DNA melting curve
Predicted melting curve and Tm of a short DNA duplex, from nearest-neighbour thermodynamics.
Predicted melting curve and Tm of a short DNA duplex, from nearest-neighbour thermodynamics.
The tool treats melting as a two-state equilibrium: a strand is either in a complete duplex or fully single-stranded. ΔH and ΔS of duplex formation are summed from the unified nearest-neighbour parameters of SantaLucia (1998), one term per stacked base-pair step plus initiation terms for the two ends. From them the equilibrium constant at any temperature follows, and from that the fraction of strands in the duplex, θ.
R is 1.987 cal/(K·mol) and T is in kelvin. Setting θ = 0.5 in the second line gives the Tm formula in the last line, so the curve and the reported Tm always agree. For a self-complementary sequence, which pairs with itself, θ / (1 − θ)2 = 2 K Ct, a symmetry term of −1.4 cal/(K·mol) is added to ΔS, and Tm = ΔH / (ΔS + R ln Ct).
The parameters describe 1 M NaCl. For lower salt the tool uses the correction of Owczarzy (2004), 1/Tm(Na) = 1/Tm(1 M) + (4.29 fGC − 3.95) × 10−5 ln[Na⁺] + 9.40 × 10−6 ln2[Na⁺], and applies it as a change in ΔS with ΔH held constant. The Tm therefore equals the one from the Tm calculatorat the same Na⁺ and strand concentration without Mg²⁺. Use that page for Mg²⁺ and dNTP corrections, several primers at once and annealing temperatures.
The example is the M13 forward primer GTAAAACGACGGCCAGT (17 nt) with its complement, 250 nM total strands and 50 mM Na⁺. The nearest-neighbour sums are ΔH = −133.6 kcal/mol and ΔS = −360.7 cal/(K·mol) at 1 M NaCl, which becomes −378.7 cal/(K·mol) at 50 mM Na⁺.
The curve passes 10 % melted at 45.5 °C and 90 % melted at 55.8 °C, so the transition is about 10 °C wide.
For duplexes up to about 40 bp that have no stable hairpins or alternative structures. Longer DNA melts in stages, AT-rich parts first, and needs a different model; the tool warns above 60 nt. Only A, C, G and T are accepted, because the parameters cover Watson-Crick pairs only. If several sequences are pasted, the chart shows the first and the table lists all.
A plot of how much of a DNA duplex has separated into single strands as the temperature rises. It is S-shaped: the duplex is intact at low temperature, melts over a range of roughly 10 to 20 °C for a short oligo, and is fully single-stranded above that. The midpoint, where half of the duplexes are melted, is the melting temperature Tm.
Tm is the temperature at which the fraction of single strands is exactly 50 %. For a two-state transition of two different strands at equal concentration this gives Tm = ΔH / (ΔS + R ln(Ct/4)), and the curve drawn here crosses 50 % at exactly that temperature.
Forming a duplex needs two strands to meet, so the equilibrium shifts towards the duplex when more strands are present. A tenfold increase in strand concentration raises the Tm of a typical 20-mer by about 3 to 4 °C. Salt also raises it, because cations screen the repulsion between the two phosphate backbones.
It is the same physics, but a qPCR instrument plots fluorescence, or its negative derivative, for an amplicon of 70 to 200 bp with a dye bound. Long products melt in several domains and are not two-state, so use this tool for oligos, primers and probes up to roughly 40 nt rather than for amplicons.
It is set by ΔH: the larger the enthalpy, the steeper the curve. Longer duplexes have a larger ΔH and melt over a narrower range. A measured curve much broader than the prediction suggests a mixture of products, mismatches or a transition that is not two-state.