I'm going to undertake an siRNA experiment soon, but I have only read about them. I want to address the role an enzyme plays in processing a protein.

From what I understand, I will need to pick two specific siRNA's from a screen of multiple siRNA candidates. As a negative control, I should also make a non-sense scramble siRNA. Of course the specific siRNA should show a significant knock-down of the enzyme; the control should not affect expression. As well, there should also be a validated positive control (such as GAPDH) and a mock treated control.

Is this experimental design sufficient? What are some common things to keep in mind, or pitfalls to avoid when doing siRNA studies?

  • $\begingroup$ In which system are you working, I'd guess cell culture? And I assume you know about the rules of designing siRNA, so that the correct strand will be picked? $\endgroup$ Nov 7, 2012 at 8:40
  • 1
    $\begingroup$ @MadScientist I'm going to do this in mammalian cell culture (PC12 cells). Can you explain what you mean by correct strand? $\endgroup$
    – user560
    Nov 7, 2012 at 12:44
  • 3
    $\begingroup$ The stability of the ends of the RNA duplex determines which of the two strands is chosen as guide strand. You have to take this strand bias into account when designing your siRNA. $\endgroup$ Nov 7, 2012 at 14:34

1 Answer 1


First you should decide whether you want to design an shRNA or use siRNAs.

If you want to use shRNA you should look at the rules that "Mad Scientist" mentioned. You can insert mismatches but make sure not to disrupt the secondary structure. You can use RNAFold to verify.

shRNA always has issues and you have to optimize your design, test it by realtime PCR or Northern blot etc to confirm the production and the concentration.

You can design siRNAs and there are online algorithms to help you design one. You have to choose a sequence which is specific to your gene. This step is important for both shRNA and siRNAs in order to minimize/avoid off-targeting.

A common practice is to use a pool of siRNAs for a single gene. Pooled siRNAs are commercially available for many genes.

For the experimental controls you would need to do two experiments:

  • mock siRNA treated; which wont affect any gene
  • Your siRNA treated

then test for a reference gene such as GAPDH and your gene. A good reference gene should be the one which is not expected to change expression. GAPDH may not be a good reference always (for example in cases where metabolism is affected). The choice of reference gene is intuitive and depends on your prior knowledge of the system. Instead of GAPDH you can also transfect GFP and quantitate that as a control (spike in).


You must log in to answer this question.