How to Calculate Primer Tm

The melting temperature (Tm) of a primer is the temperature at which half of it is bound to its complement and half is not. It is the number every PCR annealing temperature starts from, and you can calculate it two ways: the quick Wallace rule, which counts bases, and the SantaLucia 1998 nearest-neighbour method, which sums the stacking energy of every base-pair step and corrects for salt and primer concentration. This guide derives both, works a real 18-mer through the nearest-neighbour calculation step by step, explains the salt and concentration terms, and shows how to turn Tm into the annealing temperature and gradient you actually set on the cycler. The Primer Tm calculator runs both methods and shows its ΔH and ΔS so you can check the arithmetic.

What is primer Tm?

Tm is a property of a duplex, not of a single strand: it describes the temperature at which the primer and its complementary target are half-annealed at equilibrium. Above Tm the duplex falls apart; below it, it forms. Tm depends on four things — the primer’s length, its GC content (G·C pairs stack more stably than A·T pairs), the salt concentration (cations shield the backbone’s negative charge and stabilise the helix), and the primer concentration (more strands means the duplex holds to a higher temperature). Any method that ignores one of these is an approximation.

Tm is not the annealing temperature. The annealing temperature (Ta) is what the cycler runs, and it sits a few degrees below Tm so the primer can find its target within the seconds the step allows.

The Wallace rule

The Wallace rule is the back-of-the-envelope shortcut:

Tm = 2·(A + T) + 4·(G + C)

Count the A and T bases, multiply by 2; count the G and C bases, multiply by 4; add. It assumes every A/T pair contributes about 2 °C and every G/C pair about 4 °C, and it ignores salt, concentration and base order entirely.

Worked example — the 8-mer ACGTTGCA (4 A/T, 4 G/C):

• A/T: 4 × 2 = 8
• G/C: 4 × 4 = 16
• Tm = 8 + 16 = 24 °C

That is far below any usable annealing temperature, which tells you immediately this oligo is too short to work as a PCR primer. The Wallace rule is best for primers shorter than about 14 nt; beyond that, and especially for GC-rich sequences, it drifts and the nearest-neighbour method is markedly more reliable.

The nearest-neighbour method

The nearest-neighbour (NN) model treats the duplex as a stack of base-pair steps rather than independent pairs, because a G next to a C stabilises the helix differently depending on which bases flank it. Each of the ten unique dinucleotide steps has a published enthalpy (ΔH) and entropy (ΔS), measured by SantaLucia and colleagues and unified in the 1998 PNAS paper. The Tm comes from thermodynamics:

Tm = (ΔH × 1000) / (ΔS + R·ln(Ct/4)) − 273.15

where ΔH is in kcal/mol, ΔS in cal/(mol·K), R = 1.9872 cal/(mol·K), and Ct is the total primer strand concentration in molar. The result is in °C.

Step 1 — sum the stacking terms

Walk the sequence 5′→3′ and for each adjacent base pair look up its ΔH and ΔS. Because a step and its reverse complement stack identically, the sixteen possible dinucleotides collapse to ten unique values (AC and GT, for example, share one entry). Add an initiation term for each end of the duplex — a terminal G·C or A·T base pair contributes its own small ΔH/ΔS. If the primer is self-complementary (palindromic), add the −1.4 entropy-unit symmetry correction and use Ct instead of Ct/4.

Step 2 — apply the salt correction

The published parameters are for ~1 M NaCl, so correct the entropy for your actual monovalent salt:

ΔS = ΔS(NN + init) + 0.368·(N − 1)·ln[Na⁺]

with N the primer length in bases and [Na⁺] in molar. Higher salt makes the ln term less negative and raises Tm.

Step 3 — add the concentration term and solve

Compute R·ln(Ct/4), add it to the salt-corrected ΔS, and divide ΔH × 1000 by the sum. Subtract 273.15 to convert kelvin to Celsius.

Worked example: a real 18-mer

Take the primer 5′-GCGCATGCGCATACGTAG-3′ (18 nt, 11 G/C, 61.1% GC), at 250 nM primer and 50 mM Na⁺.

1. Stacking sum. Adding the 17 nearest-neighbour steps plus the two terminal initiation terms gives ΔH = −149.7 kcal/mol and ΔS(NN + init) = −402.1 cal/(mol·K).
2. Salt correction. 0.368 × (18 − 1) × ln(0.05) = 0.368 × 17 × (−2.996) = −18.7. So ΔS = −402.1 − 18.7 = −420.8 cal/(mol·K).
3. Concentration term. Ct/4 = 250 nM / 4 = 62.5 nM = 6.25 × 10⁻⁸ M. R·ln(6.25 × 10⁻⁸) = 1.9872 × (−16.59) = −33.0.
4. Solve. Denominator = −420.8 − 33.0 = −453.8. Tm = (−149.7 × 1000) / (−453.8) − 273.15 = 329.9 − 273.15 = 56.7 °C.

For comparison, the Wallace rule on the same primer gives 2·7 + 4·11 = 58 °C — within about 1.3 °C of the nearest-neighbour value, which is typical agreement for a clean, mid-length primer. When the two methods disagree by much more than that, the sequence is one where the shortcut is lying to you, and you should trust the nearest-neighbour number.

From Tm to annealing temperature and gradient

The recommended starting annealing temperature is Ta = Tm − 5 °C, so for the example above start at about 51.7 °C. Because every prediction carries error — different tools use different parameter sets, salt models and divalent corrections, and can legitimately differ by 1–3 °C — the honest workflow is a gradient PCR across roughly Tm − 10 to Tm + 5 °C (about 46.7 to 61.7 °C here). Let the gel, not the prediction, pick your final number.

Salt and magnesium both move Tm

Doubling monovalent salt raises Tm measurably. The 18-mer CGTACGTCGATGCATGCA predicts 55.2 °C at 50 mM Na⁺ but 58.4 °C at 100 mM Na⁺ — about 3 °C for a doubling. Always quote the salt your Tm was computed at.

Magnesium matters even more, because divalent cations stabilise the duplex strongly, but Mg²⁺ corrections are model-dependent and most of the effect in a real reaction is bound up in the polymerase buffer. Rather than apply an unverified divalent formula, compute your free [Mg²⁺] = [Mg²⁺] − [dNTP] (dNTPs chelate magnesium) and treat it as guidance: a high free Mg²⁺ means the real Tm sits above what a Na⁺-only model predicts. The Primer Tm calculator reports free [Mg²⁺] without folding it into Tm, and states plainly that it uses the monovalent model.

Common mistakes

• Setting Ta equal to Tm. Anneal a few degrees below Tm or the primer never binds and you get no product.
• Comparing across salt conditions. A Tm computed at 50 mM Na⁺ is not the same number as one at the 100 mM your buffer uses.
• Trusting Wallace for a long or GC-rich primer. The nearest-neighbour method is markedly more accurate there.
• Forgetting Mg²⁺. A Na⁺-only prediction sits below the real melting point in a magnesium-rich master mix.
• Using one predicted Tm for a multiplex without checking that every primer shares a workable annealing temperature.

Where does this fit in the wider workflow?

Tm is driven mostly by composition, so check the sequence first with the GC content calculator — the how to calculate GC content guide explains the counting rules — then predict melting behaviour with the Primer Tm calculator. Both live in the PCR & qPCR calculators hub. When you go to set up the reaction itself, the primer and dNTP concentrations are ordinary dilution maths, covered by the laboratory calculators.