Ata trradius, a ridge emanates in three differ- cet directions. An arch has zero triradi, a loop has one, and a whorl has two. In Bonnevie's method of counting, a line is drawn from a triradius to the center of the fingerprint, The ridges that touch this line are then counted, (Note: The triradius ridge itself is not counted, and the last ridge is not counted if it forms the center of the fingerprint.) With this method, one can obtain a ridge count for all 10 fingers. Bonnevie conducted a study on a small population and found that ridge count correla tion coefficients were relatively high in genetically related individuals Sarah Holt, who was also interested in the inheritance ofthis quantitative trait, carried out a more extensive analysis of ridge counts by examining the fingerprint patterns of a large group of people and their close relatives. In the experiment of Figure 28.8, the ridge counts for pairs of related individuals were determined by the method deseribed in Figure 28.7. The correlation coeffi- cients for ridge counts were then calculated among the pairs of related or unrelated individuals. To estimate the narrow-sense heritability, the observed correlation coefficients were then di- ‘vided by the expected correlation coefficients based on the known, ‘genetic relationships. THE HYPOTHESIS Dermal ridge count has a genetic component. The goal of this, experiment was to determine the contribution of genetics to the ‘variation in dermal ridge counts,TESTING THE HYPOTHESIS Trad (©) Loop tractus) ‘Starting material: A group of human subjects from Great Britain, 2, Roll the person's finger onto a secording surface to obtain a pint. 3. With a low-power binocular microscope, count the number of ridges, using the ‘method described in Figure 28.7. 4, Caleulate the correlation coefficients betwicen different pairs of individuals as described earlier in this chapter Experimental level See The Data. Trrsoius (0 Whot 2 troy FIGURE 28.7 Human fingerprints and the ridge count method of ‘The ridge counts 220. () This print has one ira ‘A straight line is drawn from the triads tothe center ofthe print. The numberof ridges dissecting this straight line i 13.) This print has two tri- radi, Straight lines are drawa, ftom both tradi to the centr. ‘There are 6 ridges touching the lef line and 7 touching the right lin, giving atta ridge count of 23, FIGURE 28.8 Heritability of human fingerprint patterns. Conceptual level ‘This isa method to measure a quantitative trait. ‘The correlation coefficient provides ‘way to determine the heritability forthe quantitative tail, nam2 (CHAPTER 28 © COMPLEX AND QUANTITATIVE TRAITS THE DATA Type of Number of Correlation Heritability Relationship _Pairs Examined _Coefficent (Pay) _(Fow/Tas) _ Parenthild S10 0.48 + 008 0.96 ‘Parent-parent 200 0.05 + 0.07 - Sibling: sibling oa 0.50 + 0.04 1.00 entical wins 30 095 + 0.01 0.95 Fraternal twins 92 049 + 0.08 0.98 Average heritability = 0.97 ‘The value flowing she standard enor ofthe mean, “Wecanot alta herimblty vain his ate Deas the ale fr rg 0 DOH Neverthe vale fo sis ey Tow, sagging tat here isa eligible earclaon| Sout: Adapted tom SB. Hot (1961), Quanative gents of ingepsat pars, Br Med Bul 17, 267-250, INTERPRETING THE DATA ‘As seen inthe data, the results indicate that geneties plays the major role in explaining the variation inthis trait. Genetically unrelated in dividuals (namely, the parent-parent relationships) have a negligible correlation fr ths trait. By comparison, individuals who are geneti- cally related have a substantially higher correlation coefficient ‘When the observed correlation coefficient is divided by the expected correlation coefficient based on the known genetic relationships, the average heritability value is 0.97, which is very close to LO. ‘What do these high heritability values mean? They indicate ‘that nearly all of the phenotypic variance in fingerprint pattern is, due to genetic variance. Significantly, fraternal and identical twins have substantially different observed correlation coefficients, even though we expect they have been raised in very similar environ- ‘ments, These results support the idea that genetics is playing the ‘major role in promoting variation and the results are not biased ‘heavily by environmental similarities that may be associated with ‘genetically related individuals. From an experimental viewpoint, the results show us how the determination of correlation coe cients between related individuals can provide insight into the relative contributions of genetics and environment to the variation of a quantitative trait 28.5 COMPREHENSION QUESTIONS 41. Ina population of squirels in North Carolina, the heritability for body weightis high. This means that a, body weight's primarily controlled by genes b. the environment has Ile influence on body welght the vatiance in body weights mostly due to genetic variation 4, both a and b are correct 2. If two or more diferent genotypes do not respond to environ- ‘mental valation inthe same way, this eutcome is due to a. a genotype-environment association, b. a genotype-envronment interaction the additive effects of alles 4, bothaand. 