Searching QTL by gene expression: analysis of diabesity

Abstract

Background: Recent developments in sequence databases provide the opportunity to relate the expression pattern of genes to their genomic position, thus creating a transcriptome map. Quantitative trait loci (QTL) are phenotypically-defined chromosomal regions that contribute to allelically variant biological traits, and by overlaying QTL on the transcriptome, the search for candidate genes becomes extremely focused.

Results: We used our novel data mining tool, ExQuest, to select genes within known diabesity QTL showing enriched expression in primary diabesity affected tissues. We then quantified transcripts in adipose, pancreas, and liver tissue from Tally Ho mice, a multigenic model for Type II diabetes (T2D), and from diabesity-resistant C57BL/6J controls. Analysis of the resulting quantitative PCR data using the Global Pattern Recognition analytical algorithm identified a number of genes whose expression is altered, and thus are novel candidates for diabesity QTL and/or pathways associated with diabesity.

Conclusion: Transcription-based data mining of genes in QTL-limited intervals followed by efficient quantitative PCR methods is an effective strategy for identifying genes that may contribute to complex pathophysiological processes.

Figure 1

Mirrored tissue-specific expression (red) and absolute EST representation (green) for whole ExQuest chromosomes. Maximum scales set for specific or absolute expression is within the circle at the centromere. (a) Pancreas expression for mouse chromosome 10 showing peak position of the LOD for the QTL Tafat over a pancreas expression peak that contains the gene Ela2. (b) Quantitative tissue specific expression profile for the gene Ela2. Results are derived from the number of Ela2-aligned EST hits normalized for library density. (c) Pancreas expression for mouse chromosome 19. Genes in the Tinidd1 and T2dm2 region are pancreas specific (red) and highly expressed (green). Arrow shows a 20X zoom of 3 individual genes Pnlip (1810018F18Rik), Pnliprp1 and Pnliprp2, which contribute to the large pancreatic histogram peak.