Unraveling the complex relationship between mRNA and protein abundances: a machine learning-based approach for...

Area of Science:

• Molecular Biology
• Bioinformatics
• Immunology

Background:

• The relationship between messenger RNA (mRNA) and protein levels is complex and debated.
• RNA sequencing (RNA-seq) is a common method for analyzing gene expression, but translating these findings to protein changes remains challenging.
• Understanding this correlation is crucial for interpreting biological dynamics, particularly in areas like immune response and wound healing.

Purpose of the Study:

• To investigate the correlation between mRNA and protein abundances for 17 immune and wound healing mediators in canine volumetric muscle loss.
• To develop and validate a machine learning model for predicting protein abundance from RNA-seq data.
• To assess the model's ability to correct protein assay outliers and detect post-translational modifications.

Main Methods:

• Analysis of mRNA and protein levels for 17 key mediators in canine muscle tissue samples.
• Comparison of mRNA and protein abundance correlations under identical versus varying experimental conditions.
• Development of a machine learning model to predict protein levels from RNA-seq data.

Main Results:

• A general agreement between mRNA and protein levels was observed when analyzing samples from the same experimental condition.
• A lack of correlation was found for individual genes between mRNA and protein levels across different conditions.
• The developed machine learning model achieved high accuracy in predicting protein abundances from RNA-seq data and corrected assay outliers.

Conclusions:

• While mRNA and protein levels show agreement within consistent conditions, predicting protein changes across varying conditions from RNA-seq alone is challenging.
• The novel machine learning approach effectively bridges the gap between RNA-seq and protein abundance, offering a powerful tool for biological interpretation.
• This predictive model holds potential for detecting post-translational modifications, exemplified by accurate estimation of activated transforming growth factor β1.