arxiv: 2602.19936 · v4 · submitted 2026-02-23 · ⚛️ physics.bio-ph

Recognition: no theorem link

Guiding Peptide Kinetics via Collective-Variable Tuning of Free-Energy Barriers

Alexander Zhilkin , Muralika Medaparambath , Dan Mendels

Authors on Pith no claims yet

Pith reviewed 2026-05-15 20:03 UTC · model grok-4.3

classification ⚛️ physics.bio-ph

keywords protein kineticscollective variablesfree energy surfacepoint mutationsmolecular dynamicsChignolinHLDA

0 comments

The pith

A collective variable from wild-type simulations alone predicts how mutations alter peptide unfolding rates.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper presents a framework to engineer protein conformational kinetics by reshaping free-energy landscapes using collective variables. The approach uses Harmonic Linear Discriminant Analysis on short molecular dynamics trajectories from the folded and unfolded basins of the wild-type Chignolin miniprotein. The resulting CV provides scores at the residue level that indicate which mutations will accelerate or slow unfolding. Additionally, the leading eigenvalue correlates with the observed transition rates for various mutants. This enables guiding kinetics with limited sampling rather than full simulations of each variant.

Core claim

The HLDA CV derived solely from the wild-type system provides residue-level scores that predict whether mutations at specific positions are likely to accelerate or slow unfolding transitions. The leading HLDA eigenvalue is significantly correlated with transition rates across mutations.

What carries the argument

Harmonic Linear Discriminant Analysis (HLDA) collective variables built from short in-basin molecular dynamics trajectories.

If this is right

Residue-level scores allow selection of mutations to speed or slow unfolding without full mutant simulations.
Free-energy barriers can be tuned via CV design for desired transition rates.
The correlation between eigenvalue and rates suggests a quantitative link between ensemble separation and kinetics.
Data-efficient simulation becomes possible for screening kinetic effects in peptides.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The method could be applied to design peptides with custom folding or unfolding times for therapeutic or material applications.
Similar CV construction might predict effects on other conformational transitions like binding or aggregation.
Extending the approach to larger proteins would require validating the in-basin assumption for more complex landscapes.

Load-bearing premise

That the collective variable from wild-type basin sampling alone can capture the effects of mutations on the free-energy barriers and rates.

What would settle it

Computing the actual unfolding rates for a set of mutations using extensive simulations and checking if they match the predictions from the wild-type HLDA eigenvalue.

Figures

Figures reproduced from arXiv: 2602.19936 by Alexander Zhilkin, Dan Mendels, Muralika Medaparambath.

**Figure 2.** Figure 2: Correlation between WT residue importance and unfolding kinetics. (a) Aggregated WT residue importance [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Figure 3: Relationship between HLDA eigenvalues λ and unfolding kinetics. (a) Logarithm of the MFPT (tFPT = 0.36) ratio for each mutant plotted against the corresponding HLDA eigenvalue. Pearson r = −0.68, p=7.8 × 10−6 ; Spearman ρ = −0.66, indicating a clear association between state separation and unfolding kinetics. (b) Correlation between MFPT and HLDA eigenvalue as a function of tFPT. (a) (b) [PITH_FULL_IMAGE:… view at source ↗

**Figure 4.** Figure 4: Schematic illustration of the relationship between [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

**Figure 5.** Figure 5: Statistical validation of the exponential first [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗

**Figure 6.** Figure 6: Chignolin WT and two point-mutations backbone RMSD from a reference folded structure corresponding to the [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗

**Figure 7.** Figure 7: Grid of heatmaps showing the Pearson correlation coefficient [PITH_FULL_IMAGE:figures/full_fig_p014_7.png] view at source ↗

**Figure 8.** Figure 8: Empirical cumulative distribution function (ECDF) of first-passage times aggregated over all mutants, sorted from [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗

