Difference between revisions of "System complexity scaling of self-assembly"

Revision as of 13:10, 16 November 2024

As of time of writing (late 2024) we may still be far from having hit the ultimate limits of self-assembly. Even with the constraints of remaining fro the most part fully in the realm of topological atomic precision. I.e. not veering out into the realm of synthetic biology aiming to carbon copy the workings of molecular biology with heavy use of vesicles and membranes.

Note: The wording "scaling of system complexity" is chosen intentionally here as
just "scaling" could be easily mistaken with size and a mere scaling in size may be
much less relevant for quantifying progress than system complexity.
For details on that see page: Quantifying progress by scaling in achievable complexity

The example of 3D structural DNA nanotechnology (3D-SDN)

3D structural DNA nanotechnology already managed to scale complexity by a fair bit.
Perhaps a good part of the way to the scale of a foldamer printer.
Well, 3D-SDN is clearly not stiff enough though for positional atomic precision.
It's stiffness could perhaps suffice for things like (as of 2024 not yet experimentally shown) Tether assisted assembly
which can accommodate for quite large structural deformations.
Tether assisted assembly would be an in-between between positional-assembly and self assembly that
has not yet been experimentally explored as of time of writing (2024).

3D structural DNA nanotechnology has been scaled up to gigadalton scale (hundret thousand atoms) and almost micron scale.
See ^[1] (not an open access paper unfortunately).
Granted that this lacks in termination control to the here in this wiki used definition of it
where full unbroken spherical symmetry is considered a generalized form of non-termination.
But there is other experimental work showing more termination control
at the same hierarchical self-assembly level and only slightly smaller size scale.
See: ^[2] (also not an open access paper unfortunately).

Known limits / challenges

• exponential slowdown due to part depletion thin-out
• slower diffusion speeds of larger parts
• Limits of persistence length

Tricks & workaround methods

Hierarchical selfassembly

Already experimentally demonstrated
as mentioned in the preceding 3D-SDN section.

Iterative self-assembly & Multi pot self-assembly

These may be especially helpful when making orthogonal sets of binding interfaces is hard(er). As e.g. with de-novo proteins vs 3D-SDN. Orthogonal set meaning both high (re)activity and high specificity. Matching faced binding strongly mismatching faces binding weakly. Prominent binding matrix diagonal.

Circumsembly

Removing a factor of exponential drop-off in yield by providing parallel redundant pathways. 3D-SDN already does this. Internal missing pieces can be irrelevant in many cases. Related page: Steric trap.

Squigglesembly

For building 2D structures with mere 1D assembly capabilities.
Note that this is possibly not yet experimentally investigated as of tome of writing (2024).
It should be combinable with other techniques for higher dimensionality (3D).
Again a technique that may be especially helpful when making orthogonal sets of binding interfaces is hard(er).
As with de-novo proteins.

Increase of effective concentration

In some cases 3D-SDN does this by providing a long scaffolding/templating/seed strand.
Similar strategies might be possible with de-novo proteins eventually.
The scaffolding strand method is specific to SDN.

Advanced nucleation control

Some few nucleation seeds can be intentionally brought in
in order to to knock down an intentionally too high collaborative nucleation barrier.
This way one gets a few fully completed assemblies rather than many partially completed assemblies when the
fee-stock is nigh depleted.

Weak pre-bonding based misassembly self correction

In SDN wrongly self-assembled sites can come apart again by iterative zipping.
Coorect assemblies get stabilized by collaborative bonding then there-after.
Flexibility of the self-assembled chains is critical for these sort of processes.
Unfortunately this is direct opposition to the goal of eventually getting to structures of higher stiffness.

Note that this this is (can be?) a critical prerequisite for algorithmic self assembly.

Algorithmic selfassembly

In principle this can give termination control over much larger size scales.
It is no longer full "random access" termination control though.
Also the technique is generally more challenging.
As of time of writing (late 2024) the author is unaware of
exprimental work on algorithmic selfassembly in the space of de-novo proteins.

Tether assisted assembly

This kind of
– an easier form of self-assembly (smaller space to address) mixed with
–an easier form of positional assembly (much reduced requirements of positional accuracy)
Very much unexplored territory as of 2024.

Non-thermal self-assembly

At scales that no longer support fast enough diffusion transport
there is still non-thermal self-assembly left as an option (with some constraints).

DNA vs protein vs other

Some stuff that only works with DNA (or similar).
Some stuff is more for proteins.
(wiki-TODO: Maybe make this clearer somehow by some repeated analysis pattern.)

Related

• Hierarchical selfassembly
• Iterative self-assembly & Multi pot self-assembly
• Circumsembly (SDN ready does this, internal missing pieces can be irrelevant in many cases)
• Squigglesembly (possibly not experimentally investigated as of 2024)

• Non-thermal self-assembly

• Tether assisted assembly mining in some positional assembly into self-assembly

External links

• https://en.wikipedia.org/wiki/Persistence_length

Papers

1. ↑ Wagenbauer, K. F., Sigl, C., & Dietz, H. (2017). Gigadalton-scale shape-programmable DNA assemblies. Nature, 552(7683), 78–83. doi:10.1038/nature24651
2. ↑ Zhou, Yihao & Dong, Jinyi & Zhou, Chao & Wang, Qiangbin. (2022). Finite Assembly of Three‐Dimensional DNA Hierarchical Nanoarchitectures through Orthogonal and Directional Bonding. Angewandte Chemie International Edition. 61. 10.1002/anie.202116416.