The surface of the Affymetrix array is like a giant checkerboard. The 10K array, which is a piece of glass about the size of a thumbnail, has over 400,000 squares. These squares are called features. The features on Affymetrix arrays are incredibly small — about 8 microns across. By way of comparison, a human hair is about 50 microns wide.

Each of these squares holds one unique type of DNA strand — called a probe. So, in our example, one square would hold the ATTCATG probe we built, and another square would hold the ATTTATG probe. But, while each square holds one type of probe, there isn't just one probe in each square, but millions of identical copies of the same probe.

Affymetrix builds these probes one layer at a time, using the same type of manufacturing technology that is used to build computer semiconductors. The molecules are built one layer at a time, one stacked on top of another, like bricks. Multiple probes are synthesized in parallel.

For our example, we are just going to look at one corner, or a single feature, of an imaginary array. Normally, each probe is 25 bases long, but for purposes of illustration here, let's abbreviate this standard probe length and say the probe is only a 7 base sequence, ATTCATG. The SNP is in the middle position and is highlighted in red. Somewhere else on the array, there is a probe representing the other possible SNP genotype, ATTTATG.