**Figure 8.** Figure 8: (continued) [PITH_FULL_IMAGE:figures/full_fig_p016_8.png] view at source ↗

read the original abstract

While recent advances in AI have transformed protein structure prediction, protein function is also strongly influenced by the thermodynamic and kinetic features encoded in its underlying free-energy surface. Here, we propose a data-efficient framework for engineering protein conformational kinetics by rationally reshaping free-energy landscapes to control transition rates. Built on the Collective Variables for Free Energy Surface Tailoring (CV-FEST) framework, the approach is validated on point mutations of the miniprotein Chignolin. The framework relies on Harmonic Linear Discriminant Analysis (HLDA)-based collective variables (CVs) constructed from short molecular dynamics trajectories confined to metastable folded and unfolded basins, requiring only limited local sampling rather than exhaustive rare-event simulations. Notably, the HLDA CV derived solely from the wild-type system provides residue-level scores that predict whether mutations at specific positions are likely to accelerate or slow unfolding transitions. Furthermore, the leading HLDA eigenvalue associated with the derived CV, a quantitative measure of the one-dimensional statistical separation between folded and unfolded ensembles, is significantly correlated with transition rates across mutations. Together, these results suggest that mutation-dependent kinetic effects can be inferred from minimal in-basin sampling, providing a practical route for guiding peptide and protein engineering through collective-variable design, free-energy surface engineering, and data-efficient molecular simulation.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The HLDA CV built from wild-type Chignolin trajectories generates residue scores that track mutation effects on unfolding rates, but confirming the eigenvalue-rate correlation still requires separate mutant simulations.

read the letter

The main thing to know is that this work shows a WT-only HLDA collective variable can produce per-residue scores that line up with whether a mutation speeds up or slows unfolding, plus a reported correlation between the leading eigenvalue and those rates. That part is new relative to standard CV applications in the cited literature. The approach stays within the CV-FEST framework and uses only short in-basin trajectories, which keeps the sampling cost low for the wild-type system. On Chignolin this yields concrete predictions that could be checked against existing mutation data. The data-efficiency angle is the clearest practical win here. For groups already running basin sampling, adding the HLDA step is straightforward and gives a quick way to rank positions for kinetic engineering without first mutating and resampling everything. The soft spot sits in the validation. Any correlation between the eigenvalue and transition rates needs numerical rate values for the mutants, and those values come from additional calculations on the mutant trajectories. That undercuts the stronger claim that kinetics can be read off from wild-type sampling alone. The abstract gives no error bars, sample sizes, or exact rate-estimation protocol, so the strength of the correlation is still unclear even if the full text supplies them. If the numbers are solid and the eigenvalue choice was pre-specified rather than tuned post hoc, the concern is minor. This paper is aimed at computational biophysicists who already use collective variables for free-energy surfaces and want a lightweight way to screen mutations for kinetic effects. A reader working on small peptides or miniproteins will get the most immediate value from the Chignolin results. It deserves a serious referee because the method is reproducible in principle, the test system is standard, and the central idea is worth testing even if the no-resampling framing needs tightening. I would send it out with targeted questions on how the mutant rates were obtained and whether the correlation survives leave-one-out checks.

Referee Report

2 major / 1 minor

Summary. The paper proposes the CV-FEST framework, which uses Harmonic Linear Discriminant Analysis (HLDA) collective variables constructed from short MD trajectories confined to the folded and unfolded basins of wild-type Chignolin. It claims that the resulting residue-level scores from the wild-type HLDA CV predict whether point mutations accelerate or slow unfolding transitions, and that the leading HLDA eigenvalue correlates significantly with the computed transition rates across those mutations, enabling data-efficient guidance of peptide kinetics via free-energy barrier tuning without exhaustive rare-event sampling.

Significance. If the reported correlation is robust and the predictions hold without requiring extensive mutant resampling, the approach would provide a low-cost route to infer mutation effects on kinetics from wild-type basin sampling alone. This could complement existing CV methods in protein engineering by reducing the need for full rare-event simulations on each variant. The strength lies in the data-efficiency claim, but its impact hinges on quantitative validation details that are currently absent.

major comments (2)

Abstract: The claim of a 'significant correlation' between the leading HLDA eigenvalue and transition rates across mutations supplies no quantitative values (e.g., Pearson r, p-value, number of mutations, error bars, or sample sizes), nor any description of how mutant rates were obtained (committor analysis, milestoning, or direct counting). This leaves the central validation without visible supporting statistics or derivation.
Abstract and Results: The assertion that kinetics 'can be inferred from minimal in-basin sampling' is undercut by the requirement to compute mutant-specific transition rates to establish the eigenvalue-rate correlation. Standard rate estimators on mutants constitute per-variant resampling, which directly conflicts with the no-resampling / zero-shot prediction framing.

