Protocol Online logo
Top : New Forum Archives (2009-): : Macroarray and Microarray

Microarray Study Design - (Mar/18/2012 )


I am writing to seek advice regarding microarray study design.

I have 5 cancer cell lines: 1 cell line is normal and 4 are cancer related. My question relates to whether a normal versus cancer analysis using these arrays would have sufficient power to detect a difference between normal and cancer cell lines. Would this study be worth doing? I have previously read that you need at least 5 samples per group to find differences between groups. Is this correct?

Please advise what is the best course of action in this instance.

Thank yo in advance for your help.

-Casey Wright-

hi Casey,

how similar are those cancer cell liines? would you expect to find any differences among them compared to normal cells? i don't work with cancer stuff, but i assume that the mechanisms might be similar among those cells...

about the question that you need at least 5 samples, do you mean biological replicates? it's like any other analysis, the more replicates you have, the more reliable your data will be. after working with microarray data I found many significant differences between treatments by using only 3 replicates, however, each replicate consisted of the RNA exctracted from 2 independent individuals from the same treatment.

currently microarray chips are lot cheaper than they used to be. is your study novel? I could almost place a bet that there are already cancer-related microarrays publicly available. so investing in an experiment of this kind is only as worth as the novelty that it represents.

hope this is useful. good luck!