Review a Cloning Construct Before Anyone Orders Oligos

A design review workflow for cloning constructs is a repeatable pass/fail check on the expected assembled map before anyone orders oligos. The reviewer inspects frame, junctions, unique sites, primers, and host or backbone fit, records a decision on each row, and signs a hold-or-order instruction. This is not a software feature list, not colony QC, and not a promise the later clone will work. It is the last cheap place to stop a bad plan.

Outcome, Prerequisites, and Review Boundary

The outcome: a signed review record. It names the construct, the five technical checks, pass or fail for each, the signer, and a clear order-or-hold instruction. A map with margin comments is not the outcome. A meeting that "looked fine" is not the outcome.

Prerequisites: the expected assembled sequence already exists. That object is the product of a cloning planner, not of a viewer that only opened the fragments. SeqBench's cloning workflow puts the same gate in different words: build the intended assembled sequence before a single oligo is synthesized, then QC that assembled construct, not only the parts. If you do not yet have the assembled map, stop and plan. Do not review a sketch.

The boundary: this page is the pre-order review. It is not Addgene's intake QC on a physical sample, not a diagnostic digest after transformation, and not a full cloning protocol. Those later checks still matter. They are more expensive. They are not this workflow.

Review the Construct in Order

Do not reorder the technical checks to match a favorite window. Frame errors hide in junctions; unique-site failures hide in the assembled sequence; primer problems hide until someone tries to verify the clone. Sign-off is last because it is a decision, not a sixth sequence test.

1. Frame. Translate every coding span that is supposed to stay in frame, including tag and fusion junctions. A one-base insertion at a seam shifts everything downstream. Fail if the intended ORF is not an open frame from the intended start to the intended stop. Checkpoint: the reviewer can point at the translated junction, not only at a colored arrow on the map.
2. Junctions. Inspect every seam — insert to backbone, fragment to fragment, tag to cargo. Addgene's QC notes treat cloning junctions as the place most sequence-assembly errors occur, and they include in-frame fusions among the features that distinguish a plasmid from its predecessor. Fail if a junction sequence was not actually inspected, or if it does not match the stated assembly method. Checkpoint: each seam has a recorded look, not a global "map looks clean."
3. Unique sites. Confirm the diagnostic or screening sites you plan to use are unique on the assembled construct, and that method-critical sites — a Type IIS site for Golden Gate, a restriction site you need later — were not created inside an insert by the join. SeqBench flags restriction sites that appear only after stitching. Fail if a planned unique site is duplicated, or if an internal site would wreck the method. Checkpoint: the reviewer lists the sites they relied on and shows they occur once, or not at all where they must be absent.
4. Primers. Check the oligos you are about to order, and the sequencing primers you will need later. Addgene's sequence-analysis protocol requires a sequencing primer that anneals once on the entire construct and sits a usable distance from the target. Fail a cloning primer that is not specific to its template, or a pair whose melting temperatures are mismatched enough that one half will not anneal with the other. Fail a verification primer that hits twice. Checkpoint: each primer has a uniqueness result on the assembled map.
5. Host and backbone fit. Ask whether the backbone can do this experiment in this host. A cloning vehicle that only propagates DNA is the wrong backbone for expression; the switch rules live on the cloning-vector versus expression-vector page. If the construct is expression-bound, codon usage bias is a host-preference problem that can fail the same row. Fail if the promoter, origin, marker, or coding-sequence design does not match the intended host job. Checkpoint: the reviewer can name the host and the backbone job in one sentence.
6. Sign-off. A named reviewer who did not design the construct records pass or fail on each row and writes order or hold. The designer may present the map. They do not sign their own pass. Checkpoint: the record shows a second name and a hold-or-order line.

Once those checks are executable on any annotated map, the review needs a place both people can see the same object. After the procedure is clear, ZettaGene is one example of a shared map with visualization and editing, plasmid construction, and primer design — a place to inspect features, not a substitute for the signer. The review is the table and the name. The window is optional.

Check, Pass Standard, and Fail Action

Fill this table on the expected assembled construct. A single fail blocks the order until that row is re-reviewed.

The table records a decision. It does not predict transformation efficiency, expression, or a clean Sanger trace.

Expected Result and Verification

The review is finished when the record contains all of the following: construct identity, the five technical rows, pass or fail on each, the signer's name, the date, and an order-or-hold instruction that a purchasing person can follow without asking the designer. If any field is missing, the review is not done.

Verification of this workflow is verification of the review, not of the later clone. A second person should be able to read the record and state, without a meeting, whether oligos may be ordered. Physical verification — colony screens, diagnostic digests, sequencing against the expected reference — still happens later. Those tests compare a molecule to this approved map. They cannot rescue a map nobody signed.

What a Fail Means Before Anyone Orders Oligos

A fail freezes the oligo and fragment order. The designer owns the revision. The reviewer owns the re-check on the revised expected sequence. Ordering through a fail is a process breach, not a shortcut. The cheap property of this workflow is timing: the fail happens before synthesis. After oligos exist, the same error is a wasted reaction and a delayed project, not a table row.

Do not convert a fail into optional commentary. "Looks fine except the junction" is a fail on the junction row. "Order the primers and we will fix the backbone later" is a fail on host and backbone. The hold stays until that row is a pass and the signer writes order.

Frequently Asked Questions

Who should sign off a cloning construct review?

Someone who can reject the order and who did not design the construct. The designer may walk the reviewer through the map. They do not sign their own pass. If the lab has only one molecular biologist, the PI or a trained second reader still has to put a name on the hold-or-order line.

Do we review the assembled construct or only the parts?

The assembled construct. Frame shifts, junction scars, and restriction sites created by a join will not show when you only inspect the original fragments. If the assembled sequence does not exist yet, you are not in review — you are still planning.

What does a fail mean before oligos are ordered?

The order is frozen until that check is a pass on a revised expected sequence and the signer writes order. It is not a suggestion, and it is not permission to order "the safe oligos" while the rest is fixed.

When should host or backbone fit fail the review?

When the backbone cannot do the experiment's job in the intended host — a cloning vehicle asked to express, a promoter built for the wrong species, or a coding sequence the host is known to punish for codon usage. Use the vector-job and codon-usage pages for the biology; this row only decides whether that mismatch blocks the order.