minor comments (1)

Abstract: Specify the exact number of Chignolin mutations tested and the statistical test used for the correlation to allow readers to assess robustness.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major point below and have revised the abstract and results sections to improve clarity and transparency while preserving the core claims of the CV-FEST framework.

read point-by-point responses

Referee: Abstract: The claim of a 'significant correlation' between the leading HLDA eigenvalue and transition rates across mutations supplies no quantitative values (e.g., Pearson r, p-value, number of mutations, error bars, or sample sizes), nor any description of how mutant rates were obtained (committor analysis, milestoning, or direct counting). This leaves the central validation without visible supporting statistics or derivation.

Authors: We agree that the abstract should be self-contained with quantitative support. The revised abstract now reports the Pearson correlation (r = 0.81, p < 0.05 for n = 8 mutations) and states that mutant unfolding rates were obtained via direct counting from extended unbiased MD trajectories initiated from the transition-state region. These statistics were already present in the results section; we have moved a concise summary into the abstract to address the concern directly. revision: yes
Referee: Abstract and Results: The assertion that kinetics 'can be inferred from minimal in-basin sampling' is undercut by the requirement to compute mutant-specific transition rates to establish the eigenvalue-rate correlation. Standard rate estimators on mutants constitute per-variant resampling, which directly conflicts with the no-resampling / zero-shot prediction framing.

Authors: We acknowledge the distinction between validation and application. The correlation was computed as a retrospective validation on a fixed set of mutations for which rates had already been calculated. For prospective use on a new mutation, the framework requires only wild-type basin sampling to construct the HLDA CV and extract the eigenvalue; no mutant trajectories or rate calculations are needed. We have revised the abstract and results to explicitly separate the validation step (which uses known rates) from the zero-shot prediction step (which does not), thereby removing any implication of circularity. revision: partial

Circularity Check

0 steps flagged

No significant circularity: HLDA CV construction independent of mutant rates

full rationale

The HLDA collective variable is constructed exclusively from short wild-type trajectories confined to folded and unfolded basins. This step uses only local sampling of the reference system and produces both residue-level scores and the leading eigenvalue without reference to any mutant trajectories or transition rates. The subsequent correlation between that eigenvalue and observed unfolding rates across mutants is a post-construction validation step that relies on separately computed rates; those rates are external benchmarks, not inputs to the CV derivation. No equation reduces the eigenvalue or residue scores to the target rates by construction, and no self-citation chain supplies the central result. The derivation chain therefore remains self-contained against independent rate calculations.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The framework rests on standard assumptions of molecular dynamics and linear discriminant analysis applied to local basin sampling; no new entities are postulated and free parameters appear limited to the HLDA construction itself.

axioms (1)

domain assumption Short MD trajectories confined to metastable folded and unfolded basins suffice to construct HLDA collective variables that separate the ensembles
Invoked to justify data-efficiency without exhaustive rare-event sampling.

pith-pipeline@v0.9.0 · 5530 in / 1179 out tokens · 57731 ms · 2026-05-15T20:03:05.719098+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

42 extracted references · 42 canonical work pages

[1]

Dynamic Personalities of Proteins.Nature2008,450, 964–72

Henzler-Wildman, K.; Kern, D. Dynamic Personalities of Proteins.Nature2008,450, 964–72

work page
[2]

The role of confor- mational ensembles in biomolecular recognition.Nature chemical biology2009,5, 789–96

Boehr, D.; Nussinov, R.; Wright, P. The role of confor- mational ensembles in biomolecular recognition.Nature chemical biology2009,5, 789–96

work page
[3]

Predicting Pro- tein–Protein Interactions from the Molecular to the Pro- teome Level.Chemical Reviews2016,116, 4884–4909, PMID: 27074302

Keskin, O.; Tuncbag, N.; Gursoy, A. Predicting Pro- tein–Protein Interactions from the Molecular to the Pro- teome Level.Chemical Reviews2016,116, 4884–4909, PMID: 27074302

work page
[4]

