Question

1)Lets do a simple bayesian calculation.

60% of frogs are female, the rest male.

All male frogs sing, but only 10% of female frogs sing.

You hear a frog singing.

What is the posterior probability that this singing frog is male?

(hint! The probability of any frog singing is the probability  of a singing female (.6 x .1) plus that of a singing male (.4 x 1))

1)23%

2)35%

3)75%

4)87%

2)Which is the most complicated model for molecular sequence evolution?  

1)Jukes Cantor

2)F81

3)HKY

4)GTR

3)You run modeltest on your data and determine that you do not have equal frequencies of bases, but you don't see any evidence that transversions are any more or less common than translations. Which model should you use?

1)JC

2)F81

3)HKY

4)GTR

4)The support values on trees generated via bayesian methods convey:

1)How likely that clade is

2)What proportion of trees that are likely to come from the provided data contain the clade in question

3)How many substitions are found in the sequences between two taxa

4)The quality of the analysis

Answer 1

Ans 1: Given that the probability of males p(M)= 0.4 and probability of females p(F)= 0.6.

Now probability that a female sings p(S/F)= 0.1 and probability that a male sings p(S/M)= 1.

According to Baye's theorem: p(M/S)= p(M)*p(S/M)/(p(M)*p(S/M)+p(F)*p(S/F)

So putting the values in the given equation we get the probability that the frog singing is male and it is equal to 0.869 or we can say that ~ 87%. So option 4 is the right answer.

Ans 2: GTR model is the most complicated model for molecular sequence evolution because it deals with 633 parameters when it is used for studying codingregions of the genome. Option 4 is right.

Ans 3: In JC model there were several assumptions like equal base frequencies and equal mutation rate. F81 is the modification of JC model in which the base frequencies are not equal with each other but mutation rate is same. This is all we required in the question so thats why when base frequencies are not same but mutation rates are equal then we will use F81 model. So option 2 is right.

Ans 4: Option 2, i.e. what proportion of trees that are likely to come from the provided data contain the clade in question is the right answer because bayesian methods calculate the posterior probability for a single clade on a given tree.