COMPUTER VISION ENABLES EFFECTIVE DETECTION OF COMPOSITIONAL DIFFERENCES IN COMPLEX SAMPLES: VALIDATION OF A WORKFLOW BASED ON CHROMATOGRAPHIC FINGERPRINTING AND PATTERN RECOGNITION

Within Artificial Intelligence (AI) approaches Computer Vision conceptually allows “computers and systems to derive useful information from digital images” giving access to higher-level information and “take actions or make recommendations based on that information” [1]. By comprehensive two-dimensional chromatography, we have access to highly detailed, accurate yet unstructured information on the sample's chemical composition, the possibility to exploit the AI concepts at the data processing level (e.g., by Computer Vision) for rationalizing raw data explorations could quickly drive toward the understanding of the biological phenomena interrelated to a specific/diagnostic chemical signature. The novel workflow for Computer Vision based on pattern recognition algorithms (i.e., UT fingerprinting) includes: 1) Generation of composite class images for representative samples' classes; 2) Generation of a feature template with reliable peaks and peak-regions by processing composite class images from Step 1; 3) Feature template pruning/editing to include targeted components, and eliminate bleeding peaks and artifacts; 4) Application of the feature template from Step 3 to all samples’ images and composite class images from Step 1; 5) Pair-wise comparisons to highlight quantitative pattern differences -> link to the chemistry (targeted compounds) and tracked features (untargeted compounds). As an explanatory application, a sample set originated from a Research Project on artisanal butter (from raw milk to ripened butter) is explored, capturing the evolution of volatile components patterns along the production chain. [1] https://www.ibm.com/topics/computer-vision