BIO 111: FOUNDATIONS OF BIOLOGY

LECTURE 14: "Linkage and Mapping"

 

 

Assumed Knowledge:

a completely firm grasp of meiosis (particularly prophase I and metaphase I) and independent assortment.

Lecture Notes:

I thought that it might be helpful to provide these notes on the "how's and why's" of the Test of Independent Assortment. We will do another one in class, but here is an example for you to return to...:

 

    A. Complete Linkage

            Two loci that are so close that they are always inherited together.
            Dihybrid Cross: AABB x aabb = 100% AaBb
            Test Cross: AaBb x aabb. If AB and ab are linked, then progeny = 50%AB and 50% ab; no Ab or aB

    B. Incomplete Linkage 


            a. Basic Principles:
                    - Crossing over within a region is rare (so cross-over products are rare)
                    - Most of the time it will not occur, so parental types will be most common gametes produced.

            b.Test Cross: AaBb x aabb:
                        Suppose we observe the following phenotypic ratio in the offspring: 35 AB, 28 ab, 12 Ab, 20 aB.

AND NOW, SUPPOSE WE WANT TO ANSWER THE QUESTION: "ARE THE GENES ASSORTING INDEPENDENTLY OR ARE THEY LINKED??"

                        Well, that doesn't look like a 1:1:1:1 ratio like we might expect from independently assorting genes. But how different IS this ratio from what we would expect? Is it different enough to REJECT the hypothesis of independent assortment?

            c. ANOTHER OVERVIEW OF HYPOTHESIS TESTING:

                        a. PROBLEM: How deviant do your results have to be to reject the hypothesis of independent assortment?

                        b. SOLUTION: Determine the probability that your results could be produced by genes that are truly assorting independently. If this probability is really low, your genes probably AREN'T assorting independently - they are probably linked. (READ THIS AGAIN UNTIL IT MAKES SENSE....)
                        c. By analogy, suppose you have a coin and you want to know if it is perfectly balanced. Suppose you flip it 20 times and find a 15:5 ratio of Heads:Tails. For a perfectly balanced coin, you expect a 10:10 ratio. Suppose you can determine that the probability that A PERFECTLY BALANCED COIN yielding a 15:5 split, just by chance, is only 0.01%. That tells you that there is less than 1 chance in 100 that a perfectly balanced coin would deviate this much from a 10:10 expectation. But your coin DID yield this 15:5 split. That suggests that your coin is NOT perfectly balanced; it is behaving in a manner that is very unsualy (1/100) for a perfectly balanced coin to behave. Of course, if you conclude that the coin is not balanced, there is still that nagging 0.01% chance that it WAS balanced and that your conclusion is wrong. This process is called hypothesis testing; it is central to almost all statistical tests. The probability that chance could produce your results is called the "alpha" level. It is also called the "type I error rate" - defined as the probability of being wrong if you reject the hypothesis of equality. (If you conclude that the coin is imbalanced (that your observed results ARE NOT EQUAL TO those expected by your hypothesis), this is the probability that your conclusion is wrong - that you've made an ERROR). UNDERSTAND THIS LAST PARAGRAPH INSIDE AND OUT - IT WILL HELP YOU IN ALL YOUR SCIENCE COURSES

                        d. DETERMINING THAT PROBABILITY FOR IND. ASSORT. GENES: The Chi-Square Test of Independence:

                        - Suppose you did your Dihybrid test cross and got: 35 AB, 28 ab, 12 Ab, 20 aB.

