Single-cell proteomics reveals changes in expression during hair-cell development.

Hearing and balance rely on small sensory hair cells that reside in the inner ear. To explore dynamic changes in the abundant proteins present in differentiating hair cells, we used nanoliter-scale shotgun mass spectrometry of single cells, each ~1 picoliter, from utricles of embryonic day 15 chickens. We identified unique constellations of proteins or protein groups from presumptive hair cells and from progenitor cells. The single-cell proteomes enabled the de novo reconstruction of a developmental trajectory using protein expression levels, revealing proteins that greatly increased in expression during differentiation of hair cells (e.g., OCM, CRABP1, GPX2, AK1, GSTO1) and those that decreased during differentiation (e.g., TMSB4X, AGR3). Complementary single-cell transcriptome profiling showed corresponding changes in mRNA during maturation of hair cells. Single-cell proteomics data thus can be mined to reveal features of cellular development that may be missed with transcriptomics.

Large-scale modeling of condition-specific gene regulatory networks by information integration and inference.

Understanding how regulatory networks globally coordinate the response of a cell to changing conditions, such as perturbations by shifting environments, is an elementary challenge in systems biology which has yet to be met. Genome-wide gene expression measurements are high dimensional as these are reflecting the condition-specific interplay of thousands of cellular components. The integration of prior biological knowledge into the modeling process of systems-wide gene regulation enables the large-scale interpretation of gene expression signals in the context of known regulatory relations. We developed COGERE (http://mips.helmholtz-muenchen.de/cogere), a method for the inference of condition-specific gene regulatory networks in human and mouse. We integrated existing knowledge of regulatory interactions from multiple sources to a comprehensive model of prior information. COGERE infers condition-specific regulation by evaluating the mutual dependency between regulator (transcription factor or miRNA) and target gene expression using prior information. This dependency is scored by the non-parametric, nonlinear correlation coefficient η(2) (eta squared) that is derived by a two-way analysis of variance. We show that COGERE significantly outperforms alternative methods in predicting condition-specific gene regulatory networks on simulated data sets. Furthermore, by inferring the cancer-specific gene regulatory network from the NCI-60 expression study, we demonstrate the utility of COGERE to promote hypothesis-driven clinical research.