Supplemental Data for:
Baskerville and Bartel, RNA 11, pp. 241-247


Supplemental Material
Figure S1. (A) Pilot microarray. miRNA probes were printed to test hybridization and wash conditions. Samples for hybridization to the array were individually labeled with either (Cy3, green) or (Cy5, red), mixed in equamolar amounts, and hybridized to the array at the temperatures indicated. Specific hybridization was observed as either red or green spots on the array, whereas crosshybridization was observed as shades of yellow. Each miRNA probe was selected to test the resolution of the array for single and multiple mismatches between probe sequences. For example, fug-mir-197 and hsa-mir-197 test the resolution of single nucleotide differences within three nucleotides of the 3_ end of the probe sequence. Test probes were also designed to span a range of melting temperatures. As the hybridization temperature was increased, the specificity of the test array was generally improved to a point near 55-60 C (see hsa-mir-148 and hsa-mir-136). (B) Representative sample grid of a microarray. A more extensive microarray was printed with probes covering a set of non-redundant human and mouse miRNA sequences. As a test of the array specificity, the oligonucleotides used to construct the reference sample were split into two different sets and labeled as described above, mixed and hybridized to the array at 57 C. This subgrid contains replicates samples of individual probes (top four rows and bottom four rows); test samples reproducibly hybridize within the array, and little cross-hybridization was observed, although it is visible in some cases (yellow spots in subgrid shown).

Figure S2. Correlation between cloning frequency and microarray score.
Small RNAs sequences were amplified from total RNA isolated from C. elegans mixed-stage worms and then dye-labeled for hybridization to the array. Microarray scores for mixed-stage C. elegans samples were compared with published cloning frequencies from the same samples. In general, cloning frequency and chip score are positively correlated (R2. = 0.66), similar to the results comparing cloning frequency to molecular abundance (R2 = 0.78) (Lau et al., 2001).

Figure S3. Biological replicates of microarray data. Microarray sample libraries were prepared from primary human hepatocytes enriched from two separate transplant-grade livers. Each library was cloned independently and hybridized to the array. The reproducibility of the array data between these samples (R2 = 0.93) suggests that the profile of miRNA expression within organs is consistent and reproducible, as are the methods for amplifying and labeling the miRNA sequences.

Supplemental Table 1. Microarray Data

Supplemental Table 2. Microarray Probe Sequences