Illumina Use Structured Illumination For Super-Resolution

In January of 2020 Illumina announced their super-resolution instruments. Since then I’ve had a draft post sitting here waiting for me to write something about Illumina’s super-resolution sequencing approach.

However I could never find any definitive statements, or IP with convincing demonstrations showing exactly how it worked. Illumina’s IP on super-resolution sequencing is all over the place. They have IP covering field based deactivation, STED approaches, stochastic photo-switching and other approaches. Most of this didn't seem very practical, and it seemed most likely that Illumina was using some kind of structured illumination approach.

Structured Illumination isn’t the “best” super-resolution approach. It’s limited to resolving features down to ~100nm (the best can get down to single digit nanometers). But it seemed to me that this was the only approach that could be integrated into the current platform practically.

Well, I was googling around and lo and behold Illumina have finally disclosed that they are using structured Illumination! A post from late last year describes the process and issues in some detail. It stops just short of saying “and this is what we use in the NextSeq 2000 and NovaSeq”, but the message is pretty clear.

Structured Illumination is exactly what it sounds like. Rather than illuminating the flow cell with a flat field the light is “structured”. By combining images created under different illumination a super-resolved image can be formed, breaking the diffraction limit and allowing you to image much smaller features:

The structured Illumination approach from this post by Illumina.

The remaining question however, is how big are Illumina’s wells on super-resolution systems? The last (pre-super-resolution) SEM images I saw showed wells of ~400nm diameter. Taking a look at more recent patents show features in the 200 to 300nm range:

Fig 9B “Cross sectional SEM image of the imprinted structure” from this patent. Fig 10F from this patent (an imprinted structure, possibly from a flow cell template?).

Overall, I kind of suspect we’re currently in the 200 to 350nm range. But will continue looking for more explicit information. Let’s dig a little more into the implications structured illumination based super-resolution after the break!