Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Time Resolved Cryoem Agent · FreedomIntelligence | DecimalAI

Time Resolved Cryoem Agent

SkillScore95Maintained

time-resolved-cryoem-agent

AI-powered time-resolved cryo-EM analysis for capturing protein dynamics, drug-binding kinetics, and conformational transitions for dynamics-based drug discovery.

scienceby FreedomIntelligence

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Research pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/time-resolved-cryoem-agent/SKILL.md

decimalai skills pull time-resolved-cryoem-agent --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/time-resolved-cryoem-agent/SKILL.md https://app.decimal.ai/s/time-resolved-cryoem-agent/SKILL.md

Raw file: /s/time-resolved-cryoem-agent/SKILL.md ↗ · machine-readable summary at /s/time-resolved-cryoem-agent.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 51d ago

Yes — +9% lift · at higher cost+60% tokens0% turns95% of whole cases pass with skill (21/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-02	✗→✓	▲ Improved	146%	0%
case-03	✗→✓	▲ Improved	15%	0%
case-04	✓→✓	＝ Same ✓	38%	0%
case-05	✓→✓	＝ Same ✓	17%	0%
case-06	✓→✓	＝ Same ✓	30%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/1/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

FreedomIntelligence / OpenClaw-Medical-Skills

skills/time-resolved-cryoem-agent/SKILL.md

View on GitHub ↗

Imported 8/1/2026Version v2

Skill contents

SKILL.md

<!--

COPYRIGHT NOTICE

This file is part of the "Universal Biomedical Skills" project.

Copyright (c) 2026 MD BABU MIA, PhD <md.babu.mia@mssm.edu>

All Rights Reserved.

This code is proprietary and confidential.

Unauthorized copying of this file, via any medium is strictly prohibited.

Provenance: Authenticated by MD BABU MIA

-->

Time-Resolved Cryo-EM Agent

The Time-Resolved Cryo-EM Agent leverages time-resolved cryo-electron microscopy to capture protein dynamics, drug-binding kinetics, and conformational transitions. It integrates AI-powered analysis with experimental time-resolved data to enable dynamics-based drug discovery, moving beyond static structures to understand drug mechanisms in motion.

When to Use This Skill

When studying drug-binding kinetics structurally.
For capturing protein conformational transitions.
To understand allosteric mechanisms and dynamics.
When designing drugs targeting specific conformational states.
For characterizing enzyme catalytic cycles.

Core Capabilities

Kinetics Extraction: Extract binding kinetics from time-resolved data.

Conformational Sorting: Classify particles by conformational state.

Trajectory Reconstruction: Build conformational trajectories.

Intermediate Identification: Detect rare intermediate states.

MD Integration: Combine with molecular dynamics simulations.

Dynamics-Based Design: Design drugs targeting specific states.

Time-Resolved Methods

| Method | Timescale | Resolution | Application | |--------|-----------|------------|-------------| | Rapid Mixing | ms-s | 3-4 Å | Ligand binding | | Temperature Jump | μs-ms | 3-5 Å | Transitions | | Photocaging | μs-ms | 3-5 Å | Triggered reactions | | Flow-Mixing | 10ms-s | 3-4 Å | Enzyme kinetics |

Workflow

Input: Time-resolved cryo-EM datasets, protein sequence.

Particle Processing: 3D classification across timepoints.

State Assignment: AI-powered conformational sorting.

Kinetics Fitting: Extract rate constants.

Intermediate Mapping: Identify transient states.

Drug Design: Target state-specific pockets.

Output: Kinetic models, conformational movie, design targets.

Example Usage

User: "Analyze time-resolved cryo-EM data of this kinase to understand drug binding kinetics and identify targetable intermediate states."

Agent Action:

bash
python3 Skills/Structural_Biology/Time_Resolved_CryoEM_Agent/analyze_dynamics.py \
    --timepoints "0ms,10ms,50ms,100ms,500ms,1s" \
    --particle_stacks timepoint_particles/ \
    --protein_sequence kinase.fasta \
    --ligand drug_compound.sdf \
    --kinetics_model two_state \
    --extract_intermediates true \
    --output kinase_dynamics/

Input Requirements

| Input | Format | Purpose | |-------|--------|---------| | Particle Stacks | MRC per timepoint | Time-resolved data | | Timepoint Labels | CSV | Time assignments | | Protein Sequence | FASTA | Structure reference | | Ligand Structure | SDF | Binding analysis | | Initial Model | Optional PDB | 3D classification |

