KRAS G12D Cryptic Pockets

ISEF Category: Biochemistry

Subcategory: Structural Biochemistry  ·  Difficulty: Advanced  ·  Setup: University Lab  ·  Time: Full Year

The Hook

KRAS is one of cancer research’s hardest targets, but it can briefly reveal hidden pockets as it moves. Those short-lived openings can give natural products a place to bind. Your project asks whether molecular dynamics can find those pockets before static docking misses them. That turns a hard protein target into a moving search problem.

What Is It?

A cryptic binding pocket is a site on a protein that does not look open in a single static structure, but appears when the protein flexes. Think of it like a backpack clasp that only pops open when you shake it. In this project, you use molecular dynamics, or MD, which is a computer simulation that follows atoms as they move over time, to watch KRAS G12D change shape and reveal those hidden sites.

KRAS G12D is a cancer-linked version of the KRAS protein. Scientists care about it because blocking it could help stop tumor growth. Natural-product libraries add another layer, because many plant and microbe compounds have unusual shapes that may fit pockets standard drug libraries miss. Your job is to connect protein motion, pocket shape, and compound fit into one clear analysis.

Why This Is a Good Topic

This is a strong science fair topic because it has a clear question, measurable outputs, and a real biomedical use case. You can test whether transient pockets appear, how often they open, and whether natural products rank better in those frames than in a static structure. A student can learn structural biology, docking, basic simulation analysis, and data comparison without needing a wet lab.

Research Questions

• How does simulation length affect the number of transient pockets found in KRAS G12D?
• What is the effect of starting from different KRAS G12D conformations on pocket frequency?
• Does filtering natural products by shape complementarity improve the enrichment of top-scoring poses?
• To what extent do pocket-opening frames change docking scores compared with the starting structure?
• Which physicochemical features of natural products predict binding to transient KRAS G12D pockets?
• How does the chosen force field change pocket stability and pocket volume?

Basic Materials

• Laptop or desktop computer with at least 16 GB RAM.
• Stable internet access for downloading structures and compound files.
• RCSB Protein Data Bank coordinates for KRAS G12D and related KRAS structures.
• PubChem or other public natural-product files in SDF or MOL2 format.
• Spreadsheet software for tracking scores, frame IDs, and compound names.
• External storage or cloud storage for trajectory files and result tables.

Advanced Materials

• Access to a Linux workstation or compute cluster.
• GPU-enabled node for longer molecular dynamics runs.
• Curated natural-product library from PubChem, ZINC, or a similar public source.
• Shared storage for large trajectory files and ensemble outputs.
• Protein prep and parameter files for KRAS G12D analysis.
• Reference KRAS structures for conformational comparison and validation.

Software & Tools

• GROMACS: Runs molecular dynamics simulations and helps you track pocket opening over time.
• OpenMM: Provides an open-source molecular simulation engine with flexible workflows.
• AutoDock Vina: Scores how well each compound fits the pocket frames you extract.
• UCSF ChimeraX: Lets you inspect protein conformations, pocket changes, and docked poses.
• RDKit: Filters natural products and calculates simple molecular features for ranking.

Experiment Steps

1. Define the KRAS G12D structure set you will compare and the pocket metric you will measure.
2. Choose the simulation strategy that will give you repeated views of protein motion, not just one trajectory.
3. Build a compound filtering rule so your natural-product library matches the size and shape of your target pocket.
4. Decide how you will compare docking on a static structure versus docking on transient pocket frames.
5. Set up controls that show whether pocket hits beat random compounds and whether the signal repeats across runs.
6. Pick the final ranking rule before you start analysis so the scoring does not change after you see the results.

Common Pitfalls

• Relying on one starting structure, which misses pockets that only appear after KRAS shifts shape.
• Treating the best docking score as proof of binding, which ignores whether the pose stays inside the pocket.
• Mixing pocket definitions across frames, which makes pocket volume trends hard to compare.
• Skipping negative controls, which leaves you unable to tell whether the natural-product set beats random molecules.
• Running too few simulation repeats, which can make one lucky trajectory look like a real transient pocket.

What Makes This Competitive

A class-level version of this project stops at one docking run and one pocket snapshot. A stronger version uses several trajectories, tracks pocket occupancy and stability, and compares hits against matched decoy compounds. You can raise the level again by testing whether the same natural products rank well across multiple KRAS conformations and by using statistics to show the pattern is not random. That kind of design shows real control over the modeling, not just a lucky result.

Project Variations

• Swap the natural-product library for FDA-approved compounds and see whether known drugs prefer the same transient pockets.
• Compare KRAS G12D with KRAS G12C or wild-type KRAS to test whether the pocket opens only in the mutant.
• Replace docking score alone with a consensus score that includes pocket volume, hydrogen-bond persistence, and pose stability.

Learn More

• RCSB Protein Data Bank: Search KRAS structures and download coordinate files for modeling.
• PubMed: Search review articles on KRAS G12D, cryptic pockets, and molecular dynamics.
• NIH PubChem: Find compound records, 3D conformers, and property data for natural products.
• GROMACS Manual: Read the official tutorials and documentation for simulation setup and analysis.
• OpenMM Documentation: Explore an open-source molecular simulation engine with clear examples.
• MIT OpenCourseWare: Search for computational chemistry and molecular modeling lecture notes.