The Polymerase Step Buys Length, Not Primer Binding

The extension step in PCR is the polymerase interval that synthesizes the new strand. It is not primer annealing, and it is not a calculated melting temperature. Prefer setting extension time from amplicon length and the polymerase you actually use. Reject a copied thirty-second step on every product.

NEB’s Taq optimization note treats extension time as a function of product length for that enzyme class. Tm versus annealing still owns the calculated number versus the cycler step. PCR versus qPCR versus RT-qPCR still owns the three assay names. PCR controls still own the lane map. This page only names the polymerase step. It does not choose primers. It does not invent a Zetta cycling rate.

Extension Is the Polymerase Interval

The extension step in PCR is the portion of the cycle when the polymerase adds nucleotides from the annealed primers along the template. The entity is a timed synthesis interval, not a binding event. If the interval is too short for the length and the enzyme, the product is incomplete even when primers bound well. If the interval is far longer than the chemistry needs, the cycle still “works” in a narrow sense and wastes time; that waste is not proof the step was understood.

The search asks what extension in PCR is. The short answer is synthesis time. Annealing is the step that lets primers bind. Denaturation is the step that opens the template. Extension is the step that buys length. Those three jobs are easy to collapse on a copied program that lists one temperature and hopes. Collapse is how a 3 kb clone inherits a 30-second extension written for a colony screen.

Extension Is Not Annealing and Not Tm

Keep the three numbers in their own jobs.

Prefer reading Tm when you are choosing or checking primers. Prefer reading annealing when you are setting the binding temperature. Prefer reading extension when you are asking whether the enzyme had time to finish the amplicon. Reject a protocol that uses one number for all three. The Tm sibling already exists so this page does not rematch that math.

Time Follows Length and Speed

Vendor tables for Taq-class polymerases commonly scale extension with amplicon length, often near one minute per kilobase as a starting heuristic. That figure is a manufacturer guideline for a named enzyme class. It is not a universal law, not a Zetta rate, and not a promise that every polymerase behaves the same. Fast enzymes are sold with shorter recommended times. High-fidelity mixes publish their own tables. Long amplicons fail first when someone copies a colony-screen program onto a cloning product.

A worked fail is a 2.8 kb insert amplified with a 30-second extension copied from a 400 bp screen. Primers can be perfect. Annealing can be in range. The gel is empty or shows a smear of unfinished products. The missing object is time, not “PCR.” Another fail is treating qPCR’s short cycling block as proof that endpoint cloning PCR should use the same extension. Assay taxonomy stays on the three-name sibling. The polymerase still has to finish the fragment you asked for.

Do not invent an optimal time for a named brand here. Open the polymerase’s current note. Write the length. Set the interval. If the product is GC-rich or the enzyme is processive, the note—not this page—owns the adjustment.

Record the Program After the Step Is Named

Name the step before anyone stores a cycler screenshot. If the lab only needs a binding temperature, stop. An extension page is not the subject. If the lab must reconstruct why a long product failed, name a record surface after denature, anneal, and extend are written as separate lines. ZettaNote is one such surface: official product language includes structured experiment documentation. A dated entry can hold the cycler program. That is a place the three steps may be written. It is not a thermal cycler. It is not a polymerase. Golden Gate is not a Zetta feature and does not belong on this definition.

If the group still pastes one 30-second block onto every amplicon, you do not have an extension rule. You have a habit. Write the length. Write the enzyme. Then set the interval.

Colony screens and cloning amplicons fail for different reasons when extension is copied. A short screen can finish in a brief interval and still look clean. The same interval on a multi-kilobase insert produces no band or a ladder of unfinished products that a tired reader calls primer-dimer. Primer-dimer is a different sibling. Do not diagnose a short extension as a dimer problem until the polymerase had time to finish the intended length. Conversely, an extra-long extension does not repair a primer that never bound. Set annealing on the Tm page. Set extension here.

Final extension, sometimes added as a last extra interval after cycling, is still the same job: leftover synthesis and, for some polymerases, a chance to fill ends. It is not a new definition. If the program includes it, write it as an extra polymerase interval with a reason. If the reason is “the kit said so,” cite the kit. Do not treat a final extension as proof that the per-cycle interval was understood. The per-cycle interval is what scaled with length. The final extra is a footnote unless the manufacturer says otherwise for that mix.

Multiplex and nested programs still use the same definition. Extra primer pairs do not invent a new polymerase interval. They can change which product must finish first. Set time for the longest intended amplicon unless the manufacturer says the mix is limited by a shorter product. Nested second rounds often inherit the first-round time and then fail on a longer inner product, or the reverse. Write each round’s length. Do not treat “nested” as a license to copy one block. Controls stay on the control sibling. This page only keeps the polymerase job named.

Frequently Asked Questions

Is PCR extension the same step as annealing?

No. Annealing binds primers. Extension is the polymerase interval. A correct annealing temperature cannot finish a long product if extension is too short.

Can one copied extension time serve every amplicon?

No. Time follows length and the polymerase you actually use. Vendor tables are heuristics, not a single lab-wide number.