In this experiment we had one control group and several test groups. The cell cultures were initially seeded under identical conditions, but had different initial cell numbers. The medium for all groups contained 25 mM glucose + substrates. For the first 24 hours all groups remained under these conditions. After that three wells per group were counted using the CASY Counter (day 0). After that the control group continued to receive 25 mM glucose, and the test groups received different glucose concentrations. Further substrates were added to all groups. Cell counts were then taken daily for 4 more days. Independent samples were used, with three replicates per group.
Since the initial cell numbers vary between groups, the absolute cell counts are not directly comparable. I am seeking an appropriate statistical method to analyze the data, specifically to determine whether glucose concentration has a significant effect on cell growth over time. The goal is to find a method that accounts for the different starting cell numbers while still providing meaningful analysis of the cell growth data collected over the 5-day period.
I have not done this myself before, but what I could think of is to use non-linear regression XY analyses in GraphPad Prism and choose exponential growth. Regardless of the differing starting number, you will get a value called doubling time. You can then use the doubling time of every replicates to analyze the data further based on the different input of glucose such as by using ANOVA. Alternatively, you can also plot the number of cells you counted as log and then plot it against time. Then, you will obtain a best fit curve along with a slope. Based on the slope itself, you can already tell which concentration of glucose has better growth stimulating effect, i.e., better when the slope is near 1. If you desire statistical analysis, you can pick an end-point or two (e.g., day 3 and day 4) and then again analyze them with ANOVA. Hope these make sense and useful to you.
- Badges
- Science topic