Smolke:Protocols/Primer Design

Josh's Version

 * Use IDT to estimate Tm's.
 * NB: Use the correct concentrations - significantly, we usually have 0mM Na, ~2mM Mg, 0.8mM dNTPs. The magnesium makes a big difference in the estimated Tm.
 * Use Nupack to estimate primer interactions (inter- and intra-molecular)

Sequencing/amplification primers
These are the easiest primers you're going to design. They're usually short (~18-24bp), which means there are few interactions to worry about. Some notes:
 * I aim for an initial Tm of ~60C
 * Try to avoid repeats (>3 of the same nucleotide) and regions with particularly low or high GC content.
 * Vary the length and, if possible, the location of the primer to get a good length, Tm, and GC content. Check Nupack to make sure there are no problems (hairpins or dimers)

Primers with tails
These are a little harder - they tend to be longer (since you're adding the tail) and often have dimerization regions (just from the palindromic restriction sites that you're probably adding in the tails)
 * I again aim for a Tm of ~60C for the initial binding region (the part that's homologous to the template). Also calculate the Tm for the entire primer. You may be able to raise the annealing temperature partway through your PCR.
 * Nupack is very important here, since you'll probably have inter and intra molecular binding. Make sure to consider both of your primers simultaneously (so 2 strand species with a maximum complex size of >2)
 * Intramolecular binding is often okay. As long as your 3' end is unpaired, the primer will probably work. But, if possible, avoid strong hairpins even in the middle of the primer
 * Intermolecular binding is sometimes okay. If the binding is head to head (leaving the 3' ends hanging loose), the worst that's likely to happen is that the polymerase chews back that unpaired 3' end. If the binding is tail to tail (leaving the 5' end unpaired), you'll get extension off the 3' end, leading to primer dimers. That's a bigger problem, and one you should almost always avoid
 * It often takes me a couple rounds of iteration to get decent primers - get the Tm's close, then check interactions. Then modify the primers to remove the obvious interactions, recheck Tms, recheck interactions
 * If you're adding a restriction site, add >4bp of random sequence flanking the site - restriction enzymes often don't cut well near the end of a linear piece of DNA