More nCounter Notes
Summary
As previously noted, the nCounter uses a simple optical system, essentially a regular Epi-fluorescence microscope. While it’s a relatively simple system, the raw single molecule data is very clean.
The nCounter SPRINT (which combines the optics and fluidics, and uses a non-immersion objective), likely simplifies the system even further.
While not a huge commercial success, the nCounter shows how a simple optical system can be used in the context of single molecule (but not fluorophore) detection.
This may provide lessons for other single molecule detection approaches…
Hacking Around…
In order to further explore the optical system I decided to try and grab some images through the nCounter optics. Unfortunately I couldn’t get the Mercury Lamp to fire… So I pulled the whole thing apart and used a light source I have knocking around…
As I wasn’t going to be using the built in light source and excitation filters. I decided to take a look:
I noted down the filter markings (listed below). But it looks like we’re looking at filters centered around 474nm, 547nm, 558nm, 642nm. In addition to this there a filter stack marked (UVND) I assume for taking brightfield images, knocking out the UV from the lamp.
Images
I removed the CoolSNAP ES2 camera and shoved a FLIR IMX183 camera in its place. This offsets the sensors ~5mm further back than the original camera. I assume this causes some issues. But images looked reasonable.
Here’s an image from the HiSeq X calibration slide (well it fell out of a HiSeq X, I think it’s just a regular BeadArray) taken through this system:
And here’s a USAF1951:
Pixels on this camera are 2.4 micron. Based on the USAF1951 the image should be 319 by 212 microns on the surface. The image is 5472x3648 pixels, so 13132.8x8755.2 microns.
That’s ~40x. As previously noted, the system has a 60x objective, so I’m unclear on where the difference is coming from!
A few more notes below…