Deep Learning-coupled Proximity Proteomics to Deconvolve Kinase Signaling In Vivo
Jha, K.; Shonai, D.; Parekh, A.; Uezu, A.; Fugiyama, T.; Yamamoto, H.; Parameswaran, P.; Yanagisawa, M.; Singh, R.; Soderling, S. H.
Show abstract
Deconvolving the substrates of hundreds of kinases linked to phosphorylation networks driving cellular behavior is a fundamental, unresolved biological challenge, largely due to the poorly understood interplay of kinase selectivity and substrate proximity. We introduce KolossuS, a deep learning framework leveraging protein language models to decode kinase-substrate specificity. KolossuS achieves superior prediction accuracy and sensitivity across mammalian kinomes, enabling proteome-wide predictions and evolutionary insights. By integrating KolossuS with CRISPR-based proximity proteomics in vivo, we capture kinase-substrate recognition and spatial context, obviating prior limitations. We show this combined framework identifies kinase substrates associated with physiological states such as sleep, revealing both known and novel Sik3 substrates during sleep deprivation. This novel integrated computational-experimental approach promises to transform systematic investigations of kinase signaling in health and disease.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Evolutionary-Scale Enzymology Enables Biochemical Constant Prediction Across a Multi-Peaked Catalytic Landscape 96%
- Mining the CRBN Target Space Redefines Rules for Molecular Glue-induced Neosubstrate Recognition 96%
- OpenCell: proteome-scale endogenous tagging enables the cartography of human cellular organization 96%
Similar papers in this journal
- Synthetic coevolution reveals adaptive mutational trajectories of neutralizing antibodies and SARS-CoV-2 95%
- Active learning of enhancer and silencer regulatory grammar in a developing neural tissue 94%
- Signaling and transcriptional dynamics underlying early adaptation to oncogenic BRAF inhibition 94%
"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.