Skip to main content
Figure 5 | BMC Genomics

Figure 5

From: Microarray amplification bias: loss of 30% differentially expressed genes due to long probe – poly(A)-tail distances

Figure 5

Probe – poly(A)-tail distance in relation to differentially expressed genes. A, Boxplot of probe – poly(A)-tail distances for 21,103 Agilent WHG probes (median 542 nt, range, 0–22,006) (all genes); B, Boxplot of probe – poly(A)-tail distances for the 872 differentially expressed genes after one round (median 289, range, 0–7090) (one round) and 839 differentially expressed genes after two rounds (median 212, range, 0–7621) (two rounds). Mann Whitney analysis revealed a significant decrease of the probe – poly(A)-tail distance after two rounds as compared to one round of amplification (p < 0.0001); C, 246 of the 872 genes (~30%) differentially expressed after one round are not identified as being differentially expressed after two rounds of amplification (lost in 2nd). These 246 genes have a significant longer probe – poly(A)-tail distance as compared to the remaining 626 genes which were identified as differentially expressed for both one and two rounds of amplification (persistent in 2nd) (Mann Whitney: p < 0.0001). Ten of these inconsistent genes, with a probe – poly(A)-tail distance of more than 500 nt, were selected for validation with qRT-PCR;

Back to article page