Gibson assembly plan

A Gibson reaction drawn as the graph it is: which fragments join at which terminal homology, and in what order.

What it doesCopy Anchor Link

A Gibson assembly is not a list of fragments, it is a chain of terminal homologies. Each fragment's ends overlap its neighbours', the exonuclease chews back, the ends anneal, and the arrangement that comes out is decided by which ends are complementary rather than by the order the fragments were supplied in. So the useful question — what will actually close in this tube, and in what arrangement — is a question about a graph.

This is that graph for the construct you have selected. Each level is one join, an edge is a fragment going into the reaction below it, and the edge is labelled with the homology that makes the join: how long the overlap is, and its melting temperature.

Seeing it matters because a wrong assembly reads as a correct list. Two fragments whose ends happen to be complementary produce a plausible set of names and a construct in the wrong order, and a fragment used the other way round produces the right length with the wrong sequence. Both are visible in the shape and in neither case does the total length change.

When you'd use thisCopy Anchor Link

• Before ordering primers. The arrangement the homologies force is the one you will get, and a transposed pair is cheaper to notice now.
• When the product is the right size and the wrong thing. Length is the one thing a mis-ordered assembly preserves.
• When a fragment came out reverse-complemented. The mix searches both orientations, so that is a valid solution rather than an error, but it is often not what was intended.
• When nothing closes. The graph shows how far the chain got before it ran out of matching ends.

What this computesCopy Anchor Link

• The assembly graph of the selected construct: its levels, the fragments on each, and the edges between them, taken from the simulated reaction and not from the order the fragments were listed.
• The overlap on every edge: the length of the terminal homology that joins those two fragments.
• The annealing temperature of each homology, as a nearest-neighbour Tm with salt correction — the temperature the overlap actually anneals at, which is a different number from the naive summed-base scale an assembly's own minimum and maximum Tm constraints are expressed in. Both are reported, and which one you are reading is always named.
• The construct's own facts: its length, whether it closed circular, how many fragments it consumed, and how many of them were used reverse-complemented.

What it does not doCopy Anchor Link

It draws a simulated reaction, not an experiment. Assembly efficiency, the effect of fragment molar ratio, exonuclease chew-back beyond the designed overlap, and the frequency of assembly through a short accidental homology are all real and none is modelled here: a construct that closes cleanly on screen can still give colonies with the wrong insert. It also draws the selected construct only, so a reaction that could close in several arrangements shows them one at a time.