Primer Melting Temperature
Melting temperature by nearest-neighbour thermodynamics, with the two rough estimates alongside.
20 nt · 60.0% GC · 250 nM primer, 50 mM Na⁺
CAATGCCTCCAGGTCCAGAC
A primer’s melting temperature is where half of it is bound to its template. It depends on sequence, salt and concentration — not just length. A 20-base primer at 60% GC in standard PCR buffer melts near 57 °C, and the annealing step is usually set a few degrees below that.
Three ways to estimate it, and why they disagree
The Wallace rule, 2(A+T) + 4(G+C), is arithmetic you can do in your head. It ignores salt, concentration and the order of the bases entirely, and it was only ever intended for oligos shorter than about fourteen bases. Above that it runs several degrees high.
The salt-adjusted GC formula does better over the 14–70 base range because it accounts for length and ionic strength. But it still treats the sequence as a bag of bases, so it cannot tell GCGCGC from GGGCCC even though those two stack very differently.
Nearest-neighbour thermodynamics is what modern primer design uses. It assigns an enthalpy and entropy to each adjacent base pair step — there are only ten distinct ones — sums them along the sequence, adds an initiation term for each end, then corrects for salt and strand concentration. The tool reports all three so the gap between them is visible; where they disagree, the nearest-neighbour value is the one to trust.
Tm is not the annealing temperature
Tm describes the primer-template duplex; the annealing temperature is a setting on a thermocycler. Convention is to anneal a few degrees below the lower of the two primer Tm values, and many protocols use Ta = Tm − 5 °C as a starting point before optimising. What matters more than the absolute number is that the two primers of a pair are matched — a difference of more than about 5 °C means one anneals well while the other does not.
ΔH° and ΔS° summed over the nearest-neighbour steps; R = 1.987 cal/(mol·K); C_T the total strand concentration
- 1 Break the sequence into overlapping pairs. GTCTGG gives GT, TC, CT, TG, GG — a 20-base primer has 19 such steps.
- 2 Look up ΔH and ΔS for each step. There are only ten distinct values, because a step and its reverse complement are the same stack seen from the other strand.
- 3 Add an initiation term for each end. A terminal G or C contributes differently from a terminal A or T, since an A·T end frays more easily.
- 4 Correct the entropy for salt. Cations shield the backbone charges, so more salt stabilises the duplex and raises Tm.
- 5 Solve for the temperature. Substituting into the equation gives 56.8 °C for a 20-base, 60% GC primer at 250 nM and 50 mM Na⁺.
What each estimate is for
All three appear in the tool; they diverge most for long or unusually composed primers.
| Method | Useful range | Accounts for |
|---|---|---|
| Wallace rule | Under ~14 bases | Base counts only |
| Salt-adjusted GC | ~14–70 bases | Length, GC%, salt |
| Nearest neighbour | Any length above ~8 bases | Sequence order, salt, concentration |
What a usable PCR primer normally looks like
Guidelines rather than rules — a primer outside them can still work, and one inside them can still fail.
| Property | Typical target |
|---|---|
| Length | 18–24 bases |
| Tm | 55–65 °C |
| GC content | 40–60% |
| Tm difference within a pair | Under 5 °C |
| 3′ end | G or C helps anchor the extension |
| Avoid | Runs of four or more identical bases, and self-complementarity |
The conditions change the answer
Tm is not a property of the sequence alone, which is why a value quoted without conditions is incomplete. Raising monovalent salt from 50 mM to 1 M lifts the example primer from about 57 °C to 72 °C, because the cations screen the repulsion between the two negatively charged backbones. Dropping the primer concentration lowers Tm, since a duplex is less likely to find its partner when there is less of it about.
Two limits are worth stating. The model assumes a perfectly matched duplex, so a primer with a deliberate mismatch — adding a restriction site, say — melts lower than the calculation suggests, and the usual approach is to compute Tm for the annealing portion only. And magnesium, which PCR buffers contain and which stabilises duplexes strongly, is not included in a simple monovalent correction; tools that model it apply a separate divalent term.