The effects of different sample labelling methods on signal intensities of a 60-mer diagnostic microarray.

The effects of four different labelling methods on signal intensities of a 60-mer diagnostic microarray were studied. Eighty of virus-specific oligonucleotide probes for human influenza virus were prepared in an array of 15x16 spots. RNA samples from cultured human influenza virus strains were labelled with four different methods, including direct cDNA labelling (DL), universal primer labelling (UPL), direct cDNA labelling with restriction display (DL-RD), and Cy-dUTP incorporated cDNA labelling with restriction display (IL-RD) in a signal color format. The background-subtracted signal intensities from five replicate hybridization experiments of each labelling method were analyzed using one-way analysis of variance (one-way ANOVA) and linear regression techniques. The effect of sample labelling method on background-subtracted signal intensities was significant (p<0.001) and multiple comparisons showed the differences existed mainly between DL and the other three labelling methods. The sample labelling method explained about 4.3% of signal intensity. The results demonstrated that UPL and the RD-based methods are more efficient than the conventional DL method for sample labelling, an important variation factor affecting the signal intensities in diagnostic microarrays.