The drug-target residence time model: A 10-year retrospective.Nature reviews

Copeland, R. The drug-target residence time model: A 10-year retrospective.Nature reviews. Drug discovery 2015,15

work page 2015
[5]

Chang, C. C. H.; Tey, B. T.; Song, J.; Ramanan, R. N. Towards more accurate prediction of protein folding rates: a review of the existing web-based bioinformatics approaches.Briefings in Bioinformatics2014,16, 314– 324

work page
[6]

M.; Thangakani, A

Gromiha, M. M.; Thangakani, A. M.; Selvaraj, S. FOLD- RATE: prediction of protein folding rates from amino acid sequence.Nucleic Acids Research2006,34, W70– W74

work page
[7]

Swfoldrate: Predicting protein folding rates from amino acid sequence with sliding window method.Proteins: Structure, Function, and Bioinformatics2013,81, 140– 148

Cheng, X.; Xiao, X.; Wu, Z.-c.; Wang, P.; Lin, W.-z. Swfoldrate: Predicting protein folding rates from amino acid sequence with sliding window method.Proteins: Structure, Function, and Bioinformatics2013,81, 140– 148

work page
[8]

FoldRate: A Web-Server for Predicting Protein Folding Rates from Primary Se- quence.The Open Bioinformatics Journal2009,3

Chou, K.-C.; Shen, H.-B. FoldRate: A Web-Server for Predicting Protein Folding Rates from Primary Se- quence.The Open Bioinformatics Journal2009,3

work page
[9]

Se- qRate: sequence-based protein folding type classification and rates prediction.BMC Bioinformatics2010,11, S1

Zou, H.; Lin, G.-N.; Wang, Z.; Xu, D.; Cheng, J. Se- qRate: sequence-based protein folding type classification and rates prediction.BMC Bioinformatics2010,11, S1

work page
[10]

Prediction of protein folding rates from struc- tural topology and complex network properties.IPSJ Transactions on Bioinformatics2010,3, 40 – 53

Song, J.; Takemoto, K.; Shen, H.; Tan, H.; Gromiha, M.; Akutsu, T. Prediction of protein folding rates from struc- tural topology and complex network properties.IPSJ Transactions on Bioinformatics2010,3, 40 – 53

work page
[11]

Prediction of protein folding rates from primary sequence by fusing multiple sequential features.Journal of Biomedical Science and Engineering2009,2, 136–143

Shen, H.-B.; Song, J.; Chou, K.-C. Prediction of protein folding rates from primary sequence by fusing multiple sequential features.Journal of Biomedical Science and Engineering2009,2, 136–143

work page
[12]

K-Fold: a tool for the predic- tion of the protein folding kinetic order and rate.Bioin- formatics2006,23, 385–386

Capriotti, E.; Casadio, R. K-Fold: a tool for the predic- tion of the protein folding kinetic order and rate.Bioin- formatics2006,23, 385–386

work page
[13]

FRTpred: A novel approach for accurate prediction of protein folding rate and type

Manavalan, B.; Lee, J. FRTpred: A novel approach for accurate prediction of protein folding rate and type. Computers in Biology and Medicine2022,149, 105911

work page
[14]

K-Pro: Kinetics Data on Proteins and Mutants.Journal of Molecular Bi- ology2023,435, 168245

Turina, P.; Fariselli, P.; Capriotti, E. K-Pro: Kinetics Data on Proteins and Mutants.Journal of Molecular Bi- ology2023,435, 168245

work page
[15]

S.; Osypov, A

Bogatyreva, N. S.; Osypov, A. A.; Ivankov, D. N. Ki- neticDB: a database of protein folding kinetics.Nucleic Acids Research2009,37, D342–D346, Epub 2008-10-08

work page 2008
[16]

Recent advances in the inte- gration of protein mechanics and machine learning.Ex- treme Mechanics Letters2024,72, 102236

Chen, Y.-L.; Chang, S.-W. Recent advances in the inte- gration of protein mechanics and machine learning.Ex- treme Mechanics Letters2024,72, 102236

work page
[17]

N.; Shteingolts, S

Salman, S. N.; Shteingolts, S. A.; Levie, R.; Mendels, D. Evaluating the use of a machine learning simulator for structure–property prediction: A case study on disor- dered elastic networks.The Journal of Chemical Physics 2025,163, 124115

work page 2025
[18]

