Performance comparison of two microarray platforms to assess differential gene expression in human monocyte and macrophage cells

Abstract

Background: In this study we assessed the respective ability of Affymetrix and Illumina microarray methodologies to answer a relevant biological question, namely the change in gene expression between resting monocytes and macrophages derived from these monocytes. Five RNA samples for each type of cell were hybridized to the two platforms in parallel. In addition, a reference list of differentially expressed genes (DEG) was generated from a larger number of hybridizations (mRNA from 86 individuals) using the RNG/MRC two-color platform.

Results: Our results show an important overlap of the Illumina and Affymetrix DEG lists. In addition, more than 70% of the genes in these lists were also present in the reference list. Overall the two platforms had very similar performance in terms of biological significance, evaluated by the presence in the DEG lists of an excess of genes belonging to Gene Ontology (GO) categories relevant for the biology of monocytes and macrophages. Our results support the conclusion of the MicroArray Quality Control (MAQC) project that the criteria used to constitute the DEG lists strongly influence the degree of concordance among platforms. However the importance of prioritizing genes by magnitude of effect (fold change) rather than statistical significance (p-value) to enhance cross-platform reproducibility recommended by the MAQC authors was not supported by our data.

Conclusion: Functional analysis based on GO enrichment demonstrates that the 2 compared technologies delivered very similar results and identified most of the relevant GO categories enriched in the reference list.

Figure 1

Intra-platform reproducibility of the relative expression intensities. Scatter plot comparison of the relative expression values (log 2 ratio of gene expression between macrophage and monocyte samples) of two different samples on Affymetrix (a) and Illumina (b) platforms. The blue line on each plot represents a regression line that best fits the plotted set of points. Both array types provide high inter-replicates reproducibility of the relative gene expression intensities.

Figure 2

Volcano plots representing the relationship between fold change and statistical significance. On the x-axis are represented the log 2 fold change between the two groups (macrophages and monocytes). The vertical axis represents the log-Odds (B-Statistic) computed in Limma. Each gene is represented by a point and an up- and down-regulated gene appears symmetric. B statistics represents the log-odds that the gene is differentially expressed between the two groups. A low B-value indicates little evidence of differential expression. Highlighted genes represent the top 20 significant genes identified on Affymetrix (a), Illumina (b) and RNG-86 (c) platforms.

Figure 3

Inter-platforms agreement in gene lists. Venn diagrams showing the overlap of genes identified as differentially expressed between macrophage and monocyte samples in the 2 compared platforms and the reference. Three different criteria were used to select gene lists: (a) Pc < 0.001, (b) Pc < 0.05 combined to fold-change >2 as suggested in the MAQC project and (c) "Best-3800" probes on each platform. Only a set of transcripts common to the 3 platforms was used in this comparison. Gene lists include both up- or down- regulated genes.

Figure 4

Effect of gene list selection criteria on the degree of inter-platforms concordance. Number of overlapping genes (y-axis) in the 2 and 3 lists of DEG according to the size of the list (x-axis). For each platform, the list was constituted by selecting DEG (Pc < 0.05), then within this list genes were ranked according to decreasing fold change. The number of overlapping genes between lists was calculated for increasing list size. When the number of probes in the lists was approximately 3800, the number of overlapping genes reached a plateau. The "best 3800" set of probes was defined accordingly.

Figure 5

Correlation of fold changes. For each pair-wise comparison, Pearson's correlation coefficients of fold change were calculated and the direction of change was examined; Genes present in the top-left and bottom right quarters of each plots show changes in opposite direction. These genes are expected to overlap by chance.