I hope you are including a control (without any treatment). You can do students T test where you compare two groups at a time (compare each treated with the control). You can do this test individually across the treatments (tomoxifen, nanoparticles etc) and the readouts (trypon blue, MTT, real-time PCR). I hope this helps.

Hi,

you can use graphpad to apply the statistics to your data.

You can use, annova, variance, student T tests to show that your results are significnt.

NOTE: Don't forget to include error bars to the graphs..!!!

https://www.graphpad.com/scientific-software/prism

-- best regards

You can use student T-test if you are using Microsoft Excel.

thank you so much

Hi Mahmoud,

I see a few issues in your experimental design, but it was a very wise idea to ask your question in order to try and do the experiment correctly:

1) Statistical analysis is possible only on independent samples. Replicates are not independent samples, because they stem from the same culture. This means that if you are looking for an n=3 (that might be a bit thin for the comparison you wish to do, and since you work on cells doing a few more experiments will not be a big deal), you need to apply your treatments to three (or better more) independent cultures, which means you thaw your cells three (or more) times and repeat your experiment as you described (or almost... as has already been observed!).

2) It's good to do replicates, but they represent only 1 biological value. So from replicates you can take an average and check that the variance is below 5 %, this will give you an idea of your experimental accuracy.

3) Before applying a statistical test, you should verify how your independent values are distributed. If they are normally distributed you can analyze them with an ANOVA or a t-test or a corrected t-test (see below); if they are not normally distributed then you have to use non-parametric tests. Normality is verified with a test: if you have many data points a one-sample Kolmogorov-Smirnov test will suffice, otherwise other tests will be better, such as the Shapiro-Wilk test: in any case having more than 3 independent values will allow you a more reliable analysis.

4) With normally distributed data, multiple pairwise comparisons with a t-test are not correct, because you might easily encounter false positives (so-called type I errors). So you should use tests that introduce corrections for this problem by using a reduced significance value, such as the Tukey or Bonferroni tests. These are called post-hoc tests, because normally you apply them after an ANOVA analysis.

5) Your experimental design is missing the control condition as Ramachandramouli immediately noticed: you need to incubate cells without any treatment and count their survival, their death and do the qPCR analysis

6) Once you have your experiment with a minimum of 3 independent samples for each treatment and dose, assuming you should get normally distributed data, you should do a two-factor (treatment and dose) repeated-measures ANOVA analysis followed by the post-hoc tests mentioned above (Tukey or Bonferroni) on each of your readouts, the tripan blue count as well as the MTT absorbance and your qPCR data (designing a qPCR analysis is not trivial, take some time to plan well your calibration approach for the relative quantification).

Hi, Dr. Pietro Pilo Boyl and Dr. Ramachandramouli Budida

of course i have already a control group

when i have made Anova anlysis to mtt assay on prism i cant get the analysis because the program need more than three variable?

in the analysis of qpcr the genes tested is down regulated so i have negative ct value what i should made ?

ANOVA doesn't work where you have just two groups. You must have more than two groups to use ANOVA. However Student T Test can work but i think you need to include a control for your experiment.

Graphpad