Biomarkers associated with cheese quality uncovered by integrative multi-omic analysis

Cheese quality is determined by many dynamic complex interactions that occur between cheese micmbiota, metabolites and milk substrates. Here we report that application of multi-omics and data integration (by Data Integration Analysis for Biomarker discovery using Latent components, DIABLO) is able to identify and rank biomarkers (taxa and metabolites) which could discriminate, at the molecular level, cheddar cheeses of different quality made by the same manufacturer and to the same specifications. Samples of high-quality mature cheddar had higher amounts of proline, histidine, isoleucine and aspartic acid but lower amounts of stearic acid and octadecanol relative to the lower-quality cheddar. Furthermore, Streptococcus species (presumably S. thermophilus) were present in higher abundance relative to Lactococcus lactis in the higher-quality cheddar. Integrative analysis also revealed significant relationships between these multi-omics biomarkers. Together these results highlight the potential for this approach for identifying biomarkers that could be used to discriminate cheeses of varying quality and as a complementary approach to the current manual sensory evaluation of cheese.