Chi-square

I. Objective. To utilize the Chi-square test to determine whether

experimentally obtained data constitutes a good fit to an expected ratio or
value. To interpret a Chi-square value in terms of the probability of an event
occurring. To correctly identify the null and alternative hypotheses when
using a Chi-square test.
II. Principle. The Chi-square test is commonly used to demonstrate how
closely an experimentally derived value agrees with an expected value. More
specifically, chi-square is one among many different statistical tests designed
to provide an estimate of the probability that an observed value did or did
not occur by chance alone. This principle can best be described with an
example.
Example 1: A scientist wants to sample a population of birds for the ratio of
males to females. It is generally expected that these birds will exist in a ratio
of 50:50. In other words, if you collected one hundred of these birds 50
would be male and 50 would be female. However, when the scientist
actually collects 100 of the birds he collects 45 males and 55 females.
This issue raises the following important questions:
Are the observed values close enough to the expected values to be
scientifically accepted?
Were the scientists observed values due to chance occurrence or is there
actually a difference in the gender ratio of this bird?
Example 2: Everyone knows that when you flip a coin, you have a 50
percent chance of heads and a 50 percent chance of tails. This means that
out of 100 flips you should get 50 heads and 50 tails. However, if you
actually flip a coin 100 times, a 50:50 ratio is only one among many possible
outcomes. What if you get a ratio of 48 heads to 52 tails? Is this really the
same as 50:50? How can we be certain that what we observe agrees with
what we expect taking into account the role that chance plays?
Fortunately, we have a 2 (chi-square) formula for determining what is
chance and what is not chance:

Where O is your observed value and E is an expected value.

If you take our previous coin example:

Hypothetical Chi-Square
Coin
Observe
Expected
d value
value
Heads
48
50
Tails
52
50
Total
100
100

(O-E)

(O-E)2

(O-E)2/E

-2
2

4
4

.08
.08
.16

You get a 2 value of 0.16. However, this value means nothing without a
table of 2 values.
Degrees of
Freedom
(df)

Probability (p)

0.95

0.90 0.80 0.70 0.50 0.30

0.20

0.10

0.05

0.01

0.001

0.004 0.02

0.06

0.15

0.46

1.07

1.64

2.71

3.84

6.64

10.83

0.10

0.21

0.45

0.71

1.39

2.41

3.22

4.60

5.99

9.21

13.82

0.35

0.58

1.01

1.42

2.37

3.66

4.64

6.25

7.82

11.34 16.27

0.71

1.06

1.65

2.20

3.36

4.88

5.99

7.78

9.49

13.28 18.47

1.14

1.61

2.34

3.00

4.35

6.06

7.29

9.24

11.07 15.09 20.52

1.63

2.20

3.07

3.83

5.35

7.23

8.56

10.64 12.59 16.81 22.46

2.17

2.83

3.82

4.67

6.35

8.38

9.80

12.02 14.07 18.48 24.32

2.73

3.49

4.59

5.53

7.34

9.52

11.03 13.36 15.51 20.09 26.12

3.32

4.17

5.38

6.39

8.34

10.66 12.24 14.68 16.92 21.67 27.88

10

3.94

4.86

6.18

7.27

9.34

11.78 13.44 15.99 18.31 23.21 29.59

Non-significant

Significant

To correctly interpret our 2 value we need to know our degrees of

freedom (df).

