Hierarchal analysis of mt- & nuc-DNA gene expression arrays

          The figure shows the results of a hypothetical experiment on fish growth metabolism with a small portion of a quantitative gene expression microarray. The array includes EST tags for cDNAs from four mtDNA and four nucDNA loci that contribute to nucleo-cytoplasmic holoenzymes (ATP-synthase, Cytochrome b, Cytochrome Oxidase, & NADH) involved in mitochondrial energy transactions. There are also four ESTs for a set of unknown ("UNK") but mitochondrially-abundant mRNAs.

          In the left-hand figure, mRNA expression is measured in Early, Middle, and Late-stage organisms. “Up-regulation”, that is, increased expression with respect of a control, is indicated by a red pseudo-color, and “Down-regulation” by green. Bioinformatic analysis (such as TreeView) performs a hierarchal clustering that sorts the 12 locus-specific expression patterns into groups with similar patterns.

          Consider three lines of fish that show standard, fast, and slow growth. Four patterns are detected in fish with standard (control) growth rates:
            “early down-regulated”    (3,7,9)
            “middle down-regulated” (4,6,11)
            “middle up-regulated"      (2,5,10)
            “late up-regulated”           (8,1,12)

    In each cluster, a pair of nucleo-cytoplasmic loci are coordinately regulated, along with one of the unknown (UNK) loci.  This suggests a hypothesis, that each UNK locus is involved in the control of the corresponding holoenzyme.  Fish with altered growth rates show characteristic changes in gene regulation. In fast-growing fish, the CO system (4,6,11) is up-regulated in late-stage (e.g., post-hatching) larvae.  In slow-growing fish, the ATP-synthase system (8,1) is down-regulated. The presumptive ATP-related locus UNK-12 remains constitutive, which indicates a more complex pattern of regulation.