Tm calculator
Nearest-neighbour Tm with salt correction. One primer per line, or FASTA.
Nearest-neighbour Tm with salt correction. One primer per line, or FASTA.
The melting temperature, Tm, is the temperature at which half of the primer molecules are paired to their complementary strand and half are single-stranded. It sets the PCR annealing temperature, decides whether two primers can be used together, and predicts whether a probe or an oligo will stay bound. Tm rises with length, with GC content, with salt, and with the concentration of the strands, which is why a calculator needs more than the sequence.
The stability of a duplex is the sum of the stabilities of its ten possible base-pair neighbours, each with a measured enthalpy ΔH and entropy ΔS, plus small terms for the ends of the helix. This page uses the unified parameters of SantaLucia (1998), the set behind the NEB, IDT and Primer3 calculators. Summing the neighbours gives the melting temperature at 1 M Na⁺:
where C is the oligo concentration and R the gas constant. For a self-complementary primer the divisor is 1 instead of 4 and a symmetry term is added. The result is then corrected to the actual salt.
Cations shield the negative phosphate backbone, so more salt means a more stable duplex and a higher Tm. The correction here is the one derived by Owczarzy and colleagues, first for monovalent ions (2004) and then for Mg²⁺ with or without monovalent ions (2008). It adjusts 1/Tm using the GC fraction, the ion concentrations and the length of the duplex. Mg²⁺ counts for much more than Na⁺, and because dNTPs chelate Mg²⁺, the free Mg²⁺ is estimated from the Mg²⁺ and dNTP concentrations before the correction is applied. Enter the composition of your actual reaction buffer to get a Tm for the reaction rather than for a test tube of saline.
| Rule | Formula | Use |
|---|---|---|
| Wallace | Tm = 2(A + T) + 4(G + C) | Short probes under 14 nt in high salt. Shown for every primer as a sanity check. |
| GC percent | Tm = 64.9 + 41 × (G + C − 16.4) ÷ N | Oligos of 14 nt and longer; ignores base order and salt. |
| Nearest neighbour | ΔH ÷ (ΔS + R ln C) with salt correction | The standard for primer design. Accurate to a degree or two. |
The simple rules were derived for particular conditions and give the same answer for any permutation of the same bases. The nearest-neighbour model does not, which is the point: the order of the bases changes how the duplex stacks.
For a standard Taq reaction, anneal about 3 to 5 °C below the lower Tm of the pair, calculated with the buffer's Mg²⁺. Keep the two primers within about 5 °C of each other; the pair difference is reported above when you paste exactly two. High-fidelity enzymes come with their own rules, so check the supplier's guidance, and use a gradient cycler when a reaction is marginal. The PCR annealing and extension calculator turns Tm into cycling conditions, PCR product size checks where primers bind, and the primer designer proposes primers to begin with.
Nearest-neighbour parameters exist only for A, C, G and T (U is read as T). A primer that contains an ambiguity code, a modification or a non-base character is listed without a Tm rather than given a wrong one. Primers shorter than 8 nt fall outside the model's range and are flagged; so are sequences longer than 60 nt, which are not oligos but templates. For the Tm of a long duplex, the melting curve plotter andGC content calculator are the right tools.
Because Tm is not a single number: it depends on the thermodynamic parameters, the salt correction and the oligo concentration each calculator assumes. Two nearest-neighbour calculators fed the same sequence, salt and concentration agree within about a degree. The big gaps come from calculators that use the Wallace or GC-percent rules, or that quietly assume different Na+, Mg2+ or primer concentrations. This page shows its assumptions and lets you change them.
A common starting point is 3 to 5 °C below the lower Tm of the two primers, calculated with the Mg2+ concentration of your PCR buffer. For high-fidelity polymerases such as Q5 the manufacturer recommends Tm + 3 °C instead, because their own calculator uses a different model, so follow the polymerase supplier when in doubt, and run a gradient if the reaction is marginal.
Nearest neighbour. It accounts for the order of the bases, the salt and the strand concentration, and it is accurate to a couple of degrees for typical primers. The Wallace rule, 2 °C per A or T and 4 °C per G or C, was designed for short probes in high salt and is only a rough guide above 14 bases.
Yes, a lot. Divalent cations stabilise the duplex far more than Na+. A 20-mer that melts at 55 °C in 50 mM Na+ typically melts several degrees higher in a PCR buffer with 1.5 to 2 mM Mg2+. dNTPs bind Mg2+ and reduce its free concentration, which the calculator takes into account.
Not with this method. Nearest-neighbour parameters exist only for the four standard bases, so a primer containing N, R, Y or other ambiguity codes, or modified bases, is flagged rather than given a misleading number. For a degenerate primer, calculate the Tm of its most AT-rich and most GC-rich versions to get the range.