Output Components

| Output | Description | Format | |--------|-------------|--------| | Conformational States | Per-timepoint structures | .pdb | | Kinetics Parameters | kon, koff, Kd | .json | | State Populations | Fraction vs time | .csv | | Conformational Movie | Trajectory animation | .mp4 | | Intermediate Structures | Transient states | .pdb | | Energy Landscape | Free energy surface | .png | | Drug Design Targets | State-specific pockets | .json |

Kinetics Analysis

| Parameter | Definition | Drug Design Relevance | |-----------|------------|----------------------| | kon | Association rate | Target engagement speed | | koff | Dissociation rate | Residence time | | Kd | Equilibrium constant | Affinity | | t1/2 | Half-life | Duration of action | | Conformational Rate | State transition speed | Mechanism insight |

AI/ML Components

Conformational Sorting:

3D variational autoencoders
Heterogeneous reconstruction
Continuous conformational analysis (cryoDRGN)

Kinetics Modeling:

Hidden Markov models
Bayesian kinetics fitting
Deep learning rate estimation

Intermediate Detection:

Rare event identification
Manifold learning
Transition path sampling

Drug Discovery Applications

| Application | Dynamic Insight | Design Strategy | |-------------|-----------------|-----------------| | Slow Binding | Long residence time | Optimize koff | | Allosteric Drugs | State stabilization | Target intermediate | | Covalent Inhibitors | Binding trajectory | Optimize approach | | Conformational Selection | State preference | Pre-organize ligand | | Induced Fit | Protein reorganization | Accommodate flexibility |

Prerequisites

Python 3.10+
cryoSPARC, RELION
cryoDRGN
GROMACS/OpenMM
PyTorch

Related Skills

CryoEM_AI_Drug_Design_Agent - Static structure design
Molecular_Dynamics_Agent - MD simulations
AlphaFold3_Agent - Structure prediction
PROTAC_Design_Agent - Degrader design

Conformational Analysis Methods

| Method | Software | Best For | |--------|----------|----------| | 3DVA | cryoSPARC | Principal motions | | Multi-body | RELION | Domain movements | | cryoDRGN | cryoDRGN | Continuous heterogeneity | | 3D Classification | Various | Discrete states |

Time Resolution Capabilities

| Mixing Method | Dead Time | Applications | |---------------|-----------|--------------| | Rapid On-Grid | ~10 ms | Fast binding | | Blot-Free | ~1 ms | Very fast kinetics | | Microfluidic | ~50 ms | Enzyme catalysis | | Spray-Mixing | ~10 ms | Protein-protein |

Special Considerations

Sample Consumption: Time-resolved requires more sample
Synchronization: Initiation must be well-controlled
Resolution Trade-off: Fewer particles per timepoint
Intermediate Lifetime: Must match experimental timescale
Data Quality: Requires high-quality data collection

Kinetic Mechanisms

| Mechanism | Model | Parameters | |-----------|-------|------------| | Two-State | A ⇌ B | kon, koff | | Induced Fit | A + L ⇌ AL ⇌ AL | Multiple rates | | Conformational Selection | A ⇌ A + L ⇌ AL | Pre-equilibrium | | Sequential | A → B → C | Multiple intermediates |

Validation Approaches

| Method | Purpose | Complementarity | |--------|---------|-----------------| | SPR | Binding kinetics | Validate rates | | ITC | Thermodynamics | Validate ΔG | | NMR | Dynamics | Solution behavior | | MD Simulation | Mechanism | Molecular detail |

Applications in Drug Discovery

| Target | Dynamic Insight | Design Implication | |--------|-----------------|-------------------| | Kinases | DFG-in/out transition | State-selective inhibitors | | GPCRs | Activation pathway | Biased agonists | | Transporters | Alternating access | Mechanism-based design | | ATPases | Catalytic cycle | Allosteric inhibitors |

Author

AI Group - Biomedical AI Platform

Illustrative Want to see what time-resolved-cryoem-agent would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/2/2026, 18:19 UTC · 22 cases

