# Thread: Cell growth - One way or two way anova?

1. ## Cell growth - One way or two way anova?

Hi all!

I'm having questions regarding some results I've got. I have been counting cells twice a day for a week. There are three different cell lines counted 11 different times (including plating).

I have done a one way anova in SAS with celltype and the number of counted cells, which shows there is no significance between the three cell lines. However, when I do a two way including the counts, there is significance between two of them. (Which also correlates to what one would assume from the graph made from the results)

Example of my input.
"
celltype countnr alive
1 1 6000
(...)
1 11 39083

2 1 6000
(...)
2 11 10208

3 1 6000
(...)
3 11 52063
"

Is it best to use the one way or two way in my results section?

2. ## Re: Cell growth - One way or two way anova?

Will there be more or less cells over the observation times? So are you starting with say, 60,000, and at every observation there are few cells (or deaths).

Were you using repeated measures ANOVA? Can you post your SAS code or output

3. ## Re: Cell growth - One way or two way anova?

Originally Posted by hlsmith
Will there be more or less cells over the observation times? So are you starting with say, 60,000, and at every observation there are few cells (or deaths).

Were you using repeated measures ANOVA? Can you post your SAS code or output
Originally Posted by hlsmith
Will there be more or less cells over the observation times? So are you starting with say, 60,000, and at every observation there are few cells (or deaths).

Were you using repeated measures ANOVA? Can you post your SAS code or output
I started with 6000 cells for each cell line. There will be more cells as time progress until they reach maximum confluency and the death rate increases.

My SAS codes are:
title ‘One way anova’;
proc anova data=cells;
class alive celltype;
model alive=celltype;
means celltype / t bon tukey cldiff alpha=0.05;
run;

title 'Two way anova ';
proc anova data=cells;
class Alive celltype countnr;
model alive = celltype countnr;
means celltype countnr / t bon tukey cldiff alpha=0.05;
run;
quit;

I forgot to mention that I had 3 different parallels for each cell line at every count. I calculated the average from these numbers and used it in the anova.

 Tweet