Remove the bottom portion of your fly trap (the part with the banana mash in it). Put the rest of in the ice-filled cooler, upside down. Chilling flies is one way to anesthetize them; in a few minutes, they should be asleep at the bottom of the contraption.
Get an empty fly vial, a funnel, and a foam plug. Put the funnel in the vial, get your fly trap, and quickly (before the flies warm up) knock the flies into the vial. Once you have all the flies, put the foam plug in the vial.
Get the bottle of Flynap and a Flynap wand. Dip the wand in the Flynap, touch it against the side of the Flynap bottle to get off the excess, then insert it into the fly vial. Try not to touch the vial or plug with the Flynap, as it will dissolve plastic. Leave the wand in for a few minutes, since you won't need to wake these flies up.
Label a 96-well microtiter plate with a piece of tape with your name, the date, and your collecting location on it.
Once your flies are asleep, knock them onto a piece of paper. Then put one fly in each well of the microtiter plate, starting at the upper left. If you're not sure whether a fly is D. melanogaster, ask me.
With the repeating pipettor, add 25 microliters of starch gel grinding buffer to each well that contains a fly.
Put a lid on the microtiter plate; hold it on with a rubber band. Record how many flies you caught. Put the microtiter plate in the cooler. Once everyone is done, I'll put the flies in a −80°C freezer until we're ready to run gels on them.
Knock the adult flies out of your mutant+wild-type vial into a clean vial (one without food in it), using a funnel, the way you did on Tuesday.
Put the flies to sleep by dipping a Flynap wand (they look like tiny bottle brushes) in the Flynap, wiping the excess on the inside of the Flynap bottle, then inserting it in the vial. Try not to get Flynap on the foam plug or the inside of the fly vial, as it will dissolve the plastic and flies may get caught in the sticky mess. Watch the flies carefully, and once they all stop moving, gently tap them onto a piece of white card. Don't leave them in the vial with the Flynap for too long, or they might die. (It doesn't matter if today's flies die, but you'll need flies to revive for experiments later in the semester, so it's good to practice). Observe the flies under the dissecting scope and divide them into two groups, one mutant and one wild-type. For most of you, the difference should be pretty obvious; some of the mutant phenotypes may be rather subtle, however. Draw a mutant and a wild-type fly, and label the difference between them.
Next, divide each group of flies into males and females. This is a little more difficult. Have your lab partner check your results, then ask me to take a look. From the top (dorsal view), males have a more rounded, darker tip to their abdomen. With the flies on their backs (ventral view), the genitalia are noticeably different, too.
Record the number of wild-type males, wild-type females, mutant males, and mutant females. These are the parents of the next generation of flies, which you'll count in two weeks. Discard the flies in a fly morgue (a bottle with some mineral oil in it).
Return the vial of larvae to the big tray of vials.
Now comes the fun part--you get to analyze the data and see if there are any interesting results. You'll calculate genotype and allele frequencies, then do statistical tests to see whether any of the deviations you've found are so large that they're unlikely to have arisen just by chance.
Count phenotypes. The first step is to count the number of phenotypes for each locus. Do this separately for each sex at each location. I collected data on 50 cats at a shelter in Kirkwood, Delaware, and I'll use it as an example. Here are my results on them:
female male total
short hair 19 26 45
long hair 3 2 5
white 1 1 2
colored 21 27 48
>half white 5 4 9
<half white 9 13 22
no white spotting 7 10 17
unknown 1 1 2
orange 13 19
orange+black 8 0
black 0 8
unknown 1 1
The next step is to calculate the allele frequencies. For the spotting and orange loci, you can count the alleles. For example, my sample has 13 orange females (OO) and 8 orange+black (Oo) females, so there are 34 O alleles and 8 o alleles. The frequency of the o allele is 8/42=0.190.
For the longhair and white loci, you'll have to estimate the allele frequency using the Hardy-Weinberg relationship. Under Hardy-Weinberg equilibrium, the proportion of recessive homozygotes is equal to the frequency of the recessive allele squared. So to estimate the frequency of the recessive allele, take the square root of the proportion of the recessive phenotype. For example, in my sample I have 5 long-haired and 45 short-haired cats. Long hair is recessive, so I calculate the frequency of the ll genotype, 5/50=0.10, then find the square root of 0.10, which is 0.316. So my estimated allele frequencies are 0.684 for L and 0.316 for l.