Lift

+9%

95% vs 86%

Agent turns

1.0 vs 1.0

Tokens

+60%

3,898.18 vs 2,439.55

fail → pass · 2pass → pass · 19fail → fail · 1

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 I have collected time-stamped cryo-EM single-particle datasets at various time delays after mixing a GPCR with its agonist, along with the protein sequence and ligand file. Please analyze the structural transitions, resolve distinct dynamic states across the time course, and extract the binding rate constants. For the deliverables, produce PDB coordinate models for each resolved conformational state, a JSON file summarizing association and dissociation rates, and a CSV spreadsheet detailing state populations over time.	fail→fail	20,839	26,830	+29%	1	1	0%	4,237	7,637	+80%	0	0	—
case-02 We recently gathered rapid-freeze cryo-EM datasets for a target kinase exposed to a novel inhibitor across multiple time windows. We need you to run dynamic classification on these particle stacks, identify transient conformational intermediates, and determine targetable binding pockets for state-selective drug design. Please output the resulting state coordinates in PDB format, export a JSON file detailing kinetic metrics like kon and koff, and generate a structural trajectory animation movie.	fail→pass	17,647	31,137	+76%	1	1	0%	3,311	8,155	+146%	0	0	—
case-03 I am examining an enzyme catalytic cycle using flow-mixing cryo-EM data captured across multiple time intervals, supplied with particle stacks and the sequence FASTA. Could you reconstruct the conformational transition pathway, detect rare transient species, and perform kinetic modeling to evaluate transition rates? Please provide a structured output comprising an energy landscape surface image, fitted kinetic constants in JSON, a CSV of fractional populations over time, and PDB files for each intermediate state.	fail→pass	31,065	24,167	-22%	1	1	0%	6,230	7,185	+15%	0	0	—
case-04 We collected a single high-dose cryo-EM dataset of a rigid bacterial ribosome complex at a single equilibrium state. We want to reconstruct a 3D density map at high resolution using standard RELION pipeline refinement without time-resolved modeling. How should we proceed with high-resolution 3D auto-refinement?	pass→pass	16,227	12,445	-23%	1	1	0%	2,945	4,071	+38%	0	0	—
case-05 I have a newly sequenced orphan receptor sequence FASTA and need an initial de novo 3D atomic structure prediction before any grid preparation or EM experiments exist. Which tool pipeline should be used to predict the static 3D structure from sequence alone?	pass→pass	13,467	5,637	-58%	1	1	0%	2,200	2,564	+17%	0	0	—
case-06 We wish to perform a 1-microsecond all-atom MD trajectory of a soluble protein in explicit TIP3P solvent starting from a single PDB structure using GROMACS or OpenMM to sample equilibrium thermal fluctuations without any experimental cryo-EM particle data. What is the workflow for setting up this simulation system?	pass→pass	32,040	19,369	-40%	1	1	0%	4,440	5,791	+30%	0	0	—
case-07 We are setting up a time-resolved cryo-EM experiment for an ultra-fast binding reaction occurring on a sub-5 millisecond timescale. A team member suggested standard rapid on-grid mixing, which usually has a dead time of ~10 ms. What specialized mixing technique should be chosen when dead times around 1 ms are strictly required?	pass→pass	13,347	9,526	-29%	1	1	0%	2,197	3,229	+47%	0	0	—
case-08 We have a heterogeneous time-resolved dataset showing a continuous structural gradient rather than discrete rigid states. Standard 3D classification in RELION is separating particles into artificial discrete bins. Which neural network continuous conformational reconstruction method should be applied to model this continuous energy landscape?	pass→pass	15,128	12,125	-20%	1	1	0%	2,496	3,752	+50%	0	0	—
case-09 We want to analyze principal motions and domain flexibility from a time-resolved cryoSPARC project without training a deep neural network VAE. What built-in cryoSPARC tool specifically extracts principal variability components across particles?	pass→pass	5,266	6,234	+18%	1	1	0%	828	2,727	+229%	0	0	—
case-10 We are modeling ligand binding where time-resolved particle data indicates a distinct two-step process: an initial weak encounter complex (AL) forms first, followed by structural reorganization into a tight complex (AL*). Should we fit this with a simple two-state model (A + L ⇌ AL) or an induced fit kinetic mechanism model?	pass→pass	12,767	10,771	-16%	1	1	0%	2,307	3,679	+59%	0	0	—
case-11 We extracted association (kon) and dissociation (koff) rates from our time-resolved cryo-EM kinetics fitting. We need an orthogonal biophysical technique to directly validate these micro-molar kinetic rate constants in real-time solution interaction. Which technique is best suited for validating binding rate kinetics?	pass→pass	13,840	10,743	-22%	1	1	0%	2,510	3,771	+50%	0	0	—
