Novel 4D tensor decomposition-based approach integrating tri-omics profiling data can identify functionally relevant gene clusters

Integrating transcriptome, translatome, and proteome data remains challenging because changes in mRNA, ribosome occupancy, and protein abundance do not always occur simultaneously. To address this, we applied tensor-decomposition-based unsupervised feature extraction to a tri-omics dataset generated under branched-chain amino acid starvation. The three layers were organized into a tensor and analyzed to extract singular value vectors representing coordinated variation across omics, conditions, and replicates. This approach distinguished patterns consistent with ribosome stacking, in which the transcriptome and translatome increase while the proteome decreases, from those of translational buffering, in which the proteome remains stable despite variations in upstream layers. Using components associated with these signatures, we selected 1,781 genes related mainly to reduced translational efficiency and 221 related to buffer-ing. Functional interpretation via enrichment analysis, supported by generative artificial intelligence and a manual literature review, revealed six major biological units: genome replication and maintenance, extracellular matrix remodeling, mitochondrial biogenesis and oxidative phosphorylation, proteostasis and secretion, vesicle transport and signal integration, and epigenetic/RNA regulation. These results indicate that tensor decomposition can effectively integrate triomics data and uncover the biologically meaningful gene clusters and mechanisms underlying cell fate transitions. This framework is valuable for interpreting complex multilayer regulation beyond conventional pairwise analyses.