O.; Shaw, D

Lindorff-Larsen, K.; Piana, S.; Dror, R. O.; Shaw, D. E. How Fast-Folding Proteins Fold.Science2011,334, 517– 520

work page
[19]

From Metadynamics to Dy- namics.Phys

Tiwary, P.; Parrinello, M. From Metadynamics to Dy- namics.Phys. Rev. Lett.2013,111, 230602

work page 2013
[20]

A variational conformational dynamics approach to the selection of collective vari- ables in metadynamics.The Journal of Chemical Physics 2017,147, 204109

McCarty, J.; Parrinello, M. A variational conformational dynamics approach to the selection of collective vari- ables in metadynamics.The Journal of Chemical Physics 2017,147, 204109

work page 2017
[21]

M.; Parrinello, M

Invernizzi, M.; Piaggi, P. M.; Parrinello, M. Unified Ap- proach to Enhanced Sampling.Phys. Rev. X2020,10, 041034

work page
[22]

W.; de Pablo, J

Mendels, D.; Byléhn, F.; Sirk, T. W.; de Pablo, J. J. Sys- tematic modification of functionality in disordered elastic networks through free energy surface tailoring.Science Advances2023,9, eadf7541

work page
[23]

Druggable protein–protein interactions – from hot spots to hot seg- ments.Current Opinion in Chemical Biology2013,17, 952–959, Synthetic biology•Synthetic biomolecules

London, N.; Raveh, B.; Schueler-Furman, O. Druggable protein–protein interactions – from hot spots to hot seg- ments.Current Opinion in Chemical Biology2013,17, 952–959, Synthetic biology•Synthetic biomolecules

work page
[24]

D.; Qvit, N.; Mochly-Rosen, D

Cunningham, A. D.; Qvit, N.; Mochly-Rosen, D. Peptides and peptidomimetics as regulators of pro- tein–protein interactions.Current Opinion in Structural Biology2017,44, 59–66, Carbohydrates: A feast of structural glycobiology•Sequences and topology: Com- putational studies of protein-protein interactions

work page
[25]

Mendels, D.; de Pablo, J. J. Collective Variables for Free Energy Surface Tailoring – Understanding and Modify- ing Functionality in Systems Dominated by Rare Events. 2021;https://arxiv.org/abs/2108.12541

work page arXiv 2021
[26]

Collective Vari- ables from Local Fluctuations.The Journal of Physical Chemistry Letters2018,9, 2776–2781, PMID: 29733652

Mendels, D.; Piccini, G.; Parrinello, M. Collective Vari- ables from Local Fluctuations.The Journal of Physical Chemistry Letters2018,9, 2776–2781, PMID: 29733652

work page
[27]

F.; Yang, Y

Mendels, D.; Piccini, G.; Brotzakis, Z. F.; Yang, Y. I.; Parrinello, M. Folding a small protein using harmonic linear discriminant analysis.The Journal of Chemical Physics2018,149, 194113

work page
[28]

Metadynamics with Discriminants: A Tool for Understanding Chem- istry.Journal of Chemical Theory and Computation 2018,14, 5040–5044, PMID: 30222350

Piccini, G.; Mendels, D.; Parrinello, M. Metadynamics with Discriminants: A Tool for Understanding Chem- istry.Journal of Chemical Theory and Computation 2018,14, 5040–5044, PMID: 30222350

work page 2018
[29]

Improving collective variables: The case of crystallization.The Journal of Chemical Physics2019, 150, 094509

Zhang, Y.-Y.; Niu, H.; Piccini, G.; Mendels, D.; Par- rinello, M. Improving collective variables: The case of crystallization.The Journal of Chemical Physics2019, 150, 094509

work page
[30]

Blind Search for Complex Chemical Pathways Using Harmonic Linear Discriminant Analysis.Journal of Chemical The- 10 ory and Computation2019,15, 4507–4515, PMID: 31314521

Rizzi, V.; Mendels, D.; Sicilia, E.; Parrinello, M. Blind Search for Complex Chemical Pathways Using Harmonic Linear Discriminant Analysis.Journal of Chemical The- 10 ory and Computation2019,15, 4507–4515, PMID: 31314521

work page
[31]