3. One way to estimate narrow-sense hertabilt fora given traits tocompare__for. 2 variances, related pair of inciduals b. cottlation coetficents, related pats of individuals variances, unrelated pais of individuals, 4, cortelation coefficients, unrelated pais of individuals a BREEDING Learning Outcomes: 41. Describe the effects of selective breeding 2. Caleulate hettablty from the results of selective breeding ‘experiments 3. Explain how dominance and overdominance may contribute to the beneficial characteristics of hybrids ‘The term selective breeding refers to programs and procedures, designed to modify phenotypes in species of economically impor- tant plants and animals. This phenomenon, aso called artificial selection, is related to natural selection, discussed in Chapter 27 In forming his theory of natural selection, Charles Darwin was, influenced by his observations of selective breeding by pigeon fanciers and other breeders. The primary difference berween arti ficial and natural selection is how the parents are chosen. Natural selection is due to natural variation in reproduetive success, In artificial selection, the breeder chooses individuals that possess traits that are desirable from a human perspective, In this section, ‘we will examine the effects of selective breeding and consider is, relationship to heritabilitySelective Breeding of Species Can Alter Quantitative Traits Dramatically For centuries, humans have been practicing selective breeding to oblain domestic species with interesting or agriculturally useful characteristics. Avery sitking example is the dog, whieh is a common house pet. All domestic dogs are derived from the gray ‘wolf (Canis lupus). The various breeds of dogs have been obtained by selective breeding strategies that typically focus on morpho. logical traits (e.g. size, fur color, ete.) and behavioral waits (e.g. ability to hun, friendly to humans, etc). As shown in Figure 28.9, itis very striking how selective breeding ean modify the quantita tive traits in a species, When comparing a greyhound with a bll- dog, the magnitude of the differences is amazing. They hardly look like members ofthe same species ‘A QTL study in 2007 by Nathan Sutter and colleagues indi cated thatthe size of dogs is determined, in part by alleles of the Iefl gene, which encodes a growth hormone called insulin-like growth factor I. A particular allele of this gene was found to be common to all small breeds of dogs and nearly absent from very ules ‘Cocker spaniel FIGURE 28.9 Some common breeds of dogs that have boen obtained by selective breeding. ‘Genes By electing parents caring the alles hat havea desred effect on certain quanttave tals, dog breeders have produced breeds with Astincve sets oft, Fo example, the bulldog has alles tt produce Shor legs an aft face. By comparison, the corresponding genes in German shepherd have ales tat prosice longer legs and a ore pony snout Al ohe dogs shown in this Egure cary the same kinds of gees (e.g any genes tht affect thelr szs, stapes, and fur olor However. the alles Formany af these genes ae diferent among these eos, theredy posing breeds wih strangly deren phenotypes phy: Hey Alana ema ep) Ro io ats ‘CONCEPT CHECK: What ae the snares and diferences between nati ‘alselection and selective breecing? 6 na large breeds, suggesting that this allele is one of several genes that influences body size in small breeds of dogs. Likewise, most ofthe food we eat is obtained from species that have been modified profoundly by selective breeding strate- ses, These food products include grains, fruits, vegetables, meat, milk, and juices. Figure 28.40 ilkustrates how certain characteris- tics in the wild mustard plant (Brassica oleracea) have been mod: ified by selective breeding to create several varieties of important domesticated erops. The wild plant is native to Europe and Asia, and plant breeders began to modify its traits approximately 4000 ‘years ago, As seen here, certain quantitative traits in the domestic strains, such as stems and lateral buds, differ considerably from those of the original wild species. The phenomenon that underlies selective breeding is genetic variation. Within a group of individuals, allelic variation may af- feet the outcome of quantitative traits, The fundamental strategy ‘of the selective breeder is to choose parents thal will pass on to thir offspring alleles that produce desirable phenotypic charac- teristics, For example, if a breeder wants large cattle, the largest _members of the herd are chosen as parents for the next generation, ‘These large cattle will wansmit an aray of alleles to their offspring that confer large size, The breeder often chooses genetically re- lated individuals (e., brothers and sisters) as the parental stock, ‘As mentioned previously, the practice of mating between geneti- cally related individuals is known as inbreeding, Some of the con sequences of inbreeding ate also described in Chapter 27 ‘What is the outcome when selective breeding is conducted for a quantitative trait? Figure 28.11 shows the results of a pro- ‘gram begun at the Iinois Agricultural Experiment Station in 1896, before the rediscovery of Mendel’s laws. This experiment ‘began with 163 ears of corn with an oil content ranging from 4% 10 6%. In cach of 80 suceceding generations, corn plants were di vided into two separate groups. In one group, several members with the highest oil content were chosen as parents of the next _generalion, Inthe other group, several members with the lowest oi content were chosen. After 80 generations, the oil content in the first group rose to over 18%; in the other group, it dropped to less ‘than 156. These results show that selective breeding can modify ‘quantitative taits ina very directed manner. ‘When comparing the curves in Figure 28.11, keep in mind that quantitative traits are often at an intermediate value in un- selected populations, Therefor, artificial selection can increase or decrease the magnitude ofthe tit In this case, oil content can go up or down, Figure 28.11 also shows the phenomenon known as 2 selec ton li tical selection is no longer effective, selection limit may occur for two reasons. Presumably, the starting population possesses a large amount of genetic variation, which contributes to the diver sity in phenotypes. By carefully choosing the parents, each suc: ‘ceeding generation has a higher proportion of the desirable alleles. However, after many generations, the population may be nearly ‘monomorphic fr all or most ofthe desirable alleles that affect the trait of interest. At this point, additional selective breeding will have no effect. When this occurs, the heritability for the tat is, near zero, because neatly all genetic variation for the trait of interest has been eliminated from the population. Without the it after many generations a plateau ig reached where ar