Molecular Characterization of Zebrafish Embryogenesis via DNA Microarrays and Multiplatform Time Course Metabolomics Studies

• Published in: Journal of proteome research Vol. 10; no. 11; pp. 5102 - 5117
• Main Authors: Soanes, Kelly H, Achenbach, John C, Burton, Ian W, Hui, Joseph P. M, Penny, Susanne L, Karakach, Tobias K
• Format: Journal Article
• Language: English
• Published: United States American Chemical Society 04.11.2011
• Subjects: Algorithms; Amino Acids - metabolism; Animals; Blastula - metabolism; Fish Proteins - genetics; Gastrula - metabolism; Gene Expression; Gene Expression Profiling; Magnetic Resonance Spectroscopy; Metabolomics; Multivariate Analysis; Oligonucleotide Array Sequence Analysis; Principal Component Analysis; Zebrafish - embryology; Zebrafish - genetics; Zebrafish - metabolism; data fusion; NMR; embryogenesis; Danio rerio; time course metabolomics and transcriptomics; multivariate curve resolution; LC−MS
• ISSN: 1535-3893; 1535-3907; 1535-3907
• DOI: 10.1021/pr2005549

One of the greatest strengths of “-omics” technologies is their ability to capture a molecular snapshot of multiple cellular processes simultaneously. Transcriptomics, proteomics, and metabolomics have, individually, been used in wide-ranging studies involving cell lines, tissues, model organisms, and human subjects. Nonetheless, despite the fact that their power lies in the global acquisition of parallel data streams, these methods continue to be employed separately. We highlight work done to merge transcriptomics and metabolomics technologies to study zebrafish (Danio rerio) embryogenesis. We combine information from three bioanalytical platforms, that is, DNA microarrays, 1H nuclear magnetic resonance (1H NMR), and mass spectrometry (MS)-based metabolomics, to identify and provide insights into the organism’s developmental regulators. We apply a customized approach to the analysis of such time-ordered measurements to provide temporal profiles that depict the modulation of metabolites and gene transcription. Initially, the three data sets were analyzed individually but later they were fused to highlight the advantages gained through such an integrated approach. Unique challenges posed by fusion of such data are discussed given differences in the measurement error structures, the wide dynamic range for the molecular species, and the analytical platforms used to measure them (i.e., fluorescence ratios, NMR, and MS intensities). Our data analysis reveals that changes in transcript levels at specific developmental stages correlate with previously published data with over 90% accuracy. In addition, transcript profiles exhibited trends that were similar to the accumulation of metabolites over time. Profiles for metabolites such as choline-like compounds (Trimethylamine-N-oxide, phosphocholine, betaine), creatinine/creatine, and other metabolites involved in energy metabolism exhibited a steady increase from 15 hours post fertilization (hpf) to 48 hpf. Other metabolite and transcript profiles were transiently rising and then falling back to baseline. The “house keeping” metabolites such as branched chain amino acids exhibited a steady presence throughout embryogenesis. Although the transcript profiling corresponds to only 16 384 genes, a subset of the total number of genes in the zebrafish genome, we identified examples where gene transcript and metabolite profiles correlate with one another, reflective of a relationship between gene and metabolite regulation over the course of embryogenesis.