[TheDannyManCan]

Hey there,

Okay, so I understand the general idea behind the chi-squared test: to compare observed results to expected and establish whether the difference has any statistical significance i.e. whether or not it's down to chance alone.

However, I'm a little bit puzzled with something. I'm trying to do a bit of research into whether there's a gene for homosexuality - let me note at this point that I don't want this to turn into a discussion of whether there is or is not. I'm just trying to analyse some results. Sometimes in Bio we do presentations on Friday, and I'm just doing this out of interest to try and present something relevant (chi squared) in context and current (all this business about the signs on London buses).
In any case, I've looked at a few studies which seem to suggest that there's a gene in the q28 region of the X chromosome. The results apparently suggest that it is maternally linked: http://www.osti.gov/energycitations/...sti_id=5957271
It says that out of 40 sets of brothers, 33 of them shared a set of 5 markers in the q28 region. Would we expect only 20?

This, however - http://www.springerlink.com/content/r474023v2p056242/ - puzzled the hell out of me. I have no idea how they worked out x^2 for any of the results.

I'm not even sure where to start. If there were a gene for homosexuality that was passed down maternally, a son would have a 50% chance of inheriting it, right?
If so, then if in a study of 120 males whose mothers were carriers, could we expect 60 of them to have inherited it? If that's right (and I'm starting to get way out of my understanding...), then it could be compared to the observed data of the actual study and then chi squared could be calculated.

I am just in need of someone with greater biological knowledge than myself to help me figure this one out as I am in a massive muddle - any help would be much appreciated!

[ch0wm4n]

(Original post by TheDannyManCan)
Hey there,

Okay, so I understand the general idea behind the chi-squared test: to compare observed results to expected and establish whether the difference has any statistical significance i.e. whether or not it's down to chance alone.

However, I'm a little bit puzzled with something. I'm trying to do a bit of research into whether there's a gene for homosexuality - let me note at this point that I don't want this to turn into a discussion of whether there is or is not. I'm just trying to analyse some results. Sometimes in Bio we do presentations on Friday, and I'm just doing this out of interest to try and present something relevant (chi squared) in context and current (all this business about the signs on London buses).
In any case, I've looked at a few studies which seem to suggest that there's a gene in the q28 region of the X chromosome. The results apparently suggest that it is maternally linked: http://www.osti.gov/energycitations/...sti_id=5957271
It says that out of 40 sets of brothers, 33 of them shared a set of 5 markers in the q28 region. Would we expect only 20?

This, however - http://www.springerlink.com/content/r474023v2p056242/ - puzzled the hell out of me. I have no idea how they worked out x^2 for any of the results.

I'm not even sure where to start. If there were a gene for homosexuality that was passed down maternally, a son would have a 50% chance of inheriting it, right?
If so, then if in a study of 120 males whose mothers were carriers, could we expect 60 of them to have inherited it? If that's right (and I'm starting to get way out of my understanding...), then it could be compared to the observed data of the actual study and then chi squared could be calculated.

I am just in need of someone with greater biological knowledge than myself to help me figure this one out as I am in a massive muddle - any help would be much appreciated!

Yes your correct that they would have a 50% chance, however the sample used is far to small for reliability reasons.

[Turkishee]

I'm not going to read either of those links, but may I suggest going down an alternative route, perhaps rather than calculating chi squared result to demonstrate homosexuality, why not look at other phenomenon such as klienfelters Or turners or even attempt to look at environmental effects upon the epigenetic makeup and whether this could potentially lead to a case of homosexuality. You may have more luck that way and be more convincing.

[TheDannyManCan]

(Original post by Turkishee)
I'm not going to read either of those links, but may I suggest going down an alternative route, perhaps rather than calculating chi squared result to demonstrate homosexuality, why not look at other phenomenon such as klienfelters Or turners or even attempt to look at environmental effects upon the epigenetic makeup and whether this could potentially lead to a case of homosexuality. You may have more luck that way and be more convincing.

Thanks, but I am specifically trying to make a chi-squared link because it's what we've just done in class and I want to keep it relevant. I'm not really aiming to be convincing; I don't really mind what the results say, as long as I can understand and analyse them - thanks though