A Canonical-State-Conditioned Phosphorylation Residual in the HEK293F EGF Response: Reanalysis, Replicate Robustness, and Independent Cross-Experiment Support

This study presents a source-data reanalysis of time-resolved HEK293F epidermal growth factor (EGF) phosphoproteomics to test whether closely matched canonical EGF signalling states remain distinguishable elsewhere in the phosphoproteome. Using a frozen analysis framework, 49,343 complete phosphosites were examined across eight EGF-response time points. Eleven independently defined canonical EGF phosphosites were used to construct a present-state representation. Within that representation, the 9-minute and 15-minute states formed the closest eligible nonadjacent pair. After conditioning on the dominant canonical-state dimensions, however, their noncanonical phosphoproteomic separation was the largest among the eligible time-point pairs. A subsequent biological-variance-aware inferential reconstruction identified a fixed 20-site noncanonical phosphorylation panel associated with this residual difference. Replicate-level reconstruction showed strong agreement across biological replicates, and an independent HEK293F EGF experiment provided additional cross-experiment support: 14 of the 20 sites were independently measured, and 8 showed significant late-time change, exceeding an effect- and response-matched random expectation. Functional enrichment analyses did not support a specific microtubule, general cytoskeletal, or vesicle-mediated carrier mechanism. Accordingly, the results are interpreted conservatively as evidence for a limited, reproducible late-time noncanonical phosphorylation structure that remains after conditioning on a predefined canonical EGF signalling state. These findings do not establish cellular memory, a hidden molecular carrier, or formal equivalence of the 9- and 15-minute cellular states. The deposit therefore includes a prospective dual-history preregistration designed to test the stronger hypothesis: whether two HEK293F populations with different EGF stimulation histories can be brought to a closely matched canonical signalling state while retaining a reproducible difference along the frozen 20-site residual axis. The archive contains the manuscript, fixed phosphosite panels, frozen canonical PCA reference, analysis code, replicate and cross-experiment validation reports, functional-enrichment results, figures, provenance documentation, and the prospective confirmatory protocol.