# Tumor size statistics in 4 groups of transgenic mice!! - (Apr/13/2009 )

I have 4 different groups of transgenic mice (A-D) and want to compare tumor sizes each other (group by group).

A=control (wild type), n=3 mice

B=gene B overexpressed, n=4 mice

C=gene C deleted, n=4 mice

D=gene B overexpressed & gene C deleted, n=3 mice

Hypothesis of this study is that overexpression of gene B or deletion of gene C may induce tumor (by increasing tissue weight), and both may synergistically induce tumorigenisis.

I sacrificed them, took tissues and measured their weights.

For example, their weights are as followed,

(unit=gram)

A= 0.3/0.4/0.5

B= 0.6/0.7/0.8/0.6

C=0/7/0.8/0.7/0.6

D=1.1/1.2/1.2

First I want to calculate average/mean (dont know what is correct term in this case).

Second I want to compare them group by group by calculating p values to see whether or not the difference is significant (our hyp is correct or not).

group A vs B

group A vs C

group B vs D

group C vs D

In short, I want to see if B/C is heavier than A and D is heavier than B/C. (I will say their differences are significant/our hyp is correct if p<0.05)

Now questions are followed.

1. Do I have to calculate average or mean?

2. What Test do I have to use to get p values?

3. When I draw the data above in Excel Bar graph, I want to present average/mean as well as 'deviate'. What deviate do I have to use? standard error(SE) OR standard deviation(SD)?

Thank you so much in advance!!

cancerous on Apr 13 2009, 10:03 AM said:

2. What Test do I have to use to get p values?

3. When I draw the data above in Excel Bar graph, I want to present average/mean as well as 'deviate'. What deviate do I have to use? standard error(SE) OR standard deviation(SD)?

Thank you so much in advance!!

1. average = mean

2. two-way ANOVA

3. SD

Hope this may help.

I agree with Minnie. You need some kind of statistic program to do that. Because your outcome may be affected by the interaction of two genes, then you need to use two-way ANOVA. I think either SD or SE is fine. After you get a p value from two way ANOVA, you have to run a post hoc test such as Fisher's PLSD, Tukey-Rramer, exact test to see any group pairs are significantly different or not.