Calculate the melting temperature of the 20-mer primer TACGGCATCCAGTTGAGTCA under the default conditions of four common oligo calculators and the answers spread across almost 10 °C. None of them is broken. Each assumes a different reaction: different oligo concentration, different salt, with or without Mg²⁺, and in one case an older set of thermodynamic parameters. Most of the disagreement between oligo analyzers comes from those assumptions, not from the quality of the underlying science.
That changes the question worth asking. The most accurate oligo analyzer is the one that uses current nearest-neighbor thermodynamics and lets you enter your own reaction conditions. On that measure, IDT OligoAnalyzer, Primer3 and the ProteinIQ oligo analyzer are built on the same family of published models, while some quick calculators trade those corrections for simplicity.
Oligo calculators at a glance
| Calculator | Tm parameters | Salt and Mg²⁺ | Hairpins and dimers | Default conditions | Access |
|---|---|---|---|---|---|
| IDT OligoAnalyzer | Allawi and SantaLucia 1997 nearest neighbor | Owczarzy 2004 for Na⁺, Owczarzy 2008 for Mg²⁺, dNTPs modeled | mfold hairpins; self- and heterodimers with ΔG | 0.25 µM oligo, 50 mM Na⁺, no Mg²⁺ or dNTPs | Free with an IDT account |
| Primer3 (Primer3web, Primer3Plus) | SantaLucia 1998 by default; Breslauer 1986 optional | SantaLucia 1998, Schildkraut-Lifson or Owczarzy; Mg²⁺ and dNTPs modeled | Thermodynamic alignment, reported as structure Tm | 50 nM oligo, 50 mM monovalent, 1.5 mM Mg²⁺, 0.6 mM dNTPs | Free and open source, no account |
| ProteinIQ oligo analyzer | Primer3 2.6.1, SantaLucia 1998 by default; Breslauer 1986 optional | Same three Primer3 options; Mg²⁺ and dNTPs modeled | Primer3 thermodynamic alignment with Tm, ΔG, ΔH, ΔS and structure | Primer3 defaults, all editable | ProteinIQ account; 5 credits per run |
| NEB Tm Calculator | SantaLucia 1998 | Owczarzy 2004, or Schildkraut for Phusion; each buffer is one salt value | None | Set by the chosen polymerase and buffer | Free, no account |
| Thermo Fisher Multiple Primer Analyzer | Modified nearest neighbor based on Breslauer 1986 | Monovalent salt only | Complementary stretches, no hairpins | 0.5 µM primer, 50 mM salt | Free, no account; needs at least two primers |
| Promega Biomath Tm calculator | SantaLucia 1998 and Allawi and SantaLucia 1997 | Monovalent salt; Mg²⁺ optional, no dNTPs | None | 200 nM primer, 50 mM Na⁺, no Mg²⁺ | Free web page |
Two familiar names are missing from the table because they are no longer available. Sigma-Aldrich describes its OligoEvaluator as temporarily unavailable while it builds a new version.[17] OligoCalc, the Northwestern calculator described by Kibbe in 2007, has returned errors since April 2025; it used Sugimoto's nearest-neighbor values and made no Mg²⁺ correction.[16]
Why the same primer gets different Tm values
We calculated the Tm of the 20-mer above with Primer3 2.6.1 under the default conditions each calculator documents, using the Primer3 method closest to each calculator's own. These are not the vendors' outputs, which use their own code and in IDT's case a slightly different parameter table, but they isolate how much the assumed reaction alone moves the answer.
| Conditions | Oligo | Monovalent | Mg²⁺ / dNTPs | Tm |
|---|---|---|---|---|
| Promega default | 200 nM | 50 mM | 0 / 0 | 54.8 °C |
| IDT OligoAnalyzer default (Owczarzy correction) | 250 nM | 50 mM | 0 / 0 | 54.9 °C |
| Primer3 and ProteinIQ default | 50 nM | 50 mM | 1.5 mM / 0.6 mM | 58.7 °C |
| qPCR-style buffer (Owczarzy correction) | 250 nM | 50 mM | 3 mM / 0.8 mM | 61.7 °C |
| Thermo Fisher default (Breslauer 1986) | 500 nM | 50 mM | 0 / 0 | 64.3 °C |
Mg²⁺ is the biggest lever. A Taq or qPCR buffer with 1.5 to 3 mM Mg²⁺ stabilizes a duplex far more than the same molarity of Na⁺, so a calculator that assumes no Mg²⁺ underestimates the Tm in a real PCR by several degrees. The Breslauer row shows the other source of spread: SantaLucia's 1998 analysis found that Breslauer's 1986 parameters missed the measured Tm of his 108-duplex data set by 6.0 °C on average.[1][2]
The practical rule follows directly. Before comparing a Tm from one tool with a Tm from another, or with a value printed on a supplier's specification sheet, check that the oligo, salt, Mg²⁺ and dNTP concentrations match. If they do not, the comparison says nothing about which calculator is better.
