Reading and Analysis
Data is collected as with other image based technologies. A “shoot and scoot” cycle takes a picture, moves the sample, and repeats. The pictures are interpreted with software that recognizes DNA strands and (optionally) the individual label types used in a given sample.
The sample is inserted into the electron microscope (EM) on the ZSG substrate. Holders designed for the EM can contain several samples. The holder is inserted into the EM which seals the chamber and quickly creates a vacuum.
Position – Take Picture – Repeat. The automated Reader is that easy. Put it in, turn it on and the system generates the data. The system will execute the image capture routine appropriate to the product and application. The number of shots depends on the resolution (detail in the image) required by the application. If a scientist uses only a portion of the microarray or wants to capture data for only a subset of the sample molecules, this information is input before the automated imaging routine is begun.
Higher resolution applications require more images. For example, if we are imaging a nucleotide base, we want five pixels to read its pitch of one-third of a nanometer. An analogy is the effect of zooming on a consumer digital camera. You see increasing detail in the picture, but it covers a decreasing area in the shot.
The speed of digital camera shots is high. The analogy with a consumer camera is helpful. Think about personally taking a camera shot, re-position to the next shot, auto-focus then shoot again. More expensive camera systems can process faster and therefore take more shots per second. ZSG expects to average 1.5 seconds per shot for our first product and increase it significantly over time. Our throughput is only limited by the innovative engineering of the digital camera marketplace.
Image analysis software will process the digital data in real time. Expression analysis requires a count of molecules. Sequencing requires interpreting the species of the labeling atoms. Either will happen in real-time. Critically, the sequencing molecules will be over 10,000 base-pairs long, drastically reducing downstream data requirements.