The Problem:
Over 85% of cancer-driving mutations remain undruggable — not because they lack biological importance, but because their protein structures don’t show clear binding sites. These include major oncogenic mutations like KRAS G12D, TP53 Arg248Ser, and MYC Arg58Ala, which are responsible for aggressive, treatment-resistant cancers. Traditional drug discovery fails to detect transient or hidden sites on these targets.

Our Innovation:
This project introduces a first-in-class, AI-assisted platform that reveals cryptic binding pockets — hidden sites that only emerge under specific conditions like mutation, long-timescale dynamics, or water displacement. We use a novel combination of:

• Molecular dynamics simulations (100–200 ns)
  
  

• Metastable state modeling (MSM + tICA)
  
  

• Hydration thermodynamics & water displacement analysis
  
  

• Ligand contact profiling (PLIP)
  
  

• Mutant-specific druggability scoring (CDDS)
  
  

This method has already been validated in multiple high-impact targets and generates mutant-selective binding sites absent in wild-type proteins.

Our Focus:
We focus on real-world, high-priority cancer mutations, including:

• 🧬 KRAS G12D – prevalent in pediatric leukemia and solid tumors
  
  

• 🧬 TP53 Arg248Ser / C238Y – linked to loss of tumor suppressor function
  
  

• 🧬 MYC Arg58Ala – previously undruggable transcription factor
  
  

• 🧬 SRC Kinase (cryptic site outside ATP pocket) – reducing toxicity via allosteric control