How many samples (tumor vs. normal tissues) are usually used for NGS based microRNA studies in order to have statistically meaningful results?
I think more is always better - but you have to consider financial and logistical limits...
what is the goal of the study? if it is just a trial, star with e.g. 5 samples each (tumor/normal). but for publication I suggest at least 15 to 20 patients per group
I think more is always better - but you have to consider financial and logistical limits...
what is the goal of the study? if it is just a trial, star with e.g. 5 samples each (tumor/normal). but for publication I suggest at least 15 to 20 patients per group
Bart summed it up well: You have to consider your limiting factors, i.e. money, time and quality of samples. YOu can do a hundred versus a hundred but with poor sample choices you may not get better results than with a well selected 3 versus 3 cohort.
And as Bart said: What is the goal, what do you want to learn from this experiment in the first place? There is normal and there is not malignant. Often "normal" tissue from a cancer patient may not be that normal when compared to a normal sample from a healthy individual. Then, there is tumor and there is tumor. How diverse is the group of tumors you are interested in, genetically, phenotypically, histopathologically etc.
Next: once you have the data you want to confirm them in a larger cohort by, I assume, qRT-PCR. With such a split, you may get away with a very tiny group of tumor versus normal for the NGS part. But the qRT-PCRs can become quickly expensive as well... Many things to consider. In the end you have to go along with a compromise unless you really have limitless resources. "Traditionally" I would go with 3 and 3 using very well selected samples. Since you do microRNA, you can do FFPE tissues, for which you can get/already have excellent histopathological evalutaions, if you have access to the path reports.
I’m agree that it depends of the goals of the study and the response that you will have. If you will achieve the quantitative interpretation of the primary output of NGS or simply will classify the response in a categorical scale (presence or absence of expression of miRNA). If you are looking for statistically meaningful results, you must calculate your sample size, considering such things and the significance level (or confidence level), precision and power of the statistical test to be held. Think about that you will compare two groups (tumor vs normal tissues) and that your results not only will be reported with a p-value of significance, a confidence interval of the difference will be necessarily calculated.
Dear Bastian,
I totally agreed with the previous answers. You need to know the goal and budgetof your study. The out put of NGS data must be validated. In my project I used a set of normal/tumor tissue samples form 17 patients.
Samples (tumor vs. normal tissues) required as per statistically meaningful results is actually huge and differ from cancer to cancer types.However, It takes time to collect valuable samples from appropriate hospitals. Therefore you need to go on collecting, experimenting and communicating. Always validate NGS data.
If it is a trial and will publication, I think you should be consider individual differences, 6 samples each (tumor/normal) is required. Good luck!
- Badges
- Science topic
- Similar topics
- Biological Science
- Anatomy
- Tissue