PredictProtein

PredictProtein predicts protein abundance by integrating genomic, transcriptomic, and proteomic measurements into a Bayesian network that links transcriptional measurements with post-transcriptional and translational determinants for probabilistic inference of protein-level regulation (described in a 2014 study, PMID: 24532840).

Key Features:

• Data integration: Integrates transcriptomic measurements (e.g., mRNA abundance) with proteomic observations and other genomic-level information to predict protein abundance.
• Bayesian network model: Uses a Bayesian network for probabilistic modeling and inference incorporating determinants such as mRNA abundance, mRNA–protein interaction information, mRNA folding energy, mRNA half-life, and tRNA adaptation.
• Predictive accuracy: Reported analyses in Saccharomyces cerevisiae and Schizosaccharomyces pombe identified approximately twice as many cell-cycle-associated proteins compared with transcript-only baselines.
• Dynamic prediction: Produces protein abundance profiles that are more dynamic than corresponding mRNA expression profiles and that agree with observations in human cell-line datasets.
• Molecular response modeling: Incorporates condition-specific data to model molecular responses over time for studies of adaptation and regulation across conditions.
• mRNA folding energy prediction: Predicts mRNA folding energy when protein abundance data are available to support investigation of secondary-structure effects on translation.

Scientific Applications:

• Cell-cycle studies: Identification and characterization of proteins associated with temporal regulatory programs underlying cell division and growth.
• Comparative genomics: Cross-species analyses, including S. cerevisiae and S. pombe, to compare regulatory determinants of protein abundance.
• Translational efficiency research: Investigation of post-transcriptional control mechanisms, including relationships between mRNA folding and translation output.

Methodology:

Integrates genomic, transcriptomic, and proteomic measurements into a Bayesian network for probabilistic modeling and inference using determinants explicitly including mRNA abundance, mRNA–protein interactions, mRNA folding energy, mRNA half-life, and tRNA adaptation, and predicts mRNA folding energy when protein abundance data are available.