                        - Could these genes be assorting independently? Well, anything is possible, but ARE THESE RESULTS LIKELY from independently assorting genes? How likely?
                        - Determine what you would expect if the genes assorted independently:
                                - For the A locus, by itself, there is a ratio of 47:48 of B:b out of 95 offspring.
                                - For B, there is a ratio of 55:40 of B:b.
                                - So, if the genes assort independently, you can predict the probability of phenotypic combinations by multiplying the independent probabilities together.
                                - To find the number of individuals, you'd have to multiply this final probability times the sample size (95). So, your expected frequencies,
under the hypothesis of independent assortment, are:
                                            AB = 47/95 (A) x 55/95 (B) = 0.286 x 95 = 27.21 AB
                                            Ab = 47/95 (A) x 40/95 (b) = 0.208 x 95 = 19.79 Ab
                                            aB = 48/95 (a) x 55/95 (B) = 0.293 x 95 = 27.79 aB
                                            ab = 48/95 (a) x 40/95 (b) = 0.213 x 95 = 20.21 ab
                                                                                                        TOTAL = 95 offspring

                        e. You can calculate these in a 2 x 2 Chi-square Test of Independence Table (also known as a contigency table) as:
                            (Row Total x Column Total)/ Grand Total

                        Compare observed vs. expected using the chi-square computation: X2 = SUM [(o-e)2/e]:

                                    observed        expected        o-e        (o -e)2/e
                        AB             35                  27.21        7.79     2.23
                        Ab              12                  19.79      -7.79     3.07
                        aB              20                  27.79      -7.79     2.18
                        ab               28                  20.21       7.79      3.00
                                                                                SUM = 10.48

                         f.  Interpreting the Chi-Square Value of 10.48 - Statisticians have determined how likely particular chi-square values are, just by chance, if the events truly ARE independent. They have determined that, with 1 'degree of freedom' (what you have:  (rows - 1)(columns - 1) )), a score of 3.84 will only occur 5% of the time. A score of 6.63 will only occur 1% of the time, just by chance, when two events truly ARE independent. Larger scores are progressively less likely. Your score of 10.48 signifies that truly Independently Assorting genes would ONLY yield results this deviant from what you would expect less than 1% of the time. - So, independently assorting genes are unlikely to yield your results. There is less than 1 chance in 100 that independently assorting genes would produce results that deviate this much from expectations. - The most appropriate conclusion is, then, that your genes are NOT assorting independently. You should conclude that they are linked.

 

HAVING CONCLUDED THAT THE GENES ARE LINKED, WE CAN NOW ANSWER TWO OTHER QUESTIONS:

HOW ARE THE ALLELES ARRANGED ON THE PARENT'S CHROMOSOMES, and

HOW FAR APART ARE THE LOCI?

1. Gene Arrangement

We now apply the first principle - that crossing over in a region is rare. So, the products of this rare event will be rare. So, we can identify the products of crossing over as the rarest phenotypes produced in the offspring. In this case, those are Ab and aB. These are called the recombinant types.

If crossing over is rare, then most of the time it does NOT occur. So, most of the time, the alleles are passed to the offspring in their original arrangement... and these phenotypes should be the most abundant phenotypes in the offspring. In our case, these are the AB and ab phenotypes. So, the original AaBb parent had the A and B alleles on one chromosome and the a and b alleles on the other homologous chromosome. When a cross-over occurs between these loci, the recombinant types (Ab and aB) are produced.

2. How far apart are the loci?

Well, now we apply the second principle - that the frequency of crossing over is proportional to the distance between loci. So, we measure the frequency of crossing over and can use that as a direct index of the distance between loci. So, there were 12 Ab and 20 aB offspring... for a total of 32 offspring produced by recombination events. this is 33.7% (32/95). So, we say that the genes are 33.7 map units (or centiMorgans) apart. Genes that are closer together will recombine less often... genes that are farther apart will recombine more often.

 

Consider these results from the following testcross:

                AaBb    x        aabb

        F1 Phenotypes:

                    AB        15
                    Ab         22
                    aB         29
                    ab         13

Conduct a chi-square tst of independence.  Using the p = 0.05 level (and critical X2 of 3.64), make a conclusion about whether the genes are linked or assort independently.

If the genes are linked, map their positions in the heterozygous parent, showing which alleles are on which homolog, and determining the distance between loci.