Test goodness-of-fit. You will do three different statistical tests today. The first is the chi-squared test of goodness-of-fit. This tests whether observed data fit a theoretical expectation. For example, one theoretical expectation is that half of the cats will be female and half will be male. In my sample, there are 28 males and 22 females. That looks like more males than females, but before I start speculating about why there are more male cats in shelters than female cats, I need to find out if the difference could have happened just due to chance. To do this, go to http://udel.edu/~mcdonald/statchigof.html and download the spreadsheet. You can read more about the test there, as well. On the spreadsheet, I enter my observed numbers (28 and 22) and my theoretical expectation (0.50 and 0.50), then look at the P-value. The P-value is the probability of getting a deviation as big as in my observed data, if my "null hypothesis" (that half the cats are male and half female) is true. When I enter my data, I get a P-value of 0.40. The usual rule, and the one we'll use in this class, is that if the P-value is less than 0.05, the result is "significant"; there is less than a 5 percent probability of getting results that extreme just by chance, so we reject the null hypothesis and conclude that something else is going on. Since my P-value is greater than 0.05, I do not reject the null hypothesis that there are equal numbers of female and male cats in the shelter.
Once you have the allele frequencies for the spotting and orange loci, you can calculate the expected genotypes using the Hardy-Weinberg relationship. For my data on orange in female cats, the frequency of o is 0.190 and the frequency of O is 0.810. The expected frequency of oo is then 0.1902=0.036, the expected frequency of OO is 0.8102=0.656, and the expected frequency of Oo is 2x0.190x0.810=0.308.
Enter the expected proportions and the observed numbers in the spreadsheet. Because I had to estimate one parameter (the allele frequency) from my data, the degrees of freedom are reduced to 1, so enter that in the appropriate space (under "Degrees of freedom (intrinsic hypothesis)"). The result is P=0.28, so I don't reject the null hypothesis, that the data fit the Hardy-Weinberg proportions. Test your data for spotted (do the totals for all cats), and test your data on orange (females only, obviously). Do this for each location; don't combine your two geographic locations.
Tests of independence: locations. The chi-square test of goodness-of-fit compares observed frequencies to those expected from some theoretical expectation. When you are comparing frequencies between two or more samples, you should use the chi-square test of independence. Read about this test at http://udel.edu/~mcdonald/statchiind.html, download the spreadsheet, enter your data, and record the P-values. For orange and spotting, compare allele frequencies between your two locations. For longhair and white, compare phenotype frequencies.
Lab report. Your lab report, which is due Tuesday, Sept. 8, should start with a Methods section. This should not regurgitate the information on this web page. It should say which geographic locations you looked at, then describe any particular procedures or problems you had for each gene. For example, if you had a cat that was described as having "medium" hair but it looked pretty short to you, what did you do? Did you skip over any cats; if so, why?
The Results section will mainly consist of a table containing your raw data (as I've done above), a similar table with allele or phenotype frequencies, and a table showing the results of your significance tests. Your written summary should just briefly mention any significant results; you shouldn't describe everything that is in the tables.
Your Discussion section should start by talking about any significant differences you found. Speculate about why the frequencies might be different. If none of your statistical tests are significant, pick the one or two that are the closest to significant and talk about them. Then describe further experiments you could do to test your ideas. Don't just say you want to sample more cats, or cats from more locations; propose some creative, original experiments to test your hypotheses.
You should then talk about the sampling technique. What inaccuracies might result from looking at pictures of cats in animal shelters? Are the problems with the data big enough to account for any significant differences you found? How could you get better data?
Your lab report should be typed, double-spaced (the tables can be single-spaced). There is no minimum length, but I expect it should take at least one or two pages of text (not including the tables). You may work with your lab partner on the tables, but the rest of the lab report must be your own, individual work.
Return to the Genetics Lab syllabus
Return to John McDonald's home page
This page was last revised August 29, 2009. Its URL is http://udel.edu/~mcdonald/geneticslab1.html