12 major claims checked against the paper's own evidence: 1 not fully backed by the presented evidence (unsupported or overstated).
overstatedDiscussionReviewer 1
1000ExM could enable direct protein identification in intact cells and tissues.
The paper demonstrates the method on purified proteins and peptides. The claim about in situ identification is speculative and not directly supported by the data presented. The paper acknowledges this as a future application.
Evidence: The paper does not present any data on protein identification in cells or tissues. The Discussion mentions this as a potential future application.
“We highlight one future application that could, in principle, be enabled by 1000ExM: in situ protein identification.”
partialResults, mCLING 2D line-scan analysisReviewer 1
1000ExM can resolve individual amino acid residues.
The paper shows that mCLING peptide images reveal multiple fluorescein peaks that correspond to predicted lysine positions, and GFP reconstructions approach amino acid resolution (11.63 Å). However, direct visualization of individual amino acids in a protein is not demonstrated; the evidence is from a model peptide and a small protein.
Evidence: Figure 3 shows mCLING line scans with multiple peaks corresponding to lysine positions. Figure 5 shows GFP 3D reconstruction with FSC resolution of 11.63 Å.
“With increasing expansion factor, additional fluorescein intensity peaks became separable with respect to one another and with respect to the Atto647N reference”
partialResults, Simulated identifiability of the human proteome under 1000× expansionReviewer 1
Spatial amino acid coordinate patterns alone are sufficient to uniquely identify most proteins in the human proteome.
The paper provides computational simulations showing high identifiability (>99%) with three residue classes under realistic distortion. However, this is a simulation based on assumptions about labeling efficiency and distortion, and the paper acknowledges that in situ identification remains a challenge.
Evidence: Figure 6 and associated text show simulation results for 23,391 human proteins, with >99% identifiability using three residue classes at 5 Å distortion.
“Thus, these simulations demonstrate that near-complete proteome identifiability (>99%) might be achievable with only three chemically addressable residue classes under experimentally measured 1000× expansion distortion”
partialDiscussionReviewer 3
1000ExM may find wide utility in protein visualization and identification, potentially even in intact cells and tissues.
The paper demonstrates the principle on purified proteins and peptides, but does not perform experiments in cells or tissues. The claim is forward-looking and supported by the potential shown, but not directly demonstrated.
Evidence: The paper discusses future applications and the potential for in situ protein identification, but all experiments are on purified molecules.
“We anticipate that 1000ExM will find wide utility in protein visualization and identification, potentially even in intact cells and tissues.”
supportedResults, Development of 1000ExMReviewer 1
1000ExM achieves ~1000-fold linear expansion.
The paper provides direct measurements of gel expansion factors across multiple rounds, showing ~1000× linear expansion after stabilization in PBS.
Evidence: Figure 1 and associated text show expansion factors of ~18×, ~100×, ~500×, and ~1500× across four rounds, stabilizing at ~1000× in PBS.
“In short, sequential casting of four optimized SA/DMAA networks produced 18× → ~100× → ~500× → ~1500× linear expansion without intermediate neutral re-embedding gels (). Upon exposure to phosphate buffered saline (PBS), the final gel stabilized at ~1000× ().”
supportedResults, Proportional scaling of molecular distances across expansion roundsReviewer 1
1000ExM preserves molecular-scale distances and structures.
The paper provides multiple lines of evidence: nanobody fluorophore distances scale proportionally with expansion, mCLING peptide peak positions match MD predictions, and RMSD analysis shows sub-nanometer distortion.
Evidence: Figures 2-4 and associated text show proportional scaling of nanobody distances, agreement of mCLING peak positions with MD models, and RMSD of 0.573 ± 0.151 nm for mCLING peptides.
“Across all expansion factors, measured fluorophore–fluorophore distances scaled proportionally with the physical expansion factor... This normalization yielded an inferred biological distance of ~4 nm, consistent with the expected labeling distance”
supportedAbstractReviewer 1
1000ExM enables sub-nanometer precision on conventional light microscopes.
The RMSD analysis of mCLING peptides shows a mean distortion of 0.573 nm, which is sub-nanometer. This is achieved on a confocal microscope.
Evidence: Figure 4 and associated text show RMSD of 0.573 ± 0.151 nm for mCLING peptides imaged on a confocal microscope.
“Computational analysis indicates that 1000ExM resolves adjacent amino acid residues, thereby achieving sub-nanometer precision on conventional light microscopes.”
supportedResults, paragraph 5Reviewer 2
1000ExM achieves ~1000× linear expansion through a four-network interpenetrating gel architecture.
The paper provides explicit expansion measurements showing stabilization at ~1000× after the fourth round.
Evidence: Figure 2 and text state that after four rounds and exposure to PBS, the gel stabilizes at ~1000× linear expansion.
“Upon exposure to phosphate buffered saline (PBS), the final gel stabilized at ~1000×.”
supportedResults, Proportional scaling... and Evaluation of expansion isotropy...Reviewer 2
Protein and peptide structures are maintained across all expansion factors, as verified by nanobodies, mCLING, and GFP.
The paper provides quantitative evidence: nanobody distances scale linearly, mCLING RMSD matches MD ensemble, and GFP reconstruction approaches atomic resolution.
Evidence: Nanobody distance scaling (Figure 3), mCLING RMSD analysis (0.573 nm), GFP FSC resolution (11.63 Å at 100×).
“Across all expansion factors, measured fluorophore–fluorophore distances scaled proportionally with the physical expansion factor.”
supportedResults, evaluation of expansion isotropy...Reviewer 2
1000ExM resolves adjacent amino acid residues, achieving sub-nanometer precision.
The mCLING experiments show resolved fluorescein peaks correspond to amino acid positions with sub-nanometer RMSD.
Evidence: mCLING 2D line scans reveal peak positions matching MD predictions; 3D RMSD is 0.573 nm (sub-nanometer).
“RMSD values were sub-nanometer, with a mean of 0.573 ± 0.151 nm.”
supportedResults, Simulated identifiability...Reviewers 2, 3
Most proteins could be uniquely identified from their spatial amino acid coordinate patterns alone.
Simulations across the human proteome show >99% identifiability with three residue classes under experimentally measured distortion.
Evidence: Proteome simulation results (Figure 7) show 99.3-99.5% identifiability at 5 Å distortion with three residue classes.
“near-complete proteome identifiability (>99%) might be achievable with only three chemically addressable residue classes”
supportedResults, Development of 1000ExMReviewer 3
1000ExM achieves ~1000-fold linear expansion (one billion-fold in volume).
The paper provides direct measurements of expansion factors across multiple rounds, with graphs showing 18× → ~100× → ~500× → ~1500×, stabilizing at ~1000× in PBS.
Evidence: Figure 1 and associated text report expansion factors at each round and final stabilization. The gel expansion is quantified using physical measurements.
“Across four rounds, sequential casting of four optimized SA/DMAA networks produced 18× → ~100× → ~500× → ~1500× linear expansion without intermediate neutral re-embedding gels. Upon exposure to phosphate buffered saline (PBS), the final gel stabilized at ~1000×.”