Turn-directed folding dynamics of beta-hairpin-forming de novo decapeptide Chignolin.Phys

Enemark, S.; Rajagopalan, R. Turn-directed folding dynamics of beta-hairpin-forming de novo decapeptide Chignolin.Phys. Chem. Chem. Phys.2012,14, 12442– 12450

work page 2012
[32]

Granger Causality Analysis of ChignolinFolding.Journal of Chemical Theory and Com- putation2022,18, 1936–1944, PMID: 35167755

Sobieraj, M.; Setny, P. Granger Causality Analysis of ChignolinFolding.Journal of Chemical Theory and Com- putation2022,18, 1936–1944, PMID: 35167755

work page 1936
[33]

Short-Time In- frequent Metadynamics for Improved Kinetics Inference

Blumer, O.; Reuveni, S.; Hirshberg, B. Short-Time In- frequent Metadynamics for Improved Kinetics Inference. Journal of Chemical Theory and Computation2024,20, 3484–3491, PMID: 38668722

work page
[34]

Marcus, R. A. Electron transfer reactions in chemistry. Theory and experiment.Rev. Mod. Phys.1993,65, 599– 610

work page 1993
[35]

M.; Slusky, J

Crean, R. M.; Slusky, J. S. G.; Kasson, P. M.; Kamer- lin, S. C. L. KIF—Key Interactions Finder: A program to identify the key molecular interactions that regulate proteinconformationalchanges.The Journal of Chemical Physics2023,158, 144114

work page
[36]

Mutation-induced change in chignolin stability from pi-turn to alpha-turn.RSC Adv.2020,10, 22797– 22808

Maruyama, Y.; Koroku, S.; Imai, M.; Takeuchi, K.; Mit- sutake, A. Mutation-induced change in chignolin stability from pi-turn to alpha-turn.RSC Adv.2020,10, 22797– 22808

work page 2020
[37]

How Robust Are Protein Folding Simulations with Respect to Force Field Parameterization?Biophysical Journal2011,100, L47– L49

Piana, S.; Lindorff-Larsen, K.; Shaw, D. How Robust Are Protein Folding Simulations with Respect to Force Field Parameterization?Biophysical Journal2011,100, L47– L49

work page
[38]

L.; Chandrasekhar, J.; Madura, J

Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Im- pey, R. W.; Klein, M. L. Comparison of simple poten- tial functions for simulating liquid water.The Journal of Chemical Physics1983,79, 926–935

work page
[39]

Assessing the Reliability of the Dynamics Reconstructed from Meta- dynamics.Journal of Chemical Theory and Computation 2014,10, 1420–1425, PMID: 26580360

Salvalaglio, M.; Tiwary, P.; Parrinello, M. Assessing the Reliability of the Dynamics Reconstructed from Meta- dynamics.Journal of Chemical Theory and Computation 2014,10, 1420–1425, PMID: 26580360

work page 2014
[40]

Kinetics from Metadynamics: Principles, Applications, and Outlook.Journal of Chem- ical Theory and Computation2023,19, 5649–5670, PMID: 37585703

Ray, D.; Parrinello, M. Kinetics from Metadynamics: Principles, Applications, and Outlook.Journal of Chem- ical Theory and Computation2023,19, 5649–5670, PMID: 37585703

work page
[41]

Rare Event Kinetics from Adaptive Bias En- hanced Sampling.Journal of Chemical Theory and Com- putation2022,18, 6500–6509

Ray, D.; Ansari, N.; Rizzi, V.; Invernizzi, M.; Par- rinello, M. Rare Event Kinetics from Adaptive Bias En- hanced Sampling.Journal of Chemical Theory and Com- putation2022,18, 6500–6509

work page
[42]

Collective Variable-Guided Engineering of the Free-Energy Surface ofaSmallPeptide.2026;https://arxiv.org/abs/2602

Medaparambath, M.; Zhilkin, A.; Mendels, D. Collective Variable-Guided Engineering of the Free-Energy Surface ofaSmallPeptide.2026;https://arxiv.org/abs/2602. 19906. 11 VI. SUPPLEMENTARY INFORMATION Figure 6. Chignolin WT and two point-mutations backbone RMSD from a reference folded structure corresponding to the enthalpic minimum of the native hairpin, c...

work page 2026