Sartor MA et al. (2006), A new method to remove hybridization bias for i...

XB-ART-873

Nucleic Acids Res 2006 Jan 05;341:185-200. doi: 10.1093/nar/gkj413.

Show Gene links Show Anatomy links

A new method to remove hybridization bias for interspecies comparison of global gene expression profiles uncovers an association between mRNA sequence divergence and differential gene expression in Xenopus.

Sartor MA , Zorn AM , Schwanekamp JA , Halbleib D , Karyala S , Howell ML , Dean GE , Medvedovic M , Tomlinson CR .

???displayArticle.abstract???
The recent sequencing of a large number of Xenopus tropicalis expressed sequences has allowed development of a high-throughput approach to study Xenopus global RNA gene expression. We examined the global gene expression similarities and differences between the historically significant Xenopus laevis model system and the increasingly used X.tropicalis model system and assessed whether an X.tropicalis microarray platform can be used for X.laevis. These closely related species were also used to investigate a more general question: is there an association between mRNA sequence divergence and differences in gene expression levels? We carried out a comprehensive comparison of global gene expression profiles using microarrays of different tissues and developmental stages of X.laevis and X.tropicalis. We (i) show that the X.tropicalis probes provide an efficacious microarray platform for X.laevis, (ii) describe methods to compare interspecies mRNA profiles that correct differences in hybridization efficiency and (iii) show independently of hybridization bias that as mRNA sequence divergence increases between X.laevis and X.tropicalis differences in mRNA expression levels also increase.

???displayArticle.pubmedLink??? 16397297
???displayArticle.pmcLink??? PMC1325202
???displayArticle.link??? Nucleic Acids Res
???displayArticle.grants??? [+]

???attribute.lit??? ???displayArticles.show???

	Figure 1. Experimental design for the microarray studies. mRNA expression levels from three corresponding biological replicates of X.laevis (Xl) and X.tropicalis (Xt) tissues (ovary and liver) and developmental stages (egg; stage 10, St. 10 and stage 40, St. 40) were compared to each other and with a reference RNA (Ref.). The reference RNA was composed of equal amounts of total RNA of the above tissues and developmental stages for a given Xenopus species. Each double pointed arrow represents three microarray slides, one slide per biological replicate, in which one of the three slides was ‘dye flipped’.
	Figure 2. A measurement of microarray hybridization efficiencies for X.laevis versus X.tropicalis. (A) The experimental design to determine the hybridization efficiencies of X.laevis transcripts from ovary, liver, egg; stage 10 (St. 10) and stage 40 (St. 40) to the X.tropicalis DNA microarray probes. Corresponding RNA samples from three biological replicates of the tissues and developmental stages for a given Xenopus species were directly compared to each other by microarray analysis. (B) A plot of log average spot intensities from X.laevis (Y-axis) versus X.tropicalis–X.laevis probe sequence similarity (X-axis). (C) A plot of the log-ratio hybridization bias (Y-axis) versus X.tropicalis–X.laevis probe sequence similarity (X-axis). The Y-axis represents the expected offset from actual relative mRNA expression levels (X.tropicalis/X.laevis).
	Figure 3. Differential mRNA expression levels for X.laevis (Xl) and X.tropicalis (Xt) are correlated with mRNA transcript sequence divergence. (A) The experimental design to determine whether X.laevis and X.tropicalis differential gene expression for ovary, liver, egg, stage 10 (St. 10) and stage 40 (St. 40) is associated with sequence divergence. mRNA from three biological replicates of the tissues and developmental stages for a given Xenopus species were compared with a corresponding reference RNA (Ref.), in turn, the corresponding ratios from each Xenopus species were compared to each other. (B) Predicted relative gene expression levels of X.laevis to X.tropicalis determined from local regression analysis (Y-axis) versus probe sequence similarity (X-axis). Only the 1681 genes described in the text that were significantly changed relative to the corresponding reference RNA were plotted. (C) A plot of the squared correlation coefficient values (Y-axis) versus probe sequence similarity (X-axis). The Y-axis represents the square of the correlation coefficients for the 1681 genes.
	Figure 4. The gene expression profiles for X.laevis (Xl) and X.tropicalis (Xt) are similar. (A) The experimental design to determine how X.laevis and X.tropicalis compare in their global gene expression profiles for selected tissues (ovary and liver) and developmental stages (egg, stage 10 and stage 40). mRNA levels from three biological replicates from the two tissues and three developmental stages were compared to a reference RNA (Ref.) for a given Xenopus species. The estimates of log [(Xt/Xt Ref)/(Xl/Xl Ref)] for each gene were compared between X.laevis and X.tropicalis to determine correlation coefficients. (B) Histogram representing the correlation coefficients between X.laevis and X.tropicalis using 1681 transcript levels that changed significantly among the different tissues and stages.
	Figure 5. The global gene expression profiles for X.laevis (Xl) and X.tropicalis (Xt) follow parallel temporally-regulated developmental programs. (A) The part of the experimental design used to determine how X.laevis and X.tropicalis compare in their global gene expression profiles for selected developmental stages (egg; stage 10, St. 10 and stage 40, St. 40). mRNA levels from three biological replicates from the three developmental stages for a given Xenopus species were compared with a reference RNA (Ref.) and mRNA from egg was compared with stage 10 and to stage 40 via the reference RNA. (B) Hierarchical tree of genes and heat map of the developmental stages, in which corresponding stage 10 and stage 40 mRNA levels were compared to egg mRNA levels for each Xenopus species. The top 200 ranked genes in each comparison that were at least 50% changed were included. The heat map columns left to right are: X laevis stage 10 versus X.laevis egg, X.tropicalis stage 10 versus X.tropicalis egg, X.laevis stage 40 versus X.laevis egg, and X.tropicalis stage 40 versus X.tropicalis egg. The brackets to the right of the heat map numbered 1–3, designate groups of genes that are contrary to the overall clustering trend and are described in the text.

References [+] :

Adjaye, Cross-species hybridisation of human and bovine orthologous genes on high density cDNA microarrays. 2004, Pubmed