Back

Molecular atlas of key food odorants reveals structured aroma organization and enables generative aroma design

Zhang, J.; Xing, H.; Di Pizio, A.; Ke, Q.; Kou, X.; ZHANG, D.

2026-01-23 systems biology
10.64898/2026.01.21.700072 bioRxiv
Show abstract

Food aromas arise from complex combinations of odorants, yet how these combinations are organized across foods to define aroma identity remains unclear. Decoding this "aroma code" could help bridge the sensory gap between traditional foods and sustainable alternatives, which often struggle with off-notes or to replicate consumer-expected flavor profiles. Here we present KFO-Atlas, a molecular atlas of 896 key food odorants curated from 2,282 food aroma formulations. Analysis shows that food aromas are built from sparse and structured sets of odorants. Plant-derived foods span a broad and diverse aroma space, whereas animal-derived foods tend to exhibit more similar odorant sets. In specific cases, distinct plant- and animal-based foods converge on similar odorant compositions through shared reaction pathways. Building on these insights, we develop a generative AI model that produces category-targeted aroma formulations and validate its outputs by blinded human sensory evaluation. As a proof of principle, the model reconstructs meat-like aromas using exclusively plant-derived odorants, demonstrating a data-driven route to address sensory bottlenecks in sustainable food products.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.