df= (n-1)
n is generally described as the number of values that are allowed to vary
freely. However, this definition seems a little arbitrary. So maybe an
example would be better:
Imagine that you have 5 dice and you are trying to reach the number 15 by
throwing each of the five dice one time.
The first dice is a 5. The second dice is a 6. The third dice is a 1. The fourth
dice is a 2. At this point, the dice show a total of 14. However, you want a
total of 15. The last dice can only be a 1. Therefore, 4 of your values were
able to vary freely before the last value was forced or restricted to be a
1.
If the first 4 tosses of the dice were other numbers the fifth dice would still
be the one whose fate was predetermined by the first four dice. In this
way, n=5 and (df) = n-1.
In our coin example, n=2 because we have 2 categories; heads and tails.
Since, (df) =n-1, our (df) =1.
The value of 0.16 corresponds to a probability (p value) of between 0.50 and
0.70 (it is in-between two values listed on the table). At this point it is
absolutely critical that the proper conclusions be drawn from the
probability values obtained. Statisticians and scientists generally accept a P
value of .05 as the cut-off between values that are significant and nonsignificant. Since our P value is between .50 and .70 and is more than .05
we can say that the difference between our observed values and expected
values is NOT significant. What this really means is: Flipping a coin
100 times and getting 48 heads and 52 tails is due to normal chance
and is not due to some other force since a ratio of 48:52 is not
significantly different than a ratio of 50:50.
Unfortunately, things are not quite that simple. What you really did by
2
conducting a test on your coin data was to create two hypotheses.
Hypothesis 1: There is no significant difference between our observed
and expected values -or- The variability we observed in coin tossing is due to
chance since our values are not significantly different. This is also called our
null hypothesis.
Hypothesis 2: There is a significant difference between our expected and
observed values -or- The variability we observed in coin tossing is due to
some other factor since our values are significantly different. This is also
called our alternative hypothesis.

It is critical that the proper hypotheses be identified when interpreting a

value. Otherwise, it becomes very easy to confuse what is actually being
tested.

Conducting your own Chi-square

4 different kinds of beans have been mixed and placed into a large
container. The beans should all exist in equal amounts or a ratio of 1:1:1:1.
Remove a random sample of the bean mixture. Separate and count
each bean type. Fill in the table below.

Calculation of 2 for a sample removed from a population

Type of Observe Expecte (O-E)
(O-E)2
(O-E)2/E
Bean
d (O)
d (E)

Totals

2=

1) What is our null hypothesis (in terms of bean sampling) for this test?
2) What is our alternative hypothesis?
3) How many degrees of freedom are there?
4) What probability does your 2 value correspond to?
5) Is this value significant or not significant (according to the 2 table)?
6) Which hypothesis is supported by the results of your 2 test?
7) Please write your interpretation of the results of this experiment (what do
your results tell you about your sampled population of beans)?

After the P values for each class member have been tabulated on the board,
answer the following question.

8) Why do the P values vary within the class when the actual ratio of beans
is 1:1:1:1?

The Laws of Mendelian Inheritance and Maize

I. Objective: To determine the phenotype and genotype of F2 corn kernels
for two different genes. To construct a Punnetts square using monohybrid
and dihybrid crosses. To determine the expected genotypic/phenotypic ratios
of F2 corn kernels. To determine the actual phenotypic/genotypic ratio for F2
corn kernels. To apply a 2 test to your observations.
II. Principle: Corn ears are excellent models for Mendelian genetics because
each kernel represents an independent crossing event. Therefore, each ear
represents hundreds of independent crossings.
In this lab we will work with two corn genes that conform to Mendels laws of
inheritance.
A. Mendels law of random segregation: Diploid germ-line cells of
sexually reproducing species contain two copies of almost every
chromosomal gene. The two copies are located on members of a homologous
chromosome pair. During meiosis, the two copies separate, so that a gamete
receives only one copy of each gene.
Random segregation can be demonstrated by a monohybrid cross. In a
monohybrid cross, a parental cross is made between two individuals that
differ in the genotype of one gene. The offspring of the parental generation is
called the F1 (first filial) generation. The F1 generation can be allowed to
interbreed or self-fertilize (inter se cross, or selfing) to produce the F1
(second filial) generation.
For example, consider a monohybrid cross involving kernel color in maize.
There are
several genes that control seed color in maize. One gene for seed color is
designated by the letter R and has two alleles.
Seed color gene with two alleles:
R= purple (or red) allele (dominant allele)
r = yellow (or white) allele (recessive allele)
Possible genotypes and phenotypes:
R R = purple phenotype

R r = purple phenotype
r r = yellow phenotype
Consider the following parental cross in which the silk (female flower) of a
homozygous purple plant is fertilized with the pollen from a homozygous
yellow plant.
Parental (R) generation genotype:
Parental (R) generation phenotype:

RR
Purple

rr
Yellow

According to the law of random segregation, each ovule or female gamete

receives one copy of the R allele when it is formed during meiosis. Each
pollen grain or male gamete receives the r allele. Therefore, the offspring
of this cross (first filial or F1 generation) must be heterozygous Rr and will
display the purple phenotype.
F1 generation: All Rr Purple
The formation of the F1 generation is shown in the following diagram.

