Machine Learning-Assisted Early
Diagnosis and Smart
Nanotheranostic Profiling for
Cervical Cancer
Introduction
• Cervical cancer is a major cause of death in
women globally.
• High-risk HPV types (especially 16 & 18) are
primary causes.
• Current screening methods (Pap smear, HPV
DNA test) lack sensitivity and precision.
• Need for AI and nanomedicine integration in
early detection.
HPV and Cervical Cancer
Pathophysiology
• E6 and E7 HPV oncoproteins inactivate p53
and Rb tumor suppressors.
• Persistent HPV infection leads to malignancy.
• Integration of viral DNA affects gene
expression and promotes oncogenesis.
Need for Machine Learning
• ML models can analyze high-dimensional
genomic and clinical data.
• Used for predicting HPV progression and
patient risk stratification.
• Algorithms: SVM, Random Forest, XGBoost,
DNN.
Nanobioinformatics and Its Role
• Merges nanotech, bioinformatics, and data
science.
• Simulates drug delivery, ligand-protein
interactions.
• Enables personalized, precise diagnosis and
treatment strategies.
Objectives of the Study
• 1. Evaluate drug efficacy for cervical cancer.
• 2. Study E6/E7-targeted therapies and
vaccines.
• 3. Analyze ADMET of phytocompounds.
• 4. Explore nanocarrier-based drug delivery.
• 5. Bridge biomarker discovery and clinical
application.
Methodology Overview
• Machine learning models for classification.
• GC–MS for compound identification.
• Molecular docking with HPV E6.
• ADMET analysis using SwissADME, ProTox.
• Nanocluster modeling via Marvin Sketch,
Gaussian 09.
Machine Learning Models
• Models: Random Forest, XGBoost, LightGBM,
Logistic Regression, DNN, Linear SVM.
• Applied preprocessing, feature scaling, and
validation.
ML Results
• Decision Tree Classifier:
• - Accuracy: 97%
• - AUC: 0.89
• - High precision-recall balance
• Effective classification of positive and negative
cases with minimal misclassification.
GC–MS Analysis
• Plant: *Withania somnifera*
• Method: Cold maceration extraction
• 19 bioactive compounds identified
• Major compounds: Retinol acetate,
Lycoxanthin, Retinoyl-beta-glucuronide
GC–MS Peak Summary
• Top compounds with retention times and
abundance:
• - Retinol acetate
• - Lycoxanthin
• - Retinoyl-beta-glucuronide
• (Insert chromatogram if available)
Molecular Docking – Target
• Target Protein: HPV-16 E6 (PDB ID: 8B82)
• Docking Software: AutoDock 4.2
• Visualization: PyMOL, Discovery Studio,
LigPlot+.
Docking Results – Key Compound
• Compound: Retinoyl-beta-glucuronide 6',3'-
lactone
• Binding residues: Ile758, Leu780, Gly757
• Hydrogen bonds: 3.19–3.27 Å
• Favorable hydrophobic & electrostatic
interactions.
ADMET Properties
• Predicted via SwissADME & ProTox-3.0
• Toxicity Class: 5 (Low toxicity)
• Mild effects: nephro, cardiac, respiratory
• No GI or BBB permeability → safer therapeutic
profile
ADMET Boiled-Egg Plot
• (Insert boiled-egg plot image from
SwissADME)
• Interpretation: Safe, non-permeable to brain
& gut.
Nanocluster Design
• Si nanocluster modeled in Marvin Sketch
• Optimized using Chem3D & Gaussian 09
• Structure minimized to transition state and
validated.
Si Nanocluster Docking
• Docked with E6 protein (8B82)
• Moderate binding energy
• Visualized using PyMOL & Discovery Studio
• Potential nanocarrier for drug delivery
Future Scope
• HPV vaccination in underserved areas
• Use of liquid biopsy and ctDNA
• Mobile AI diagnostic tools
• Integration of AI, nanomedicine, and
phytotherapy
Challenges
• Poor infrastructure in LMICs
• Cultural resistance to screening
• Lack of centralized cancer registries
• Limited vaccine and radiotherapy access
Conclusion
• Combines ML, phytochemicals, and nanotech
for cervical cancer management.
• Retinoyl-beta-glucuronide & Si nanocluster =
promising theranostics.
• Preclinical validation needed before clinical
use.