Nearest-neighbor parameters
Every modern oligo analyzer predicts Tm from nearest-neighbor thermodynamics: the stability of a duplex is the sum of contributions from each pair of adjacent base pairs, plus initiation terms. Breslauer and colleagues published the first complete DNA table in 1986 from 19 oligomers and 9 polymers.[2] In 1998 SantaLucia reconciled the parameters from seven laboratories into a unified set derived from 108 oligonucleotide duplexes.[1]
The unified parameters are the default in Primer3, NEB, Promega and ProteinIQ, and the closely related Allawi and SantaLucia (1997) table is what IDT uses for DNA.[18][9] With his salt correction, SantaLucia reported that they predicted the Tm of 26 oligonucleotide duplexes in 0.01 to 0.3 M NaCl with an average deviation of 2.2 °C.[1] That is the realistic accuracy floor for any of these calculators on a well-behaved primer, and it is why differences under about 2 °C between two modern tools are not meaningful.
Basic formulas based only on length and GC content, such as the Wallace rule, ignore sequence order entirely. They are fine for a rough number but cannot tell two primers with the same composition apart. The GC content calculator reports that kind of quick estimate; a nearest-neighbor tool is the one to use for PCR conditions.
Salt, Mg²⁺ and dNTP corrections
Nearest-neighbor parameters are measured in 1 M NaCl, so every calculator has to correct them to the reaction. The simplest correction, from Schildkraut and Lifson, adds a term proportional to the logarithm of Na⁺. SantaLucia's 1998 correction adjusts the entropy instead.[1] Owczarzy and colleagues measured 92 duplexes between 69 mM and 1.02 M Na⁺ and derived a quadratic correction that predicted an independent validation set with an average error of 1.6 °C.[3]
Mg²⁺ needs its own treatment. Owczarzy et al. (2008) showed that once the ratio of the square root of Mg²⁺ to monovalent cations exceeds 0.22, magnesium dominates duplex stability, and that typical PCR and sequencing buffers with 1.5 to 5 mM Mg²⁺ and 20 to 100 mM monovalent cations fall in that range.[4] IDT switches between its monovalent and Mg²⁺ corrections at exactly that ratio, after correcting free Mg²⁺ for binding to dNTPs.[9]
Primer3 handles Mg²⁺ in two ways. With the SantaLucia or Schildkraut-Lifson correction it converts Mg²⁺ to an equivalent Na⁺ concentration with the von Ahsen formula, 120 times the square root of Mg²⁺ minus dNTPs (in mM), and ignores Mg²⁺ entirely when dNTPs exceed it.[7] Von Ahsen and colleagues reported a standard error of 1.76 °C for this approach on 217 matched duplexes measured under PCR conditions.[5] With the Owczarzy option, Primer3 applies the 2004 monovalent correction or the 2008 Mg²⁺ correction depending on which ion dominates.[7]
Calculators without a Mg²⁺ input are not wrong, but they answer a different question. Thermo Fisher's Multiple Primer Analyzer takes only primer and salt concentrations.[13] NEB reduces each polymerase buffer to a single salt value and adds polymerase-specific rules for the annealing temperature, which makes it a good choice when you are using an NEB enzyme and less useful as a general Tm reference.[14]
Hairpins and primer dimers
Hairpin and dimer checks differ more between tools than Tm does. IDT finds hairpins with Zuker's mfold algorithm and lists self- and heterodimers with their ΔG, and its guidance treats a ΔG of −9 kcal/mol or more negative as likely to be a problem.[10][11] Primer3 and ProteinIQ use Primer3's thermodynamic alignment, which scores the most stable intramolecular or intermolecular structure with nearest-neighbor parameters for matches, mismatches, dangling ends, bulges and loops.[7][6]
Primer3 reports each structure as a melting temperature as well as a ΔG, and its primer design defaults reject hairpins and dimers that melt above 47 °C.[7] A structure Tm is easy to compare with an annealing temperature: a dimer that melts at 20 °C will not survive a 60 °C annealing step, while one at 50 °C may. The 3′ end matters most, because a polymerase can extend a duplex anchored there, so ProteinIQ also reports the 3′ end stability of each oligo against each partner.
Thermo Fisher's analyzer works differently. It looks for exact complementary stretches between primers at a user-set sensitivity, has no hairpin check, and states that its dimer result is a preliminary guide and "not conclusive data".[13] NEB and Promega calculate Tm only.[14][15]
For the primer pair used above, Primer3's alignment gives self-dimer ΔG values of −2.7 and −4.5 kcal/mol and a heterodimer ΔG of −2.6 kcal/mol at 37 °C, all with structure Tm values below 0 °C. By IDT's −9 kcal/mol rule and by Primer3's 47 °C limit, the pair is clean.
