Wikipedia:Reference desk/Archives/Science/2018 October 1

= October 1 =

Toluene, water and isopropanol miscibility
I mixed one part toluene with one part 70% isopropanol and it separated into two phases. How do I determine what is in the lower phase? I'm guessing it's the water? — Preceding unsigned comment added by 185.230.100.66 (talk) 16:56, 1 October 2018 (UTC)
 * Density density density. More dense liquid sinks, less dense liquid floats. If you google fotr density of isopropanol water you will find various tables and specific values covering a wide range of percents. "70%" happens to be a standard product, and is directly available from simple product literature (i.e., no need to use scientific journals or paywalled sources), and probably even in the snapshot that Google provides on its list of hits. Of course our toluene article has the density of that pure substance. But once mixed, the more the IPA migrates into the toluene, the lower the %IPA in the aqueous phase and the denser the aqueous phase becomes. DMacks (talk) 17:59, 1 October 2018 (UTC)

Proportion of predicted progeny for a test cross of ABC/abc involving linkage
help i embarrassed myself as a tutor today. I don't have time to draw a diagram and a picture of the question would not be free content. Suppose there are three loci, A, B and C on a single chromosome, where A/B are 20 map units apart, and B/C are 10 map units apart:

• A — 20 — B — 10 — C •

Suppose we have a rat with the following genotype:

A B C ——— a b c

The question asks for the proportion of progeny with the genotype Aabbcc.

Whichever way I did it, I did not get 9%, which is what was shown on the answer key:

The only way to produce Aabbcc progeny is through Abc and abc gametes.

Assuming no double crossovers between same loci (but allowing for double crossovers between different loci): the gamete frequency for abc is 0.9 * 0.8 * 0.5 = 36% (chance of no crossover betwen either pairs of loci * half the gametes) and for Abc is 0.2 * 0.9 * 0.5 = 9% (chance of crossover between A/B * chance of no crossover between B/C * half the gametes) so the expected frequency should be 3.24%.

How is the answer key (9%) correct? Yanping Nora Soong (talk) 18:11, 1 October 2018 (UTC)


 * Disclaimer: I just discovered the topic of crossovers (but not probabilities).
 * There is an error in your reasoning when you assumed the first gamete must be abc and the second Abc (it can be the other way around) so you miss a factor of two. (Out of the 36 outcomes when you cast two dice, only one is a double 1, but two contain one 1 and one 2 - if that is not obvious to you, or for further reading, see .)
 * Furthermore, the probability of crossover is not strictly proportional to the distance so it is not exactly 80% and 90% chances of no crossover. The formula at Centimorgan should be valid only if double crossovers are allowed (since at infinite distance the probability of recombination tends to 1/2, which comes from the fact that an odd or even number of crossovers is equally likely), but yours (P = d/100) is pretty much never right I think (except for small-d approximations). Using the article formula I get crossover properties of P10=9.06% and P20=16.48% (instead of 10% and 20%).
 * Your reasoning for gamete frequencies seems correct to me and gives $$P(abc)=(1-P_{20})*(1-P_{10})*0.5=37.97\%, P(Abc)=P_{20}*(1-P_{10})*0.5=7.49\%$$. The final probability of genotype is then $$P(Abc+abc)=2*P(Abc)*P(abc)=5.69\%$$. While it brings one closer to the result, that is still not the answer key. Tigraan Click here to contact me 14:21, 2 October 2018 (UTC)
 * It's been a while since I've dealt with something like this but since we are talking about rats, which don't naturally self fertilise, have I misunderstood something or is there no way to give a single progeny diploid genotype frequency based on the question as worded? There is only one rat mentioned. We have no idea what the genotype of the other parent is. It could be anything e.g. abc|abc, ABC|ABC, a'b'c'|A'B'C'. You could only give a range from 0% (if the other parent has none of those alleles) to the maximum possible (assuming the most favourable parent to produce that combination when mating with the parent who's genotype is mentioned). Are you sure that the question didn't say something like 'assume both parents are ABC|abc' or 'if the rat self fertilises' (which is a very odd suggestion but I'm sure far from the weirdest test question) or something which would clarify the situation? Nil Einne (talk) 19:26, 2 October 2018 (UTC)


 * Your original calculation came up with a 9% figure for Abc. So the way to get the book's answer is to suppose they threw in the word "backcross" somewhere, or a comment about the phenotype of the other parent being abc, etc.  The assumption of no double crossovers between adjacent genes is reasonable, though the statement that they are 20 and 10 centimorgans apart (rather than having this observed probability of recombination) would make it an approximation. Wnt (talk) 01:45, 3 October 2018 (UTC)


 * I asked my genetics professor from eight years ago, and she made me realize I confused a test cross with a self-cross. Oops. Yanping Nora Soong (talk) 16:31, 3 October 2018 (UTC)
 * Remember that providing key question details in the subject heading only on the RD is quite confusing since it's fairly common that people will only skim over or even just not notice the heading. Nil Einne (talk) 19:33, 3 October 2018 (UTC)


 * Indeed so! Also, people tend not to Wikilink from headers since it is "frowned upon", whereas if the OP had linked test cross in the question, he or she might well have gotten to his own answer in the process. Wnt (talk) 01:51, 4 October 2018 (UTC)


 * I wrote this question from my phone, which is quite different in "feel" than writing from my laptop. Yanping Nora Soong (talk) 15:46, 4 October 2018 (UTC)


 * I didn't intend to criticize anyone but myself for not RTFQ. Wnt (talk) 12:28, 6 October 2018 (UTC)