Fishers exact test on 1x2

#1
Hello everyone,

I have 2 groups, cases and controls, and I observed 16 variants in group of cases and 3 variants in group of controls.

I would like to test is there a significant difference between 16:3. I thought of doing Fishers exact test, but not sure how it can be performed on 1x2 table.

All suggestions are welcomed. Thank you.
 

Karabiner

TS Contributor
#2
Did you observe 16 variants, or did you observe 16 subjects with variants?

If it was the latter, then for a 2*2 table you need to add the two groups cases without variants and controls without variants.

With kind regards

Karabiner
 
#3
Did you observe 16 variants, or did you observe 16 subjects with variants?

If it was the latter, then for a 2*2 table you need to add the two groups cases without variants and controls without variants.

With kind regards

Karabiner
I observed 16 variants in group of cases, while I 3 in controls. My main goal would be to conclude that there is a overrepresentation of detected variants in cases.
 
Last edited:

Karabiner

TS Contributor
#4
I do not know what variants are. Do 16 variants mean that 16 different subjects had 1 variant each?

You iniitially said 3 variants in the control group Now you say there were no variants in the control group?
 
#5
I do not know what variants are. Do 16 variants mean that 16 different subjects had 1 variant each?

You iniitially said 3 variants in the control group Now you say there were no variants in the control group?
My mistake, it is 16:3.

It is 16 variants in 16 individuals, 3 variants in 3 individuals. It is the analysis in genetics, so missense variants were observed with high pathogenicity score.

Preferably not to include number of individuals in case/control group, because the main observation is 16:3 missense variants. The goal is to conclude, as expected, we observed more damaging variants in cases vs controls, and p value is significant.
 

Karabiner

TS Contributor
#6
I am not familiar with a statistically significant p-value as the research goal,
instead of valid and reproducable findings. So I'm afraid I don't have anything
to add. But it is interesting to get to know how genetics works as a science.
 
Last edited:

hlsmith

Less is more. Stay pure. Stay poor.
#7
Are cases and controls matched?

So you have 3 variants out of ??? controls and 16 variants out of ??? cases? Please post denominators. Did you look for any other genes or SNPs in these people? Making sure you don't need to penalize yourself for double dipping. Also, present Odds Ratios with precision estimates not pvalues.