Outputs beyond Tm
| Calculator | Physical properties | Chemistry beyond plain DNA | Batch and automation |
|---|---|---|---|
| IDT OligoAnalyzer | MW, nearest-neighbor extinction coefficient, nmol/OD and µg/OD | RNA, LNA, 2′-O-methyl, phosphorothioates, named modifications, degenerate bases, single-mismatch Tm | SciTools Plus API for organizations |
| Primer3 | None | Degenerate bases can be read as N | Command-line program and library |
| ProteinIQ oligo analyzer | MW, additive extinction coefficient, nmol/OD260 and µg/OD260 | Unmodified DNA only | Up to 96 oligos per run, every pair of up to 24 screened, CSV and JSON export, API on paid plans |
| NEB Tm Calculator | None | Up to 3 degenerate bases, reported as the minimum Tm | Batch mode for primer pairs |
| Thermo Fisher Multiple Primer Analyzer | MW, extinction coefficient, nmol/OD260 and µg/OD260 | Degenerate bases accepted | Paste several primers at once |
| Promega Biomath | None | None | None documented |
IDT's analyzer covers the most chemistry, which is no surprise from a company that synthesizes modified oligos. If a primer carries a dye, a quencher, LNA bases or phosphorothioate bonds, use it or your own supplier's calculator, because the modification changes mass and absorbance and, for LNA, the Tm itself.[9] For modifications it has no thermodynamic parameters for, IDT reports the Tm of the unmodified sequence.
ProteinIQ calculates the extinction coefficient by adding per-base values, while IDT uses a nearest-neighbor model that accounts for stacking between adjacent bases.[9] The additive value is the right order of magnitude for quantifying a primer stock, but for exact concentrations of a modified or structured oligo, the supplier's value or a measured one is better. Molecular weight and extinction coefficient for single sequences are also available in the molecular weight and extinction coefficient calculators.
Which oligo analyzer to use
For a single unmodified primer and a known buffer, any calculator built on the SantaLucia parameters with a Mg²⁺ correction gives a Tm within the 2 °C or so that the underlying model supports. The choice then comes down to what else you need.
Use IDT OligoAnalyzer when the oligo is modified, contains RNA or LNA, or you need a Tm against a single-mismatch target, and you are willing to sign in. Use the NEB calculator when you want an annealing temperature for a specific NEB polymerase. Use Primer3 directly, or Primer3 on ProteinIQ, when you are designing primers from a template rather than checking ones you already have.
Use the ProteinIQ oligo analyzer when you want Primer3's thermodynamics without installing it, need to screen a primer panel for every pairwise dimer at once, or want each analysis saved with its conditions so a colleague can rerun it. It does not check specificity against a genome; NCBI Primer-BLAST does that.
Frequently asked questions
What is the most accurate oligo analyzer?
There is no single most accurate tool, because accuracy depends mostly on matching the calculation to your reaction. Calculators that use the SantaLucia 1998 or Allawi and SantaLucia 1997 nearest-neighbor parameters with Mg²⁺ and dNTP corrections, such as IDT OligoAnalyzer, Primer3 and the ProteinIQ oligo analyzer, reach the roughly 2 °C accuracy those models were validated to.[1][5] IDT reports that its own algorithm was validated on more than 100 independent sequences of 8 to 60 base pairs.[8]
Why does my Tm differ between IDT and another calculator?
Usually because the defaults differ. IDT assumes 0.25 µM oligo, 50 mM Na⁺ and no Mg²⁺, while Primer3 assumes 50 nM oligo, 50 mM monovalent cations, 1.5 mM Mg²⁺ and 0.6 mM dNTPs.[9][7] For the 20-mer in this guide, that alone moves the Tm from 54.9 to 58.7 °C.
Is IDT OligoAnalyzer free?
IDT describes it as free, but it now requires signing in with an IDT account.[12] Primer3web, Primer3Plus, the NEB and Promega calculators and the Thermo Fisher analyzer work without an account.
What ΔG is a problem for primer dimers?
IDT's guidance flags dimers and hairpins with a ΔG of −9 kcal/mol or more negative.[11] Primer3 expresses the same risk as a structure melting temperature and by default rejects structures that melt above 47 °C.[7] Interactions at the 3′ end deserve the most attention because a polymerase can extend them.
Can I check RNA or modified oligos?
IDT OligoAnalyzer handles RNA, LNA, 2′-O-methyl bases, phosphorothioates and many named modifications.[10] Primer3 and the ProteinIQ oligo analyzer are limited to unmodified DNA; for RNA folding, RNAfold predicts secondary structure from RNA-specific parameters.
