Complementary Regulatory Networks Enabled By Parallel Transcriptome and Proteome Dynamics

Systems genomics can take advantage of system dynamics to elucidate functional relationships between genes. Transcriptome dynamics have long been used for this purpose and more recently proteome and metabolome dynamics have been used. We created a protein atlas of maize that exploits developmental dynamics to provide global descriptions of the spatiotemporal patterns of protein accumulation and post-translational state throughout development. The atlas data was used to identify new genes and to refine gene models through the emerging field of proteogenomics. In parallel with protein measurements we used RNAseq to define the transcriptome from 23 of the same 33 samples used for the protein atlas. The Pearson correlation between global mRNA and protein abundance was approximately 0.5 for each tissue. In some cases high protein abundance was associated with non-detectable mRNA which could be partially explained by cycling mRNA with stable protein; by early expression followed by rapid mRNA decay but stable protein; and by transport of protein to adjacent tissue. We compared gene regulatory networks made using only RNAseq data to networks made using protein data for transcription factors and mRNA data for target genes. The RNA-only network was more extensive, representing 41,000 genes (2100 transcription factors). The RNA-protein hybrid network was limited by the proteome coverage to only 20,000 genes (760 transcription factors). Interestingly, very few edges were shared between the two networks indicating that mRNA and protein measurements detect different but complementary regulatory relationships between genes.