Project description:Transcription signatures have been used to stratify breast cancer patients into clinically distinct subgroups. However, transcription alone does not determine protein expression. Of potentially equal importance for determining the tumor phenotype is the rate at which transcripts are translated to form protein. Protein translation is controlled to a major degree by miRNA, and cancer cells may deregulate the expression of key genes by altering the activity of relevant miRNAs. The importance of miRNA deregulation and the extent to which multiple miRNAs coordinately deregulate key proteins in breast cancer is only partly understood. To gain such insight, we analyzed genome-wide miRNA expression and mRNA/protein expression for a panel of 105 selected cancer related genes in breast carcinomas from 283 patients. The miRNA-mRNA-protein interactome for the selected genes was constructed by modeling protein expression as a joint function of mRNA and miRNA expression, considering the effect of both one miRNA at a time, and all studied miRNAs simultaneously. The interactome represents a map of the global effects of miRNAs on protein expression, capturing direct as well as indirect effects. The results reveal extensive association between miRNA and protein expression as well as coordinated effects of multiple miRNAs on individual proteins. Applying the model onto two other independent primary breast cancer cohorts confirmed the generalizability of key aspects of the interactome map. The mRNA expression profiling of 283 breast cancer samples was performed using the SurePrint G3 Human GE 8x60K one-color microarrays from Agilent (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer’s protocol (One-Color Microarray-Based Gene Expression Analysis, Low Input Quick Amp Labeling, v.6.5, May 2010). For each sample, 100 ng of RNA was amplified and hybridized on the array. Scanning was performed with Agilent Scanner G2565A, using AgilentG3_GX_1Color as profile. Signals were extracted using FE v.10.7.3.1 and protocol GE1_107_Sep09 (Agilent Technologies). Arrays were log2-transformed, quantile normalized and hospital-adjusted by subtracting from each probe value the mean probe value among samples from the same hospital. The Oslo Breast Cancer Consortium (OSBREAC)

Project description:Protein expression is regulated by the production and degradation of mRNAs and proteins but the specifics of their relationship are controversial. Although technological advances have enabled genome-wide and time-series surveys of mRNA and protein abundance, recent studies have shown paradoxical results, with most statistical analyses being limited to linear correlation, or analysis of variance applied separately to mRNA and protein datasets. Here, using recently analyzed genome-wide time-series data, we have developed a statistical analysis framework for identifying which types of genes or biological gene groups have significant correlation between mRNA and protein abundance after accounting for potential time delays. Our framework stratifies all genes in terms of the extent of time delay, conducts gene clustering in each stratum, and performs a non-parametric statistical test of the correlation between mRNA and protein abundance in a gene cluster. Consequently, we revealed stronger correlations than previously reported between mRNA and protein abundance in two metabolic pathways. Moreover, we identified a pair of stress responsive genes (ADC17 and KIN1) that showed a highly similar time series of mRNA and protein abundance. Furthermore, we confirmed robustness of the analysis framework by applying it to another genome-wide time-series data and identifying a cytoskeleton-related gene cluster (keratin 18, keratin 17, and mitotic spindle positioning) that shows similar correlation. The significant correlation and highly similar changes of mRNA and protein abundance suggests a concerted role of these genes in cellular stress response, which we consider provides an answer to the question of the specific relationships between mRNA and protein in a cell. In addition, our framework for studying the relationship between mRNAs and proteins in a cell will provide a basis for studying specific relationships between mRNA and protein abundance after accounting for potential time delays.

Project description:Transcription signatures have been used to stratify breast cancer patients into clinically distinct subgroups. However, transcription alone does not determine protein expression. Of potentially equal importance for determining the tumor phenotype is the rate at which transcripts are translated to form protein. Protein translation is controlled to a major degree by miRNA, and cancer cells may deregulate the expression of key genes by altering the activity of relevant miRNAs. The importance of miRNA deregulation and the extent to which multiple miRNAs coordinately deregulate key proteins in breast cancer is only partly understood. To gain such insight, we analyzed genome-wide miRNA expression and mRNA/protein expression for a panel of 105 selected cancer related genes in breast carcinomas from 283 patients. The miRNA-mRNA-protein interactome for the selected genes was constructed by modeling protein expression as a joint function of mRNA and miRNA expression, considering the effect of both one miRNA at a time, and all studied miRNAs simultaneously. The interactome represents a map of the global effects of miRNAs on protein expression, capturing direct as well as indirect effects. The results reveal extensive association between miRNA and protein expression as well as coordinated effects of multiple miRNAs on individual proteins. Applying the model onto two other independent primary breast cancer cohorts confirmed the generalizability of key aspects of the interactome map. The mRNA expression profiling of 283 breast cancer samples was performed using the SurePrint G3 Human GE 8x60K one-color microarrays from Agilent (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer’s protocol (One-Color Microarray-Based Gene Expression Analysis, Low Input Quick Amp Labeling, v.6.5, May 2010). For each sample, 100 ng of RNA was amplified and hybridized on the array. Scanning was performed with Agilent Scanner G2565A, using AgilentG3_GX_1Color as profile. Signals were extracted using FE v.10.7.3.1 and protocol GE1_107_Sep09 (Agilent Technologies). Arrays were log2-transformed, quantile normalized and hospital-adjusted by subtracting from each probe value the mean probe value among samples from the same hospital. The Oslo Breast Cancer Consortium (OSBREAC)

Dataset Information

Strong association between mRNA folding strength and protein abundance in S. cerevisiae.

Publications

Strong association between mRNA folding strength and protein abundance in S. cerevisiae.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure