Specificity

One of most important parameters to ensure the generation of high quality microarray data is specificity. This is especially true for analyzing environmental samples since there may be numerous sequences for each gene present in a sample. GeoChip specificity is controlled by probe design and hybridization conditions. Our studies showed that hybridizations at 50 oC and 50% formamide could differentiate sequences with < 90-92% identity (Rhee et al., 2004; Liebich et al., 2006; Deng et al. 2008) (Figure 1). With such high resolution, GeoChip is able to differentiate microorganisms at the species-strain level (Tiquia et al. 2004; Liebich et al., 2006; Deng et al. 2008). Based on these results, we systematically tested and experimentally established probe design criteria by considering sequence identity, continuous sequence stretches and free energy in order to increase specificity (He et al., 2005; Liebich et al., 2006). These criteria have been implemented in a novel software tool, CommOligo, for microarray probe design (Li et al., 2005). Probes for GeoChip 2.0 were designed using this newly developed software. Evaluation of GeoChip 2.0 revealed only a very small percentage of false positives (0.002-0.004%) and no false negatives under the experimental conditions examined (He et al., 2007). Thus, GeoChip has been extensively validated, and proven to be a highly specific microarray.

(A) Partial image from the hybridization of the DNA mixture of clones of six functional genes (dsrA, nirK, nirS, nifH, pmoA, and amoA) with a prototype GeoChip. The layout of the array spots is shown in red and specific hybridizations are shown in green. (B) A small portion of GeoChip hybridization with a marine sediment sample to demonstrate specificity with environmental samples. (C) The relationship between probe-target identity and relative signal intensity at different hybridization temperatures. At 50 oC and 50% formamide, the probes with < 92% sequence identity to target samples (red arrow) have less than 10% signal (background signal level).