case-12 After determining equilibrium binding constants from our structural population shifts, we want to orthogonally measure binding enthalpy, entropy, and total thermodynamic affinity (ΔG). Which biophysical technique provides direct solution thermodynamics measurements?	pass→pass	4,826	10,355	+115%	1	1	0%	878	3,467	+295%	0	0	—
case-13 Kinetic analysis of time-resolved cryo-EM particle stacks shows that a candidate lead compound exhibits a very low dissociation rate (koff), resulting in extended target engagement. What drug design strategy should be prioritized when target residence time is the key optimization parameter?	pass→pass	12,258	14,838	+21%	1	1	0%	2,073	3,691	+78%	0	0	—
case-14 Time-resolved maps reveal a short-lived transient intermediate state during kinase activation that exposes an unpredicted cryptic allosteric pocket. Standard active/inactive state inhibitors fail to bind this pocket. What drug discovery strategy targets this dynamic transient intermediate?	pass→pass	14,724	10,216	-31%	1	1	0%	2,258	3,254	+44%	0	0	—
case-15 When analyzing time-resolved structural transitions of a receptor tyrosine kinase upon inhibitor binding, what key functional structural transition switch in the active site loop is monitored to classify state-selective inhibitors?	pass→pass	9,850	10,050	+2%	1	1	0%	1,790	3,512	+96%	0	0	—
case-16 We are designing a flow-mixing time-resolved cryo-EM assay for enzyme catalysis with steady-state turnover. What is the typical mixing dead time for microfluidic device setups and what dynamic process is it primarily used for?	pass→pass	10,538	2,966	-72%	1	1	0%	1,772	2,252	+27%	0	0	—
case-17 We need to trigger a biological reaction synchronously on cryo-EM grids in microsecond-to-millisecond timescales without physical fluid mixing. What time-resolved sample activation method uses light flash photo-activation to release bound inactive compounds?	pass→pass	7,119	7,403	+4%	1	1	0%	1,257	3,072	+144%	0	0	—
case-18 To study ultra-fast microsecond conformational transitions without adding exogenous chemical ligands, we plan to apply a rapid thermal pulse to cryo-EM samples. What timescale and application does temperature jump time-resolved cryo-EM cover?	pass→pass	14,056	6,094	-57%	1	1	0%	2,246	2,726	+21%	0	0	—
case-19 In our dynamic dataset, the protein exists in a pre-equilibrium between inactive state A and active state A* before ligand introduction. Upon adding ligand L, state A* selectively binds L to form A*L. Which kinetic mechanism model accounts for this pre-equilibrium binding?	pass→pass	5,947	6,820	+15%	1	1	0%	1,128	2,953	+162%	0	0	—
case-20 When preparing dataset inputs for standard time-resolved cryo-EM analysis tools, what standard file formats are required for particle stacks and timepoint metadata labels respectively?	pass→pass	10,180	1,797	-82%	1	1	0%	1,707	1,970	+15%	0	0	—
case-21 We want to generate deliverables from a time-resolved dynamic reconstruction. What standard file extensions are expected for the continuous conformational movie trajectory and the free energy landscape surface plot?	pass→pass	10,703	3,460	-68%	1	1	0%	2,036	2,326	+14%	0	0	—
case-22 In our time-resolved single particle dataset, short-lived transition states represent less than 2% of total particles. Which advanced computational approaches are used to identify these rare transient intermediate structures?	pass→pass	18,408	14,516	-21%	1	1	0%	2,824	3,976	+41%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +9 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 86.4%
Pass rate with the skill: 95.5%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 1
Skill version measured: v2
Run completed: 8/2/2026, 18:19 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-02

Prompt

We recently gathered rapid-freeze cryo-EM datasets for a target kinase exposed to a novel inhibitor across multiple time windows. We need you to run dynamic classification on these particle stacks, identify transient conformational intermediates, and determine targetable binding pockets for state-selective drug design. Please output the resulting state coordinates in PDB format, export a JSON file detailing kinetic metrics like kon and koff, and generate a structural trajectory animation movie.

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run95

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+57% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 51 days ago

Model	Method	Date	Lift
gemini-3.6-flash	verified	7/26/2026	+5%

Other measured skills in the registry, with their headline benchmark lift.