INTRODUCTION TO BEHAVIORAL GENETICS

PSYCH 3102   Dr Hewitt

HOMEWORK #4  ANSWER KEY

 

1. Type of relative                   Expected correlation                        Numerical value

                                                        in terms of h2                                                if h2 = 0.6

MZ twins                                                    h2                                                                 0.6

DZ twins                                                   ½ h2                                                              0.3

Parent w. adopted child                               0                                                0

Parent w. biological child                        ½ h2                                                0.30

Adoptive siblings                                         0                                                   0       

Biological full sibs                                    ½ h2                                                               0.3

Biological half sibs                                   ¼ h2                                    0.15

Cousins                                                     1/8h2                                             0.075

 

 

 

 

 

  

 

 

 

2.                                                                                         TRAIT

                                                         HT                 IQ                    N                           EXT

i.Is the trait influenced by genes?            yes to all

Reasons:-            in each case,  MZ correlation > DZ correlation

 

ii.Is there evidence of shared         no                  yes                     no                        no

family environmental effects?

                   Reasons:-              only for IQ  does  DZ correlation exceed ½ MZ correlation

 

iii.Is there any evidence for

non-additive genetic effects?             no           no                    no                      yes

                 Reasons:-     only for EXT is the DZ correlation less than ½ the MZ correlation

 

iv.Does the non-shared environment

account for more than 10% of variance? no      yes                  yes                 yes

                Reasons:-    only for HT does the MZ correlation exceed 0.9

 

 3. Heritability for this trait is     0

            Because   no variation in the phenotype is due to genes in this population

 

 

 

4. a, False – heritability is NOT the proportion of a phenotype that is passed on – phenotype is never passed on, only genes are. Whilst roughly 50% of alleles are passed on to each offspring by each parent, we never know exactly which alleles. The actual 50% is unique to each offspring.

 

 

 

  b. False – high heritability does NOT mean genetic determinism. Whilst a high heritability implies most variation for the trait comes from genes not environment and also that the phenotype of the person is a good reflection of their genotype, the environment can (and does) change or can be manipulated so that phenotype is changed

  examples of this :  the change in IQ, height over time – mean levels of both have been rising over time, heritability for both remains high

 

  c. False – heritability estimated from one group can NOT be used to imply anything about a second group – each heritability that is calculated refers to the group from which the data was drawn. Heritabilites may vary across groups, as might mean level of trait – but we cannot say anything about the causes of those group differences. The environment might differ between the groups, for example, and this may be the cause of the differences.

 

 d. False – a large heritability does NOT imply genes of large effect. We can say that genetic variation is important in determining level of trait but this could come about through the action of a small number of genes of large effect each, or from a large number of genes each with very small effect.

 

5.   you can work the c2on the allele frequencies or the genotype frequencies but you have to take into account the facts that genotype (allele) frequencies and sample sizes are all different . You would use a null hypothesis of NO Association and calculate the expected on the basis that each class will have a frequency that reflects sample size and genotype frequency ONLY  ie there is NO effect of genotype on case versus control.

  

Obs                                                      DAT1 genotype

                                                9/9                   9/10                 10/10

                        Controls          9                      55                    98        /162

                        Cases               10                    71                    129      /210

                                                19                    126                  227                  Total = 372    

 

 

Exp                  Controls          19x162                        126x162          227x162         

                                                 372                 372                 372

 

                        Cases               19x210                        126x210          227x210

                                                 372                 372                 372

 

Obs                  Exp                  (Obs-Exp)2/Exp

9                      8.17                             0.084

10                    10.64                           0.038

55                    54.08                           0.016

71                    70.56                           0.003

98                    98.85                           0.007

129                  127.12                         0.028

                                                e = 0.176    5df        Prob > 100%    There is NO association