Hi Brendan, your reasoning is theoretically fine and there will be cases in which you will need to use it, although in general in PCR it is very difficult to predict what is better in advance due to the very complex kinetics of the reaction.

I would suggest to always start with a simple protocol, in your case with the annealing temperature given by your 3' 20 nts. As Jesper says, it is most often sufficient, because although it is true that after the first cycles your tail will become part of the template, it is also true that after 10 cycles (to use your numbers) your specific template will be easily around 1000 times more abundant than any other sequence, that's why the annealing temperature will not matter much anymore.

The critical steps in PCR are always the first cycles, in those you have to ensure maximal specificity in order to increase the specific template efficiently avoiding producing at the same time unspecific material. That's why Mark suggests the step-down approach, which in case you obtain a dirty PCR at lower annealing temp, but no product at all or too little if you raise it, can work very well.

Keep in mind that independently from the global Tm of your primers, the way you design the 3' end of your primers (essentially the dG given by the annealing of the last 7-8 3' nts) will strongly affect the annealing temperature for your PCR reaction and the specificity of your primers. So, the most care in the design should be placed on the 3' side of the primers, avoiding sequences that can give rise to primer-dimers (GC, or GG and CC on FW and RV primers, etc.), avoiding too many GC pairing all together (try to have no more than two adjacent ones), etc...

Good luck!

Hi Brendan

In my experience the PCR will work just fine by using the lower temperature Tm throughout the PCR.

PCR is an extremely robust method that very seldom fails. I usually use a standard protocol of 25-35 cycles and a Tm of 55. This works for almost any PCR (exceptions include very long PCRs and repetitive template DNA).

With that said, your arguments are fully valid and it is always a good idea to try and anticipate potential problems. 

Good luck

Jesper

Yes you're right.  I have seen reactions where I needed to do the step-up protocol you speak of, that would not work in any other way.  

Hi Brendan

Could you not approach it the other way and use a "touchdown" protocol i.e. start at a higher annealing temperature for a few cycles to generate some specific product and then reducing the annealing temperature to increase the yield? I use this type of protocol to good effect for amplifying bacterial sequences from field samples with low bacterial loads. I use standard primers (some 25nt long) which gave non-specific products using a normal PCR protocol. The "touchup" protocol you describe seemed to give less yield.

Good luck

Hi Brendan, your reasoning is theoretically fine and there will be cases in which you will need to use it, although in general in PCR it is very difficult to predict what is better in advance due to the very complex kinetics of the reaction.

I would suggest to always start with a simple protocol, in your case with the annealing temperature given by your 3' 20 nts. As Jesper says, it is most often sufficient, because although it is true that after the first cycles your tail will become part of the template, it is also true that after 10 cycles (to use your numbers) your specific template will be easily around 1000 times more abundant than any other sequence, that's why the annealing temperature will not matter much anymore.

The critical steps in PCR are always the first cycles, in those you have to ensure maximal specificity in order to increase the specific template efficiently avoiding producing at the same time unspecific material. That's why Mark suggests the step-down approach, which in case you obtain a dirty PCR at lower annealing temp, but no product at all or too little if you raise it, can work very well.

Keep in mind that independently from the global Tm of your primers, the way you design the 3' end of your primers (essentially the dG given by the annealing of the last 7-8 3' nts) will strongly affect the annealing temperature for your PCR reaction and the specificity of your primers. So, the most care in the design should be placed on the 3' side of the primers, avoiding sequences that can give rise to primer-dimers (GC, or GG and CC on FW and RV primers, etc.), avoiding too many GC pairing all together (try to have no more than two adjacent ones), etc...

Good luck!

Many good comments so far (Pietro's especially thoughtful).  I would try to use T_A1 for the entire PCR and see how it goes.  If you have problems with non-specific products, you can try the type of step-up protocol that you describe, but I would use fewer cycles for the first phase--just 3 to 5 cycles at the low T_A and then shift up to the higher one.  If you have access to a gradient machine,  I would also just try using a range of higher T_A's for the entire PCR-- as many primers will work well at far above their "calculated" T_A.

Use touch down PCR, I think it is the name. Anyway, I typically set first 10 cycle as touch down, with -0.5 C in each cycle, and start the annealing temp 2-3 C above the Tm of the specific portion of primer. The set another 25 cycles with annealing temp as the Tm (without your tail). It works for me, even for long fragments larger than 10k.  

Thanks everyone! I'm going to start with the estimated T_A (for Phusion, modified Breslauer thermodynamics) and use a touchdown protocol if I have difficulty with non-specific products.