Review papers with raw data transparency

Visual paper analysis — Illuminating oncogenic KRAS signaling by multi-dimensional chemical proteomics

Concise visual overview, then targeted critical analysis with recommendations and reproducibility notes. All claims cite the paper.

Key empirical findings (paper-sourced)

• Scale &open-quote;decrypt&close-quote; approach: dose-resolved decryptC (reactive cysteines), decryptM (phosphoproteome), decryptE (protein expression) and ubiquitylome across 3 KRAS-mutant lines producing ~687,954 dose–response curves and covering 25,038 cys-peptides, 69,729 phosphopeptides, 13,093 ubi-peptides and 8,505 proteins Primary dataset summary
• Target selectivity: Sotorasib and Adagrasib potently and selectively modify KRAS C12 in G12C lines (pEC50s consistent with viability EC50s), with only one weak Adagrasib off-target (EEF1A2 C31 at µM) detected by decryptC DecryptC selectivity
• KRAS core phospho-signature: 241 phospho-peptides (252 phospho-sites on 196 proteins) regulated across all three cell lines at 2 h — enriched for proline-directed MAPK motifs (SP/TP) and substrates of ERK/RSK/MNK families; in vitro MAPK1 kinase assays validated several new candidate MAPK1/3 substrates KRAS core phospho-signature
• Immediate vs adaptive separation: dose×time (1–16 h) decryptM shows most KRAS-core sites are immediate (1–2 h) and adaptive sites at 16 h are CDK-related reflecting cell-cycle exit and population shifts rather than direct KRAS→CDK phosphorylation — i.e. PTMs (phospho & ubiquitylation) change extensively while protein-expression changes are modest during quiescence entry Temporal separation of responses
• Proteostasis regulation: KRAS inhibition increases ubiquitylation on E1/E2 machinery (UBA1, UBE2N etc.) near catalytic residues — suggesting regulated attenuation of ubiquitin conjugation accompanies quiescence entry Ubiquitylation of E1/E2 enzymes

Critical appraisal — strengths

• Scale & multi-dimensionality: integrates cysteine-reactivity, phospho-, ubiq-, and proteome-level dose–response data enabling cross-layer inferences (target engagement → pathway engagement → proteostasis) — uncommon depth and direct mechanistic linking Methodological integration description
• Dose-resolved statistics (CurveCurator): stronger for target/off-target identification than single-dose studies — improves discrimination of on-pathway vs off-pathway drug effects.
• Time dimension separates direct kinase-substrate responses from downstream population shifts — a clear methodological advance for interpreting phosphoproteomics in drug studies.

Key limitations & blindspots

• Limited cellular diversity: three cell lines (two pancreatic, one lung) provide depth but limit generality across tumor types, KRAS alleles, co-mutation backgrounds (e.g. STK11/KEAP1/TP53) and microenvironment influences — authors note this limitation Cell-line selection limitation
• In vitro-only context: no in vivo validation (e.g. xenograft phospho-profiling or clinical sample comparison) to confirm which adaptive/protective PTM changes occur in tumors under therapy.
• Data-access caveat: authors state raw data available on request and via ProteomicsDB/Zenodo dashboards; for full reproducibility reviewers will need raw MS files and CurveCurator code/parameters (authors mention these are available on publication) Data availability note
• Functional causality: while many PTMs are correlated with pathway inhibition, only a small subset were mechanistically validated (in vitro kinase assays, phosphomimetic mutagenesis for CCDC86). Many other putative links remain associative and require directed perturbations (site-directed mutants, targeted E3/E2 modulation) to claim mechanistic roles.
• Off-target characterization for non-KRAS drugs: the paper uses decryptM to flag off-targets (e.g. Temuterkib→AAK1/GAK), but follow-up using chemoproteomic orthogonal methods (kinobeads, CETSA) was limited to some probes, leaving potential off-pathway effects incompletely ruled out.

What would convincingly alter conclusions (falsification tests)

1. If orthogonal in vivo tumor models (xenografts or PDXs) showed that the 241-site KRAS core signature is absent or not regulated after KRAS inhibition, it would indicate strong context-dependence and reduce the claim of a common core signature.
2. If CRISPR site-specific mutants of key PTM sites (e.g. validated MAPK substrates or UBA1 K528/K635) failed to affect cell-cycle exit or quiescence behavior under KRAS inhibition, that would weaken claims that those PTMs drive adaptation.
3. If broad proteome reprogramming (transcriptional/protein-expression level) occurred in long-term treated cells (beyond 72 h) and explained adaptive survival, the emphasis on PTMs as the primary adaptive mechanism would need re-evaluation.

Practical next experiments (concise)

• CRISPR-Cas9 knock-in of phospho-null/mimetic mutations for 5 high-confidence KRAS-core sites (e.g. USP10-T74, ADAR-T601, CCDC86-S217) and assess cell-cycle exit, nucleolar integrity, and long-term colony formation ± KRASi.
• Orthogonal target engagement in cells and tissues: CETSA/TPP or kinobeads for Sotorasib/Adagrasib across more lines and PDX-derived material to confirm in vivo selectivity and on-target EC50s.
• Single-cell phosphoproteomics or high-dimensional IF (IMC) after KRASi to test whether adaptive CDK-signature phosphorylation reflects per-cell signaling rewiring or population composition shifts.
• Functional E1/E2 attenuation test: express ubiquitylation-dead UBA1 mutants or block autoubiquitylation to test impact on proteostasis and quiescence entry after KRAS inhibition.

Key source (primary)

Full-study preprint and dataset: Primary paper (2025)