Another way to analyze the outcome of a monohybrid cross is the Punnetts

square. The Punnetts square method uses a simple grid to match all of the
possible combinations of gametes in a cross. The gamete genotypes of one
parent are listed along the top of the grid, and the gamete genotypes of the
other parent are listed along the side of the grid. By filling in each square
with the alleles from the top with the alleles from the side, the different

possible combinations of genotypes in the offspring are found in the grid. The
following diagram shows a Punnetts square analysis for the monohybrid
cross that was just covered.
Parental (P) generation genotype:
RR
Parental (P) generation phenotype : Purple

rr
Yellow

Punnetts square analysis for the parental cross:

9) If we used the above offspring (F1) in a new cross, what would be the
genotypes and phenotypes of the offspring of the F2 generation?

10) What is the genotypic ratio of the F2 generation?

11) What is the phenotypic ratio of the F2 generation?
12) What would be the genotypic and phenotypic ratios of offspring from a
cross between a homozygous recessive parent (r r) and a heterozygous
parent (R r)?

13) The ear of corn in front of you (#1) is from the F2 generation of a
controlled cross. Using what you now know about Punnetts squares,
monohybrid crosses, and phenotypic/genotypic ratios, how could you
determine the genotype and phenotype of the F1 parents?

14) What is the phenotype and genotype of the F1 parents using your
method from the above question?

15) What is the phenotype and genotype of the F1 parents of corn #2?
B. Mendels Law of Independent Assortment: When the alleles of two
different genes separate during meiosis, they do so independently of one
another unless the genes are located on the same chromosome (linked). This
is the principle of independent assortment. Mendel discovered independent
assortment by performing dihybrid crosses in the pea plant. We will examine
dihybrid crosses in maize. Consider the genes for kernel color and kernel
composition in maize.
Seed color gene
Seed composition gene
R allele, dominant, for purple kernels
T allele, dominant for smooth
(starchy) kernels
r allele, for yellow kernels
t allele, for wrinkled (sweet)
kernels
In the P generation, a homozygous plant with purple, smooth kernels was
crossed with a plant having yellow wrinkled kernels. The F1 plants were
allowed to fertilize themselves. According to the principle of independent
assortment, the color gene and the seed shape gene should not affect one
another; that is, they should behave independently.
It is also possible to analyze the dihybrid cross with a Punnetts square. For a
dihybrid cross, well need a 4 x 4 grid because there are four genotypes in
the F1 gametes. Here is the Punnetts square analysis of the F1 cross from
the above example.

F1 cross= R r

Tt

x Rr

Tt

RT

RR TT

RR Tt

Rr TT

Rr Tt

Rt

RR Tt

RR tt

Rr Tt

Rr tt

rT

Rr TT

Rr Tt

rr TT

rr Tt

rt

Rr Tt

Rr tt

rr Tt

rr tt

RT

16) Using what you have learned about Mendels law of random segregation
and independent assortment, why are all possible combinations of genes
used in the dihybrid Punnetts square? In other words, along the top and
side of the above Punnetts square, every possible combination of genes
from each of the parents (F1) is represented. Why is this necessary? What
does it account for?
17) What are the phenotypic and genotypic ratios of the F2 offspring from
the above crosses?
18) Construct a dihybrid Punnetts square for F1 parents with genotypes
different (your choice) from the above example. What are the phenotypic
ratios of the F2 offspring?
19) In front of you there is offspring (#3) from the F1 parents of a dihybrid
cross. Using a chi-square test, and what you have learned about Mendels
laws, determine the genotype and phenotypes of the F1 parents. To answer
this question properly, you will need to construct a chi-square table and
determine your expected and observed values for each phenotype. In
addition you need to correctly identify your null and alternative hypotheses
when conducting the chi-square test.